Show More
The requested changes are too big and content was truncated. Show full diff
1 | NO CONTENT: modified file |
|
NO CONTENT: modified file | ||
The requested commit or file is too big and content was truncated. Show full diff |
@@ -1,181 +1,226 b'' | |||||
1 | #!/usr/bin/env python |
|
1 | #!/usr/bin/env python | |
2 | # encoding: utf-8 |
|
2 | # encoding: utf-8 | |
3 | """ |
|
3 | """ | |
4 | Inputhook management for GUI event loop integration. |
|
4 | Inputhook management for GUI event loop integration. | |
5 | """ |
|
5 | """ | |
6 |
|
6 | |||
7 | #----------------------------------------------------------------------------- |
|
7 | #----------------------------------------------------------------------------- | |
8 | # Copyright (C) 2008-2009 The IPython Development Team |
|
8 | # Copyright (C) 2008-2009 The IPython Development Team | |
9 | # |
|
9 | # | |
10 | # Distributed under the terms of the BSD License. The full license is in |
|
10 | # Distributed under the terms of the BSD License. The full license is in | |
11 | # the file COPYING, distributed as part of this software. |
|
11 | # the file COPYING, distributed as part of this software. | |
12 | #----------------------------------------------------------------------------- |
|
12 | #----------------------------------------------------------------------------- | |
13 |
|
13 | |||
14 | #----------------------------------------------------------------------------- |
|
14 | #----------------------------------------------------------------------------- | |
15 | # Imports |
|
15 | # Imports | |
16 | #----------------------------------------------------------------------------- |
|
16 | #----------------------------------------------------------------------------- | |
17 |
|
17 | |||
18 | import ctypes |
|
18 | import ctypes | |
19 | import sys |
|
19 | import sys | |
20 |
|
20 | |||
21 | #----------------------------------------------------------------------------- |
|
21 | #----------------------------------------------------------------------------- | |
22 | # Code |
|
22 | # Code | |
23 | #----------------------------------------------------------------------------- |
|
23 | #----------------------------------------------------------------------------- | |
24 |
|
24 | |||
25 |
|
25 | |||
26 | class InputHookManager(object): |
|
26 | class InputHookManager(object): | |
27 |
"""Manage PyOS_InputHook for different GUI toolkits. |
|
27 | """Manage PyOS_InputHook for different GUI toolkits. | |
|
28 | ||||
|
29 | This class installs various hooks under ``PyOSInputHook`` to handle | |||
|
30 | GUI event loop integration. | |||
|
31 | """ | |||
28 |
|
32 | |||
29 | def __init__(self): |
|
33 | def __init__(self): | |
30 | self.PYFUNC = ctypes.PYFUNCTYPE(ctypes.c_int) |
|
34 | self.PYFUNC = ctypes.PYFUNCTYPE(ctypes.c_int) | |
31 | self._reset() |
|
35 | self._reset() | |
32 |
|
36 | |||
33 | def _reset(self): |
|
37 | def _reset(self): | |
34 | self._callback_pyfunctype = None |
|
38 | self._callback_pyfunctype = None | |
35 | self._callback = None |
|
39 | self._callback = None | |
36 | self._installed = False |
|
40 | self._installed = False | |
|
41 | self._current_gui = None | |||
37 |
|
42 | |||
38 | def get_pyos_inputhook(self): |
|
43 | def get_pyos_inputhook(self): | |
39 | """Return the current PyOS_InputHook as a ctypes.c_void_p. |
|
44 | """Return the current PyOS_InputHook as a ctypes.c_void_p. | |
40 | """ |
|
45 | """ | |
41 | return ctypes.c_void_p.in_dll(ctypes.pythonapi,"PyOS_InputHook") |
|
46 | return ctypes.c_void_p.in_dll(ctypes.pythonapi,"PyOS_InputHook") | |
42 |
|
47 | |||
43 | def get_pyos_inputhook_as_func(self): |
|
48 | def get_pyos_inputhook_as_func(self): | |
44 | """Return the current PyOS_InputHook as a ctypes.PYFUNCYPE. |
|
49 | """Return the current PyOS_InputHook as a ctypes.PYFUNCYPE. | |
45 | """ |
|
50 | """ | |
46 | return self.PYFUNC.in_dll(ctypes.pythonapi,"PyOS_InputHook") |
|
51 | return self.PYFUNC.in_dll(ctypes.pythonapi,"PyOS_InputHook") | |
47 |
|
52 | |||
48 | def set_inputhook(self, callback): |
|
53 | def set_inputhook(self, callback): | |
49 | """Set PyOS_InputHook to callback and return the previous one. |
|
54 | """Set PyOS_InputHook to callback and return the previous one. | |
50 | """ |
|
55 | """ | |
51 | self._callback = callback |
|
56 | self._callback = callback | |
52 | self._callback_pyfunctype = self.PYFUNC(callback) |
|
57 | self._callback_pyfunctype = self.PYFUNC(callback) | |
53 | pyos_inputhook_ptr = self.get_pyos_inputhook() |
|
58 | pyos_inputhook_ptr = self.get_pyos_inputhook() | |
54 | original = self.get_pyos_inputhook_as_func() |
|
59 | original = self.get_pyos_inputhook_as_func() | |
55 | pyos_inputhook_ptr.value = \ |
|
60 | pyos_inputhook_ptr.value = \ | |
56 | ctypes.cast(self._callback_pyfunctype, ctypes.c_void_p).value |
|
61 | ctypes.cast(self._callback_pyfunctype, ctypes.c_void_p).value | |
57 | self._installed = True |
|
62 | self._installed = True | |
58 | return original |
|
63 | return original | |
59 |
|
64 | |||
60 | def clear_inputhook(self): |
|
65 | def clear_inputhook(self): | |
61 | """Set PyOS_InputHook to NULL and return the previous one. |
|
66 | """Set PyOS_InputHook to NULL and return the previous one. | |
62 | """ |
|
67 | """ | |
63 | pyos_inputhook_ptr = self.get_pyos_inputhook() |
|
68 | pyos_inputhook_ptr = self.get_pyos_inputhook() | |
64 | original = self.get_pyos_inputhook_as_func() |
|
69 | original = self.get_pyos_inputhook_as_func() | |
65 | pyos_inputhook_ptr.value = ctypes.c_void_p(None).value |
|
70 | pyos_inputhook_ptr.value = ctypes.c_void_p(None).value | |
66 | self._reset() |
|
71 | self._reset() | |
67 | return original |
|
72 | return original | |
68 |
|
73 | |||
69 | def enable_wx(self, app=False): |
|
74 | def enable_wx(self, app=False): | |
70 | """Enable event loop integration with wxPython. |
|
75 | """Enable event loop integration with wxPython. | |
71 |
|
76 | |||
|
77 | Parameters | |||
|
78 | ---------- | |||
|
79 | app : bool | |||
|
80 | Create a running application object or not. | |||
|
81 | ||||
|
82 | Notes | |||
|
83 | ----- | |||
72 | This methods sets the PyOS_InputHook for wxPython, which allows |
|
84 | This methods sets the PyOS_InputHook for wxPython, which allows | |
73 | the wxPython to integrate with terminal based applications like |
|
85 | the wxPython to integrate with terminal based applications like | |
74 | IPython. |
|
86 | IPython. | |
75 |
|
87 | |||
76 | Once this has been called, you can use wx interactively by doing:: |
|
88 | Once this has been called, you can use wx interactively by doing:: | |
77 |
|
89 | |||
78 | >>> import wx |
|
90 | >>> import wx | |
79 | >>> app = wx.App(redirect=False, clearSigInt=False) |
|
91 | >>> app = wx.App(redirect=False, clearSigInt=False) | |
80 |
|
92 | |||
81 | Both options this constructor are important for things to work |
|
93 | Both options this constructor are important for things to work | |
82 | properly in an interactive context. |
|
94 | properly in an interactive context. | |
83 |
|
95 | |||
84 | But, *don't start the event loop*. That is handled automatically by |
|
96 | But, *don't start the event loop*. That is handled automatically by | |
85 | PyOS_InputHook. |
|
97 | PyOS_InputHook. | |
86 | """ |
|
98 | """ | |
87 | from IPython.lib.inputhookwx import inputhook_wx |
|
99 | from IPython.lib.inputhookwx import inputhook_wx | |
88 | self.set_inputhook(inputhook_wx) |
|
100 | self.set_inputhook(inputhook_wx) | |
|
101 | self._current_gui = 'wx' | |||
89 | if app: |
|
102 | if app: | |
90 | import wx |
|
103 | import wx | |
91 | app = wx.App(redirect=False, clearSigInt=False) |
|
104 | app = wx.App(redirect=False, clearSigInt=False) | |
92 | return app |
|
105 | return app | |
93 |
|
106 | |||
94 | def disable_wx(self): |
|
107 | def disable_wx(self): | |
95 | """Disable event loop integration with wxPython. |
|
108 | """Disable event loop integration with wxPython. | |
96 |
|
109 | |||
97 | This merely sets PyOS_InputHook to NULL. |
|
110 | This merely sets PyOS_InputHook to NULL. | |
98 | """ |
|
111 | """ | |
99 | self.clear_inputhook() |
|
112 | self.clear_inputhook() | |
100 |
|
113 | |||
101 | def enable_qt4(self, app=False): |
|
114 | def enable_qt4(self, app=False): | |
102 | """Enable event loop integration with PyQt4. |
|
115 | """Enable event loop integration with PyQt4. | |
103 |
|
116 | |||
|
117 | Parameters | |||
|
118 | ---------- | |||
|
119 | app : bool | |||
|
120 | Create a running application object or not. | |||
|
121 | ||||
|
122 | Notes | |||
|
123 | ----- | |||
104 | This methods sets the PyOS_InputHook for wxPython, which allows |
|
124 | This methods sets the PyOS_InputHook for wxPython, which allows | |
105 | the PyQt4 to integrate with terminal based applications like |
|
125 | the PyQt4 to integrate with terminal based applications like | |
106 | IPython. |
|
126 | IPython. | |
107 |
|
127 | |||
108 | Once this has been called, you can simply create a QApplication and |
|
128 | Once this has been called, you can simply create a QApplication and | |
109 | use it. But, *don't start the event loop*. That is handled |
|
129 | use it. But, *don't start the event loop*. That is handled | |
110 | automatically by PyOS_InputHook. |
|
130 | automatically by PyOS_InputHook. | |
111 | """ |
|
131 | """ | |
112 | from PyQt4 import QtCore |
|
132 | from PyQt4 import QtCore | |
113 | # PyQt4 has had this since 4.3.1. In version 4.2, PyOS_InputHook |
|
133 | # PyQt4 has had this since 4.3.1. In version 4.2, PyOS_InputHook | |
114 | # was set when QtCore was imported, but if it ever got removed, |
|
134 | # was set when QtCore was imported, but if it ever got removed, | |
115 | # you couldn't reset it. For earlier versions we can |
|
135 | # you couldn't reset it. For earlier versions we can | |
116 | # probably implement a ctypes version. |
|
136 | # probably implement a ctypes version. | |
117 | try: |
|
137 | try: | |
118 | QtCore.pyqtRestoreInputHook() |
|
138 | QtCore.pyqtRestoreInputHook() | |
119 | except AttributeError: |
|
139 | except AttributeError: | |
120 | pass |
|
140 | pass | |
|
141 | self._current_gui = 'qt4' | |||
121 | if app: |
|
142 | if app: | |
122 | from PyQt4 import QtGui |
|
143 | from PyQt4 import QtGui | |
123 | app = QtGui.QApplication(sys.argv) |
|
144 | app = QtGui.QApplication(sys.argv) | |
124 | return app |
|
145 | return app | |
125 |
|
146 | |||
126 | def disable_qt4(self): |
|
147 | def disable_qt4(self): | |
127 | """Disable event loop integration with PyQt4. |
|
148 | """Disable event loop integration with PyQt4. | |
128 |
|
149 | |||
129 | This merely sets PyOS_InputHook to NULL. |
|
150 | This merely sets PyOS_InputHook to NULL. | |
130 | """ |
|
151 | """ | |
131 | self.clear_inputhook() |
|
152 | self.clear_inputhook() | |
132 |
|
153 | |||
133 | def enable_gtk(self, app=False): |
|
154 | def enable_gtk(self, app=False): | |
134 | """Enable event loop integration with PyGTK. |
|
155 | """Enable event loop integration with PyGTK. | |
135 |
|
156 | |||
|
157 | Parameters | |||
|
158 | ---------- | |||
|
159 | app : bool | |||
|
160 | Create a running application object or not. | |||
|
161 | ||||
|
162 | Notes | |||
|
163 | ----- | |||
136 | This methods sets the PyOS_InputHook for PyGTK, which allows |
|
164 | This methods sets the PyOS_InputHook for PyGTK, which allows | |
137 | the PyGTK to integrate with terminal based applications like |
|
165 | the PyGTK to integrate with terminal based applications like | |
138 | IPython. |
|
166 | IPython. | |
139 |
|
167 | |||
140 | Once this has been called, you can simple create PyGTK objects and |
|
168 | Once this has been called, you can simple create PyGTK objects and | |
141 | use them. But, *don't start the event loop*. That is handled |
|
169 | use them. But, *don't start the event loop*. That is handled | |
142 | automatically by PyOS_InputHook. |
|
170 | automatically by PyOS_InputHook. | |
143 | """ |
|
171 | """ | |
144 | import gtk |
|
172 | import gtk | |
145 | try: |
|
173 | try: | |
146 | gtk.set_interactive(True) |
|
174 | gtk.set_interactive(True) | |
|
175 | self._current_gui = 'gtk' | |||
147 | except AttributeError: |
|
176 | except AttributeError: | |
148 | # For older versions of gtk, use our own ctypes version |
|
177 | # For older versions of gtk, use our own ctypes version | |
149 | from IPython.lib.inputhookgtk import inputhook_gtk |
|
178 | from IPython.lib.inputhookgtk import inputhook_gtk | |
150 | add_inputhook(inputhook_gtk) |
|
179 | add_inputhook(inputhook_gtk) | |
151 |
|
180 | |||
152 | def disable_gtk(self): |
|
181 | def disable_gtk(self): | |
153 | """Disable event loop integration with PyGTK. |
|
182 | """Disable event loop integration with PyGTK. | |
154 |
|
183 | |||
155 | This merely sets PyOS_InputHook to NULL. |
|
184 | This merely sets PyOS_InputHook to NULL. | |
156 | """ |
|
185 | """ | |
157 | self.clear_inputhook() |
|
186 | self.clear_inputhook() | |
158 |
|
187 | |||
159 | def enable_tk(self, app=False): |
|
188 | def enable_tk(self, app=False): | |
160 | # Creating a Tkinter.Tk object sets PyOS_InputHook() |
|
189 | """Enable event loop integration with Tk. | |
161 | pass |
|
190 | ||
|
191 | Parameters | |||
|
192 | ---------- | |||
|
193 | app : bool | |||
|
194 | Create a running application object or not. | |||
|
195 | ||||
|
196 | Notes | |||
|
197 | ----- | |||
|
198 | Currently this is a no-op as creating a :class:`Tkinter.Tk` object | |||
|
199 | sets ``PyOS_InputHook``. | |||
|
200 | """ | |||
|
201 | self._current_gui = 'tk' | |||
162 |
|
202 | |||
163 | def disable_tk(self): |
|
203 | def disable_tk(self): | |
164 | """Disable event loop integration with Tkinter. |
|
204 | """Disable event loop integration with Tkinter. | |
165 |
|
205 | |||
166 | This merely sets PyOS_InputHook to NULL. |
|
206 | This merely sets PyOS_InputHook to NULL. | |
167 | """ |
|
207 | """ | |
168 | self.clear_inputhook() |
|
208 | self.clear_inputhook() | |
169 |
|
209 | |||
|
210 | def current_gui(self): | |||
|
211 | """Return a string indicating the currently active GUI or None.""" | |||
|
212 | return self._current_gui | |||
|
213 | ||||
170 | inputhook_manager = InputHookManager() |
|
214 | inputhook_manager = InputHookManager() | |
171 |
|
215 | |||
172 | enable_wx = inputhook_manager.enable_wx |
|
216 | enable_wx = inputhook_manager.enable_wx | |
173 | disable_wx = inputhook_manager.disable_wx |
|
217 | disable_wx = inputhook_manager.disable_wx | |
174 | enable_qt4 = inputhook_manager.enable_qt4 |
|
218 | enable_qt4 = inputhook_manager.enable_qt4 | |
175 | disable_qt4 = inputhook_manager.disable_qt4 |
|
219 | disable_qt4 = inputhook_manager.disable_qt4 | |
176 | enable_gtk = inputhook_manager.enable_gtk |
|
220 | enable_gtk = inputhook_manager.enable_gtk | |
177 | disable_gtk = inputhook_manager.disable_gtk |
|
221 | disable_gtk = inputhook_manager.disable_gtk | |
178 | enable_tk = inputhook_manager.enable_tk |
|
222 | enable_tk = inputhook_manager.enable_tk | |
179 | disable_tk = inputhook_manager.disable_tk |
|
223 | disable_tk = inputhook_manager.disable_tk | |
180 | clear_inputhook = inputhook_manager.clear_inputhook |
|
224 | clear_inputhook = inputhook_manager.clear_inputhook | |
181 | set_inputhook = inputhook_manager.set_inputhook No newline at end of file |
|
225 | set_inputhook = inputhook_manager.set_inputhook | |
|
226 | current_gui = inputhook_manager.current_gui No newline at end of file |
@@ -1,34 +1,36 b'' | |||||
1 | #!/usr/bin/env python |
|
1 | #!/usr/bin/env python | |
2 | """Script to auto-generate our API docs. |
|
2 | """Script to auto-generate our API docs. | |
3 | """ |
|
3 | """ | |
4 | # stdlib imports |
|
4 | # stdlib imports | |
5 | import os |
|
5 | import os | |
6 | import sys |
|
6 | import sys | |
7 |
|
7 | |||
8 | # local imports |
|
8 | # local imports | |
9 | sys.path.append(os.path.abspath('sphinxext')) |
|
9 | sys.path.append(os.path.abspath('sphinxext')) | |
10 | from apigen import ApiDocWriter |
|
10 | from apigen import ApiDocWriter | |
11 |
|
11 | |||
12 | #***************************************************************************** |
|
12 | #***************************************************************************** | |
13 | if __name__ == '__main__': |
|
13 | if __name__ == '__main__': | |
14 | pjoin = os.path.join |
|
14 | pjoin = os.path.join | |
15 | package = 'IPython' |
|
15 | package = 'IPython' | |
16 | outdir = pjoin('source','api','generated') |
|
16 | outdir = pjoin('source','api','generated') | |
17 | docwriter = ApiDocWriter(package,rst_extension='.txt') |
|
17 | docwriter = ApiDocWriter(package,rst_extension='.txt') | |
18 | docwriter.package_skip_patterns += [r'\.fixes$', |
|
18 | docwriter.package_skip_patterns += [r'\.fixes$', | |
19 | r'\.externals$', |
|
19 | r'\.externals$', | |
20 | r'\.extensions', |
|
20 | r'\.extensions', | |
21 | r'\.kernel.config', |
|
21 | r'\.kernel.config', | |
22 | r'\.attic', |
|
22 | r'\.attic', | |
|
23 | r'\.quarantine', | |||
|
24 | r'\.deathrow' | |||
23 | ] |
|
25 | ] | |
24 |
docwriter.module_skip_patterns += [ r'\. |
|
26 | docwriter.module_skip_patterns += [ r'\.core.fakemodule', | |
25 | r'\.cocoa', |
|
27 | r'\.cocoa', | |
26 | r'\.ipdoctest', |
|
28 | r'\.ipdoctest', | |
27 | r'\.Gnuplot', |
|
29 | r'\.Gnuplot', | |
28 | r'\.frontend.process.winprocess', |
|
30 | r'\.frontend.process.winprocess', | |
29 | ] |
|
31 | ] | |
30 | docwriter.write_api_docs(outdir) |
|
32 | docwriter.write_api_docs(outdir) | |
31 | docwriter.write_index(outdir, 'gen', |
|
33 | docwriter.write_index(outdir, 'gen', | |
32 | relative_to = pjoin('source','api') |
|
34 | relative_to = pjoin('source','api') | |
33 | ) |
|
35 | ) | |
34 | print '%d files written' % len(docwriter.written_modules) |
|
36 | print '%d files written' % len(docwriter.written_modules) |
@@ -1,515 +1,517 b'' | |||||
1 | .. _development: |
|
1 | .. _development: | |
2 |
|
2 | |||
3 | ============================== |
|
3 | ============================== | |
4 | IPython development guidelines |
|
4 | IPython development guidelines | |
5 | ============================== |
|
5 | ============================== | |
6 |
|
6 | |||
7 |
|
7 | |||
8 | Overview |
|
8 | Overview | |
9 | ======== |
|
9 | ======== | |
10 |
|
10 | |||
11 | This document describes IPython from the perspective of developers. Most |
|
11 | This document describes IPython from the perspective of developers. Most | |
12 | importantly, it gives information for people who want to contribute to the |
|
12 | importantly, it gives information for people who want to contribute to the | |
13 | development of IPython. So if you want to help out, read on! |
|
13 | development of IPython. So if you want to help out, read on! | |
14 |
|
14 | |||
15 | How to contribute to IPython |
|
15 | How to contribute to IPython | |
16 | ============================ |
|
16 | ============================ | |
17 |
|
17 | |||
18 | IPython development is done using Bazaar [Bazaar]_ and Launchpad [Launchpad]_. |
|
18 | IPython development is done using Bazaar [Bazaar]_ and Launchpad [Launchpad]_. | |
19 | This makes it easy for people to contribute to the development of IPython. |
|
19 | This makes it easy for people to contribute to the development of IPython. | |
20 | There are several ways in which you can join in. |
|
20 | There are several ways in which you can join in. | |
21 |
|
21 | |||
22 | If you have a small change that you want to send to the team, you can edit your |
|
22 | If you have a small change that you want to send to the team, you can edit your | |
23 | bazaar checkout of IPython (see below) in-place, and ask bazaar for the |
|
23 | bazaar checkout of IPython (see below) in-place, and ask bazaar for the | |
24 | differences:: |
|
24 | differences:: | |
25 |
|
25 | |||
26 | $ cd /path/to/your/copy/of/ipython |
|
26 | $ cd /path/to/your/copy/of/ipython | |
27 | $ bzr diff > my_fixes.diff |
|
27 | $ bzr diff > my_fixes.diff | |
28 |
|
28 | |||
29 | This produces a patch file with your fixes, which we can apply to the source |
|
29 | This produces a patch file with your fixes, which we can apply to the source | |
30 | tree. This file should then be attached to a ticket in our `bug tracker |
|
30 | tree. This file should then be attached to a ticket in our `bug tracker | |
31 | <https://bugs.launchpad.net/ipython>`_, indicating what it does. |
|
31 | <https://bugs.launchpad.net/ipython>`_, indicating what it does. | |
32 |
|
32 | |||
33 | This model of creating small, self-contained patches works very well and there |
|
33 | This model of creating small, self-contained patches works very well and there | |
34 | are open source projects that do their entire development this way. However, |
|
34 | are open source projects that do their entire development this way. However, | |
35 | in IPython we have found that for tracking larger changes, making use of |
|
35 | in IPython we have found that for tracking larger changes, making use of | |
36 | bazaar's full capabilities in conjunction with Launchpad's code hosting |
|
36 | bazaar's full capabilities in conjunction with Launchpad's code hosting | |
37 | services makes for a much better experience. |
|
37 | services makes for a much better experience. | |
38 |
|
38 | |||
39 | Making your own branch of IPython allows you to refine your changes over time, |
|
39 | Making your own branch of IPython allows you to refine your changes over time, | |
40 | track the development of the main team, and propose your own full version of |
|
40 | track the development of the main team, and propose your own full version of | |
41 | the code for others to use and review, with a minimum amount of fuss. The next |
|
41 | the code for others to use and review, with a minimum amount of fuss. The next | |
42 | parts of this document will explain how to do this. |
|
42 | parts of this document will explain how to do this. | |
43 |
|
43 | |||
44 | Install Bazaar and create a Launchpad account |
|
44 | Install Bazaar and create a Launchpad account | |
45 | --------------------------------------------- |
|
45 | --------------------------------------------- | |
46 |
|
46 | |||
47 | First make sure you have installed Bazaar (see their `website |
|
47 | First make sure you have installed Bazaar (see their `website | |
48 | <http://bazaar-vcs.org/>`_). To see that Bazaar is installed and knows about |
|
48 | <http://bazaar-vcs.org/>`_). To see that Bazaar is installed and knows about | |
49 | you, try the following:: |
|
49 | you, try the following:: | |
50 |
|
50 | |||
51 | $ bzr whoami |
|
51 | $ bzr whoami | |
52 | Joe Coder <jcoder@gmail.com> |
|
52 | Joe Coder <jcoder@gmail.com> | |
53 |
|
53 | |||
54 | This should display your name and email. Next, you will want to create an |
|
54 | This should display your name and email. Next, you will want to create an | |
55 | account on the `Launchpad website <http://www.launchpad.net>`_ and setup your |
|
55 | account on the `Launchpad website <http://www.launchpad.net>`_ and setup your | |
56 | ssh keys. For more information of setting up your ssh keys, see `this link |
|
56 | ssh keys. For more information of setting up your ssh keys, see `this link | |
57 | <https://help.launchpad.net/YourAccount/CreatingAnSSHKeyPair>`_. |
|
57 | <https://help.launchpad.net/YourAccount/CreatingAnSSHKeyPair>`_. | |
58 |
|
58 | |||
59 | Get the main IPython branch from Launchpad |
|
59 | Get the main IPython branch from Launchpad | |
60 | ------------------------------------------ |
|
60 | ------------------------------------------ | |
61 |
|
61 | |||
62 | Now, you can get a copy of the main IPython development branch (we call this |
|
62 | Now, you can get a copy of the main IPython development branch (we call this | |
63 | the "trunk"):: |
|
63 | the "trunk"):: | |
64 |
|
64 | |||
65 | $ bzr branch lp:ipython |
|
65 | $ bzr branch lp:ipython | |
66 |
|
66 | |||
67 | Create a working branch |
|
67 | Create a working branch | |
68 | ----------------------- |
|
68 | ----------------------- | |
69 |
|
69 | |||
70 | When working on IPython, you won't actually make edits directly to the |
|
70 | When working on IPython, you won't actually make edits directly to the | |
71 | :file:`lp:ipython` branch. Instead, you will create a separate branch for your |
|
71 | :file:`lp:ipython` branch. Instead, you will create a separate branch for your | |
72 | changes. For now, let's assume you want to do your work in a branch named |
|
72 | changes. For now, let's assume you want to do your work in a branch named | |
73 | "ipython-mybranch". Create this branch by doing:: |
|
73 | "ipython-mybranch". Create this branch by doing:: | |
74 |
|
74 | |||
75 | $ bzr branch ipython ipython-mybranch |
|
75 | $ bzr branch ipython ipython-mybranch | |
76 |
|
76 | |||
77 | When you actually create a branch, you will want to give it a name that |
|
77 | When you actually create a branch, you will want to give it a name that | |
78 | reflects the nature of the work that you will be doing in it, like |
|
78 | reflects the nature of the work that you will be doing in it, like | |
79 | "install-docs-update". |
|
79 | "install-docs-update". | |
80 |
|
80 | |||
81 | Make edits in your working branch |
|
81 | Make edits in your working branch | |
82 | --------------------------------- |
|
82 | --------------------------------- | |
83 |
|
83 | |||
84 | Now you are ready to actually make edits in your :file:`ipython-mybranch` |
|
84 | Now you are ready to actually make edits in your :file:`ipython-mybranch` | |
85 | branch. Before doing this, it is helpful to install this branch so you can |
|
85 | branch. Before doing this, it is helpful to install this branch so you can | |
86 | test your changes as you work. This is easiest if you have setuptools |
|
86 | test your changes as you work. This is easiest if you have setuptools | |
87 | installed. Then, just do:: |
|
87 | installed. Then, just do:: | |
88 |
|
88 | |||
89 | $ cd ipython-mybranch |
|
89 | $ cd ipython-mybranch | |
90 | $ python setupegg.py develop |
|
90 | $ python setupegg.py develop | |
91 |
|
91 | |||
92 | Now, make some changes. After a while, you will want to commit your changes. |
|
92 | Now, make some changes. After a while, you will want to commit your changes. | |
93 | This let's Bazaar know that you like the changes you have made and gives you |
|
93 | This let's Bazaar know that you like the changes you have made and gives you | |
94 | an opportunity to keep a nice record of what you have done. This looks like |
|
94 | an opportunity to keep a nice record of what you have done. This looks like | |
95 | this:: |
|
95 | this:: | |
96 |
|
96 | |||
97 | $ ...do work in ipython-mybranch... |
|
97 | $ ...do work in ipython-mybranch... | |
98 | $ bzr commit -m "the commit message goes here" |
|
98 | $ bzr commit -m "the commit message goes here" | |
99 |
|
99 | |||
100 | Please note that since we now don't use an old-style linear ChangeLog (that |
|
100 | Please note that since we now don't use an old-style linear ChangeLog (that | |
101 | tends to cause problems with distributed version control systems), you should |
|
101 | tends to cause problems with distributed version control systems), you should | |
102 | ensure that your log messages are reasonably detailed. Use a docstring-like |
|
102 | ensure that your log messages are reasonably detailed. Use a docstring-like | |
103 | approach in the commit messages (including the second line being left |
|
103 | approach in the commit messages (including the second line being left | |
104 | *blank*):: |
|
104 | *blank*):: | |
105 |
|
105 | |||
106 | Single line summary of changes being committed. |
|
106 | Single line summary of changes being committed. | |
107 |
|
107 | |||
108 | * more details when warranted ... |
|
108 | * more details when warranted ... | |
109 | * including crediting outside contributors if they sent the |
|
109 | * including crediting outside contributors if they sent the | |
110 | code/bug/idea! |
|
110 | code/bug/idea! | |
111 |
|
111 | |||
112 | As you work, you will repeat this edit/commit cycle many times. If you work on |
|
112 | As you work, you will repeat this edit/commit cycle many times. If you work on | |
113 | your branch for a long time, you will also want to get the latest changes from |
|
113 | your branch for a long time, you will also want to get the latest changes from | |
114 | the :file:`lp:ipython` branch. This can be done with the following sequence of |
|
114 | the :file:`lp:ipython` branch. This can be done with the following sequence of | |
115 | commands:: |
|
115 | commands:: | |
116 |
|
116 | |||
117 | $ ls |
|
117 | $ ls | |
118 | ipython |
|
118 | ipython | |
119 | ipython-mybranch |
|
119 | ipython-mybranch | |
120 |
|
120 | |||
121 | $ cd ipython |
|
121 | $ cd ipython | |
122 | $ bzr pull |
|
122 | $ bzr pull | |
123 | $ cd ../ipython-mybranch |
|
123 | $ cd ../ipython-mybranch | |
124 | $ bzr merge ../ipython |
|
124 | $ bzr merge ../ipython | |
125 | $ bzr commit -m "Merging changes from trunk" |
|
125 | $ bzr commit -m "Merging changes from trunk" | |
126 |
|
126 | |||
127 | Along the way, you should also run the IPython test suite. You can do this |
|
127 | Along the way, you should also run the IPython test suite. You can do this | |
128 | using the :command:`iptest` command (which is basically a customized version of |
|
128 | using the :command:`iptest` command (which is basically a customized version of | |
129 | :command:`nosetests`):: |
|
129 | :command:`nosetests`):: | |
130 |
|
130 | |||
131 | $ cd |
|
131 | $ cd | |
132 | $ iptest |
|
132 | $ iptest | |
133 |
|
133 | |||
134 | The :command:`iptest` command will also pick up and run any tests you have |
|
134 | The :command:`iptest` command will also pick up and run any tests you have | |
135 |
written. |
|
135 | written. See :ref:`testing documentation <devel_testing>` for further details | |
|
136 | on the testing system. | |||
136 |
|
137 | |||
137 |
|
138 | |||
138 | Post your branch and request a code review |
|
139 | Post your branch and request a code review | |
139 | ------------------------------------------ |
|
140 | ------------------------------------------ | |
140 |
|
141 | |||
141 | Once you are done with your edits, you should post your branch on Launchpad so |
|
142 | Once you are done with your edits, you should post your branch on Launchpad so | |
142 | that other IPython developers can review the changes and help you merge your |
|
143 | that other IPython developers can review the changes and help you merge your | |
143 | changes into the main development branch. To post your branch on Launchpad, |
|
144 | changes into the main development branch. To post your branch on Launchpad, | |
144 | do:: |
|
145 | do:: | |
145 |
|
146 | |||
146 | $ cd ipython-mybranch |
|
147 | $ cd ipython-mybranch | |
147 | $ bzr push lp:~yourusername/ipython/ipython-mybranch |
|
148 | $ bzr push lp:~yourusername/ipython/ipython-mybranch | |
148 |
|
149 | |||
149 | Then, go to the `IPython Launchpad site <www.launchpad.net/ipython>`_, and you |
|
150 | Then, go to the `IPython Launchpad site <www.launchpad.net/ipython>`_, and you | |
150 | should see your branch under the "Code" tab. If you click on your branch, you |
|
151 | should see your branch under the "Code" tab. If you click on your branch, you | |
151 | can provide a short description of the branch as well as mark its status. Most |
|
152 | can provide a short description of the branch as well as mark its status. Most | |
152 | importantly, you should click the link that reads "Propose for merging into |
|
153 | importantly, you should click the link that reads "Propose for merging into | |
153 | another branch". What does this do? |
|
154 | another branch". What does this do? | |
154 |
|
155 | |||
155 | This let's the other IPython developers know that your branch is ready to be |
|
156 | This let's the other IPython developers know that your branch is ready to be | |
156 | reviewed and merged into the main development branch. During this review |
|
157 | reviewed and merged into the main development branch. During this review | |
157 | process, other developers will give you feedback and help you get your code |
|
158 | process, other developers will give you feedback and help you get your code | |
158 | ready to be merged. What types of things will we be looking for: |
|
159 | ready to be merged. What types of things will we be looking for: | |
159 |
|
160 | |||
160 | * All code is documented. |
|
161 | * All code is documented. | |
161 | * All code has tests. |
|
162 | * All code has tests. | |
162 | * The entire IPython test suite passes. |
|
163 | * The entire IPython test suite passes. | |
163 |
|
164 | |||
164 | Once your changes have been reviewed and approved, someone will merge them |
|
165 | Once your changes have been reviewed and approved, someone will merge them | |
165 | into the main development branch. |
|
166 | into the main development branch. | |
166 |
|
167 | |||
167 |
|
168 | |||
168 | Some notes for core developers when merging third-party contributions |
|
169 | Some notes for core developers when merging third-party contributions | |
169 | ===================================================================== |
|
170 | ===================================================================== | |
170 |
|
171 | |||
171 | Core developers, who ultimately merge any approved branch (from themselves, |
|
172 | Core developers, who ultimately merge any approved branch (from themselves, | |
172 | another developer, or any third-party contribution) will typically use |
|
173 | another developer, or any third-party contribution) will typically use | |
173 |
:command:`bzr merge` to merge the branch into the trunk and push it to the |
|
174 | :command:`bzr merge` to merge the branch into the trunk and push it to the | |
174 |
Launcphad site. |
|
175 | main Launcphad site. This is a short list of things to keep in mind when doing | |
175 |
process, so that the project history is easy to understand in the long |
|
176 | this process, so that the project history is easy to understand in the long | |
176 |
that generating release notes is as painless and accurate as |
|
177 | run, and that generating release notes is as painless and accurate as | |
177 |
|
178 | possible. | ||
178 | - When you merge any non-trivial functionality (from one small bug fix to a big |
|
179 | ||
179 | feature branch), please remember to always edit the changes_ file |
|
180 | - When you merge any non-trivial functionality (from one small bug fix to a | |
180 | accordingly. This file has one main section for each release, and if you |
|
181 | big feature branch), please remember to always edit the :file:`changes.txt` | |
181 | edit it as you go, noting what new features, bug fixes or API changes you |
|
182 | file accordingly. This file has one main section for each release, and if | |
182 | have made, the release notes will be almost finished when they are needed |
|
183 | you edit it as you go, noting what new features, bug fixes or API changes | |
183 | later. This is much easier if done when you merge the work, rather than |
|
184 | you have made, the release notes will be almost finished when they are | |
184 | weeks or months later by re-reading a massive Bazaar log. |
|
185 | needed later. This is much easier if done when you merge the work, rather | |
185 |
|
186 | than weeks or months later by re-reading a massive Bazaar log. | ||
186 | - When big merges are done, the practice of putting a summary commit message in |
|
187 | ||
187 | the merge is *extremely* useful. It makes this kind of job much nicer, |
|
188 | - When big merges are done, the practice of putting a summary commit message | |
|
189 | in the merge is *extremely* useful. It makes this kind of job much nicer, | |||
188 | because that summary log message can be almost copy/pasted without changes, |
|
190 | because that summary log message can be almost copy/pasted without changes, | |
189 | if it was well written, rather than dissecting the next-level messages from |
|
191 | if it was well written, rather than dissecting the next-level messages from | |
190 | the individual commits. |
|
192 | the individual commits. | |
191 |
|
193 | |||
192 | - It's important that we remember to always credit who gave us something if |
|
194 | - It's important that we remember to always credit who gave us something if | |
193 |
it's not the committer. |
|
195 | it's not the committer. In general, we have been fairly good on this front, | |
194 |
this is just a reminder to keep things up. |
|
196 | this is just a reminder to keep things up. As a note, if you are ever | |
195 | committing something that is completely (or almost so) a third-party |
|
197 | committing something that is completely (or almost so) a third-party | |
196 | contribution, do the commit as:: |
|
198 | contribution, do the commit as:: | |
197 |
|
199 | |||
198 | $ bzr commit --author="Someone Else" |
|
200 | $ bzr commit --author="Someone Else" | |
199 |
|
201 | |||
200 | This way it will show that name separately in the log, which makes it even |
|
202 | This way it will show that name separately in the log, which makes it even | |
201 |
easier to spot. |
|
203 | easier to spot. Obviously we often rework third party contributions | |
202 | extensively, but this is still good to keep in mind for cases when we don't |
|
204 | extensively, but this is still good to keep in mind for cases when we don't | |
203 | touch the code too much. |
|
205 | touch the code too much. | |
204 |
|
206 | |||
205 |
|
207 | |||
206 | Documentation |
|
208 | Documentation | |
207 | ============= |
|
209 | ============= | |
208 |
|
210 | |||
209 | Standalone documentation |
|
211 | Standalone documentation | |
210 | ------------------------ |
|
212 | ------------------------ | |
211 |
|
213 | |||
212 | All standalone documentation should be written in plain text (``.txt``) files |
|
214 | All standalone documentation should be written in plain text (``.txt``) files | |
213 | using reStructuredText [reStructuredText]_ for markup and formatting. All such |
|
215 | using reStructuredText [reStructuredText]_ for markup and formatting. All such | |
214 | documentation should be placed in directory :file:`docs/source` of the IPython |
|
216 | documentation should be placed in directory :file:`docs/source` of the IPython | |
215 | source tree. The documentation in this location will serve as the main source |
|
217 | source tree. The documentation in this location will serve as the main source | |
216 | for IPython documentation and all existing documentation should be converted |
|
218 | for IPython documentation and all existing documentation should be converted | |
217 | to this format. |
|
219 | to this format. | |
218 |
|
220 | |||
219 | To build the final documentation, we use Sphinx [Sphinx]_. Once you have |
|
221 | To build the final documentation, we use Sphinx [Sphinx]_. Once you have | |
220 | Sphinx installed, you can build the html docs yourself by doing:: |
|
222 | Sphinx installed, you can build the html docs yourself by doing:: | |
221 |
|
223 | |||
222 | $ cd ipython-mybranch/docs |
|
224 | $ cd ipython-mybranch/docs | |
223 | $ make html |
|
225 | $ make html | |
224 |
|
226 | |||
225 | Docstring format |
|
227 | Docstring format | |
226 | ---------------- |
|
228 | ---------------- | |
227 |
|
229 | |||
228 | Good docstrings are very important. All new code should have docstrings that |
|
230 | Good docstrings are very important. All new code should have docstrings that | |
229 | are formatted using reStructuredText for markup and formatting, since it is |
|
231 | are formatted using reStructuredText for markup and formatting, since it is | |
230 | understood by a wide variety of tools. Details about using reStructuredText |
|
232 | understood by a wide variety of tools. Details about using reStructuredText | |
231 | for docstrings can be found `here |
|
233 | for docstrings can be found `here | |
232 | <http://epydoc.sourceforge.net/manual-othermarkup.html>`_. |
|
234 | <http://epydoc.sourceforge.net/manual-othermarkup.html>`_. | |
233 |
|
235 | |||
234 | Additional PEPs of interest regarding documentation of code: |
|
236 | Additional PEPs of interest regarding documentation of code: | |
235 |
|
237 | |||
236 | * `Docstring Conventions <http://www.python.org/peps/pep-0257.html>`_ |
|
238 | * `Docstring Conventions <http://www.python.org/peps/pep-0257.html>`_ | |
237 | * `Docstring Processing System Framework <http://www.python.org/peps/pep-0256.html>`_ |
|
239 | * `Docstring Processing System Framework <http://www.python.org/peps/pep-0256.html>`_ | |
238 | * `Docutils Design Specification <http://www.python.org/peps/pep-0258.html>`_ |
|
240 | * `Docutils Design Specification <http://www.python.org/peps/pep-0258.html>`_ | |
239 |
|
241 | |||
240 |
|
242 | |||
241 | Coding conventions |
|
243 | Coding conventions | |
242 | ================== |
|
244 | ================== | |
243 |
|
245 | |||
244 | General |
|
246 | General | |
245 | ------- |
|
247 | ------- | |
246 |
|
248 | |||
247 | In general, we'll try to follow the standard Python style conventions as |
|
249 | In general, we'll try to follow the standard Python style conventions as | |
248 | described here: |
|
250 | described here: | |
249 |
|
251 | |||
250 | * `Style Guide for Python Code <http://www.python.org/peps/pep-0008.html>`_ |
|
252 | * `Style Guide for Python Code <http://www.python.org/peps/pep-0008.html>`_ | |
251 |
|
253 | |||
252 |
|
254 | |||
253 | Other comments: |
|
255 | Other comments: | |
254 |
|
256 | |||
255 | * In a large file, top level classes and functions should be |
|
257 | * In a large file, top level classes and functions should be | |
256 | separated by 2-3 lines to make it easier to separate them visually. |
|
258 | separated by 2-3 lines to make it easier to separate them visually. | |
257 | * Use 4 spaces for indentation. |
|
259 | * Use 4 spaces for indentation. | |
258 | * Keep the ordering of methods the same in classes that have the same |
|
260 | * Keep the ordering of methods the same in classes that have the same | |
259 | methods. This is particularly true for classes that implement an interface. |
|
261 | methods. This is particularly true for classes that implement an interface. | |
260 |
|
262 | |||
261 | Naming conventions |
|
263 | Naming conventions | |
262 | ------------------ |
|
264 | ------------------ | |
263 |
|
265 | |||
264 | In terms of naming conventions, we'll follow the guidelines from the `Style |
|
266 | In terms of naming conventions, we'll follow the guidelines from the `Style | |
265 | Guide for Python Code`_. |
|
267 | Guide for Python Code`_. | |
266 |
|
268 | |||
267 | For all new IPython code (and much existing code is being refactored), we'll |
|
269 | For all new IPython code (and much existing code is being refactored), we'll | |
268 | use: |
|
270 | use: | |
269 |
|
271 | |||
270 | * All ``lowercase`` module names. |
|
272 | * All ``lowercase`` module names. | |
271 |
|
273 | |||
272 | * ``CamelCase`` for class names. |
|
274 | * ``CamelCase`` for class names. | |
273 |
|
275 | |||
274 | * ``lowercase_with_underscores`` for methods, functions, variables and |
|
276 | * ``lowercase_with_underscores`` for methods, functions, variables and | |
275 | attributes. |
|
277 | attributes. | |
276 |
|
278 | |||
277 | There are, however, some important exceptions to these rules. In some cases, |
|
279 | There are, however, some important exceptions to these rules. In some cases, | |
278 | IPython code will interface with packages (Twisted, Wx, Qt) that use other |
|
280 | IPython code will interface with packages (Twisted, Wx, Qt) that use other | |
279 | conventions. At some level this makes it impossible to adhere to our own |
|
281 | conventions. At some level this makes it impossible to adhere to our own | |
280 | standards at all times. In particular, when subclassing classes that use other |
|
282 | standards at all times. In particular, when subclassing classes that use other | |
281 | naming conventions, you must follow their naming conventions. To deal with |
|
283 | naming conventions, you must follow their naming conventions. To deal with | |
282 | cases like this, we propose the following policy: |
|
284 | cases like this, we propose the following policy: | |
283 |
|
285 | |||
284 | * If you are subclassing a class that uses different conventions, use its |
|
286 | * If you are subclassing a class that uses different conventions, use its | |
285 | naming conventions throughout your subclass. Thus, if you are creating a |
|
287 | naming conventions throughout your subclass. Thus, if you are creating a | |
286 | Twisted Protocol class, used Twisted's |
|
288 | Twisted Protocol class, used Twisted's | |
287 | ``namingSchemeForMethodsAndAttributes.`` |
|
289 | ``namingSchemeForMethodsAndAttributes.`` | |
288 |
|
290 | |||
289 | * All IPython's official interfaces should use our conventions. In some cases |
|
291 | * All IPython's official interfaces should use our conventions. In some cases | |
290 | this will mean that you need to provide shadow names (first implement |
|
292 | this will mean that you need to provide shadow names (first implement | |
291 | ``fooBar`` and then ``foo_bar = fooBar``). We want to avoid this at all |
|
293 | ``fooBar`` and then ``foo_bar = fooBar``). We want to avoid this at all | |
292 | costs, but it will probably be necessary at times. But, please use this |
|
294 | costs, but it will probably be necessary at times. But, please use this | |
293 | sparingly! |
|
295 | sparingly! | |
294 |
|
296 | |||
295 | Implementation-specific *private* methods will use |
|
297 | Implementation-specific *private* methods will use | |
296 | ``_single_underscore_prefix``. Names with a leading double underscore will |
|
298 | ``_single_underscore_prefix``. Names with a leading double underscore will | |
297 | *only* be used in special cases, as they makes subclassing difficult (such |
|
299 | *only* be used in special cases, as they makes subclassing difficult (such | |
298 | names are not easily seen by child classes). |
|
300 | names are not easily seen by child classes). | |
299 |
|
301 | |||
300 | Occasionally some run-in lowercase names are used, but mostly for very short |
|
302 | Occasionally some run-in lowercase names are used, but mostly for very short | |
301 | names or where we are implementing methods very similar to existing ones in a |
|
303 | names or where we are implementing methods very similar to existing ones in a | |
302 | base class (like ``runlines()`` where ``runsource()`` and ``runcode()`` had |
|
304 | base class (like ``runlines()`` where ``runsource()`` and ``runcode()`` had | |
303 | established precedent). |
|
305 | established precedent). | |
304 |
|
306 | |||
305 | The old IPython codebase has a big mix of classes and modules prefixed with an |
|
307 | The old IPython codebase has a big mix of classes and modules prefixed with an | |
306 | explicit ``IP``. In Python this is mostly unnecessary, redundant and frowned |
|
308 | explicit ``IP``. In Python this is mostly unnecessary, redundant and frowned | |
307 | upon, as namespaces offer cleaner prefixing. The only case where this approach |
|
309 | upon, as namespaces offer cleaner prefixing. The only case where this approach | |
308 | is justified is for classes which are expected to be imported into external |
|
310 | is justified is for classes which are expected to be imported into external | |
309 | namespaces and a very generic name (like Shell) is too likely to clash with |
|
311 | namespaces and a very generic name (like Shell) is too likely to clash with | |
310 | something else. We'll need to revisit this issue as we clean up and refactor |
|
312 | something else. We'll need to revisit this issue as we clean up and refactor | |
311 | the code, but in general we should remove as many unnecessary ``IP``/``ip`` |
|
313 | the code, but in general we should remove as many unnecessary ``IP``/``ip`` | |
312 | prefixes as possible. However, if a prefix seems absolutely necessary the more |
|
314 | prefixes as possible. However, if a prefix seems absolutely necessary the more | |
313 | specific ``IPY`` or ``ipy`` are preferred. |
|
315 | specific ``IPY`` or ``ipy`` are preferred. | |
314 |
|
316 | |||
315 | .. _devel_testing: |
|
317 | .. _devel_testing: | |
316 |
|
318 | |||
317 | Testing system |
|
319 | Testing system | |
318 | ============== |
|
320 | ============== | |
319 |
|
321 | |||
320 | It is extremely important that all code contributed to IPython has tests. |
|
322 | It is extremely important that all code contributed to IPython has tests. | |
321 | Tests should be written as unittests, doctests or as entities that the Nose |
|
323 | Tests should be written as unittests, doctests or as entities that the Nose | |
322 | [Nose]_ testing package will find. Regardless of how the tests are written, we |
|
324 | [Nose]_ testing package will find. Regardless of how the tests are written, we | |
323 | will use Nose for discovering and running the tests. Nose will be required to |
|
325 | will use Nose for discovering and running the tests. Nose will be required to | |
324 | run the IPython test suite, but will not be required to simply use IPython. |
|
326 | run the IPython test suite, but will not be required to simply use IPython. | |
325 |
|
327 | |||
326 | Tests of Twisted using code need to follow two additional guidelines: |
|
328 | Tests of Twisted using code need to follow two additional guidelines: | |
327 |
|
329 | |||
328 | 1. Twisted using tests should be written by subclassing the :class:`TestCase` |
|
330 | 1. Twisted using tests should be written by subclassing the :class:`TestCase` | |
329 | class that comes with :mod:`twisted.trial.unittest`. |
|
331 | class that comes with :mod:`twisted.trial.unittest`. | |
330 |
|
332 | |||
331 | 2. All :class:`Deferred` instances that are created in the test must be |
|
333 | 2. All :class:`Deferred` instances that are created in the test must be | |
332 | properly chained and the final one *must* be the return value of the test |
|
334 | properly chained and the final one *must* be the return value of the test | |
333 | method. |
|
335 | method. | |
334 |
|
336 | |||
335 | When these two things are done, Nose will be able to run the tests and the |
|
337 | When these two things are done, Nose will be able to run the tests and the | |
336 | twisted reactor will be handled correctly. |
|
338 | twisted reactor will be handled correctly. | |
337 |
|
339 | |||
338 | Each subpackage in IPython should have its own :file:`tests` directory that |
|
340 | Each subpackage in IPython should have its own :file:`tests` directory that | |
339 | contains all of the tests for that subpackage. This allows each subpackage to |
|
341 | contains all of the tests for that subpackage. This allows each subpackage to | |
340 | be self-contained. A good convention to follow is to have a file named |
|
342 | be self-contained. A good convention to follow is to have a file named | |
341 | :file:`test_foo.py` for each module :file:`foo.py` in the package. This makes |
|
343 | :file:`test_foo.py` for each module :file:`foo.py` in the package. This makes | |
342 | it easy to organize the tests, though like most conventions, it's OK to break |
|
344 | it easy to organize the tests, though like most conventions, it's OK to break | |
343 | it if logic and common sense dictate otherwise. |
|
345 | it if logic and common sense dictate otherwise. | |
344 |
|
346 | |||
345 | If a subpackage has any dependencies beyond the Python standard library, the |
|
347 | If a subpackage has any dependencies beyond the Python standard library, the | |
346 | tests for that subpackage should be skipped if the dependencies are not |
|
348 | tests for that subpackage should be skipped if the dependencies are not | |
347 | found. This is very important so users don't get tests failing simply because |
|
349 | found. This is very important so users don't get tests failing simply because | |
348 | they don't have dependencies. We ship a set of decorators in the |
|
350 | they don't have dependencies. We ship a set of decorators in the | |
349 | :mod:`IPython.testing` package to tag tests that may be platform-specific or |
|
351 | :mod:`IPython.testing` package to tag tests that may be platform-specific or | |
350 | otherwise may have restrictions; if the existing ones don't fit your needs, add |
|
352 | otherwise may have restrictions; if the existing ones don't fit your needs, add | |
351 | a new decorator in that location so other tests can reuse it. |
|
353 | a new decorator in that location so other tests can reuse it. | |
352 |
|
354 | |||
353 | To run the IPython test suite, use the :command:`iptest` command that is |
|
355 | To run the IPython test suite, use the :command:`iptest` command that is | |
354 | installed with IPython (if you are using IPython in-place, without installing |
|
356 | installed with IPython (if you are using IPython in-place, without installing | |
355 | it, you can find this script in the :file:`scripts` directory):: |
|
357 | it, you can find this script in the :file:`scripts` directory):: | |
356 |
|
358 | |||
357 | $ iptest |
|
359 | $ iptest | |
358 |
|
360 | |||
359 | This command colects all IPython tests into separate groups, and then calls |
|
361 | This command colects all IPython tests into separate groups, and then calls | |
360 | either Nose with the proper options and extensions, or Twisted's |
|
362 | either Nose with the proper options and extensions, or Twisted's | |
361 | :command:`trial`. This ensures that tests that need the Twisted reactor |
|
363 | :command:`trial`. This ensures that tests that need the Twisted reactor | |
362 | management facilities execute separate of Nose. If any individual test group |
|
364 | management facilities execute separate of Nose. If any individual test group | |
363 | fails, :command:`iptest` will print what you need to type so you can rerun that |
|
365 | fails, :command:`iptest` will print what you need to type so you can rerun that | |
364 | particular test group alone for debugging. |
|
366 | particular test group alone for debugging. | |
365 |
|
367 | |||
366 | By default, :command:`iptest` runs the entire IPython test |
|
368 | By default, :command:`iptest` runs the entire IPython test | |
367 | suite (skipping tests that may be platform-specific or which depend on tools |
|
369 | suite (skipping tests that may be platform-specific or which depend on tools | |
368 | you may not have). But you can also use it to run only one specific test file, |
|
370 | you may not have). But you can also use it to run only one specific test file, | |
369 | or a specific test function. For example, this will run only the |
|
371 | or a specific test function. For example, this will run only the | |
370 | :file:`test_magic` file from the test suite:: |
|
372 | :file:`test_magic` file from the test suite:: | |
371 |
|
373 | |||
372 | $ iptest IPython.tests.test_magic |
|
374 | $ iptest IPython.tests.test_magic | |
373 | ---------------------------------------------------------------------- |
|
375 | ---------------------------------------------------------------------- | |
374 | Ran 10 tests in 0.348s |
|
376 | Ran 10 tests in 0.348s | |
375 |
|
377 | |||
376 | OK (SKIP=3) |
|
378 | OK (SKIP=3) | |
377 | Deleting object: second_pass |
|
379 | Deleting object: second_pass | |
378 |
|
380 | |||
379 | while the ``path:function`` syntax allows you to select a specific function in |
|
381 | while the ``path:function`` syntax allows you to select a specific function in | |
380 | that file to run:: |
|
382 | that file to run:: | |
381 |
|
383 | |||
382 | $ iptest IPython.tests.test_magic:test_obj_del |
|
384 | $ iptest IPython.tests.test_magic:test_obj_del | |
383 | ---------------------------------------------------------------------- |
|
385 | ---------------------------------------------------------------------- | |
384 | Ran 1 test in 0.204s |
|
386 | Ran 1 test in 0.204s | |
385 |
|
387 | |||
386 | OK |
|
388 | OK | |
387 |
|
389 | |||
388 | Since :command:`iptest` is based on nosetests, you can pass it any regular |
|
390 | Since :command:`iptest` is based on nosetests, you can pass it any regular | |
389 | nosetests option. For example, you can use ``--pdb`` or ``--pdb-failures`` to |
|
391 | nosetests option. For example, you can use ``--pdb`` or ``--pdb-failures`` to | |
390 | automatically activate the interactive Pdb debugger on errors or failures. See |
|
392 | automatically activate the interactive Pdb debugger on errors or failures. See | |
391 | the nosetests documentation for further details. |
|
393 | the nosetests documentation for further details. | |
392 |
|
394 | |||
393 |
|
395 | |||
394 | A few tips for writing tests |
|
396 | A few tips for writing tests | |
395 | ---------------------------- |
|
397 | ---------------------------- | |
396 |
|
398 | |||
397 | You can write tests either as normal test files, using all the conventions that |
|
399 | You can write tests either as normal test files, using all the conventions that | |
398 | Nose recognizes, or as doctests. Note that *all* IPython functions should have |
|
400 | Nose recognizes, or as doctests. Note that *all* IPython functions should have | |
399 | at least one example that serves as a doctest, whenever technically feasible. |
|
401 | at least one example that serves as a doctest, whenever technically feasible. | |
400 | However, example doctests should only be in the main docstring if they are *a |
|
402 | However, example doctests should only be in the main docstring if they are *a | |
401 | good example*, i.e. if they convey useful information about the function. If |
|
403 | good example*, i.e. if they convey useful information about the function. If | |
402 | you simply would like to write a test as a doctest, put it in a separate test |
|
404 | you simply would like to write a test as a doctest, put it in a separate test | |
403 | file and write a no-op function whose only purpose is its docstring. |
|
405 | file and write a no-op function whose only purpose is its docstring. | |
404 |
|
406 | |||
405 | Note, however, that in a file named :file:`test_X`, functions whose only test |
|
407 | Note, however, that in a file named :file:`test_X`, functions whose only test | |
406 | is their docstring (as a doctest) and which have no test functionality of their |
|
408 | is their docstring (as a doctest) and which have no test functionality of their | |
407 | own, should be called *doctest_foo* instead of *test_foo*, otherwise they get |
|
409 | own, should be called *doctest_foo* instead of *test_foo*, otherwise they get | |
408 | double-counted (the empty function call is counted as a test, which just |
|
410 | double-counted (the empty function call is counted as a test, which just | |
409 | inflates tests numbers artificially). This restriction does not apply to |
|
411 | inflates tests numbers artificially). This restriction does not apply to | |
410 | functions in files with other names, due to how Nose discovers tests. |
|
412 | functions in files with other names, due to how Nose discovers tests. | |
411 |
|
413 | |||
412 | You can use IPython examples in your docstrings. Those can make full use of |
|
414 | You can use IPython examples in your docstrings. Those can make full use of | |
413 | IPython functionality (magics, variable substitution, etc), but be careful to |
|
415 | IPython functionality (magics, variable substitution, etc), but be careful to | |
414 | keep them generic enough that they run identically on all Operating Systems. |
|
416 | keep them generic enough that they run identically on all Operating Systems. | |
415 |
|
417 | |||
416 | The prompts in your doctests can be either of the plain Python ``>>>`` variety |
|
418 | The prompts in your doctests can be either of the plain Python ``>>>`` variety | |
417 | or ``In [1]:`` IPython style. Since this is the IPython system, after all, we |
|
419 | or ``In [1]:`` IPython style. Since this is the IPython system, after all, we | |
418 | encourage you to use IPython prompts throughout, unless you are illustrating a |
|
420 | encourage you to use IPython prompts throughout, unless you are illustrating a | |
419 | specific aspect of the normal prompts (such as the ``%doctest_mode`` magic). |
|
421 | specific aspect of the normal prompts (such as the ``%doctest_mode`` magic). | |
420 |
|
422 | |||
421 | If a test isn't safe to run inside the main nose process (e.g. because it loads |
|
423 | If a test isn't safe to run inside the main nose process (e.g. because it loads | |
422 | a GUI toolkit), consider running it in a subprocess and capturing its output |
|
424 | a GUI toolkit), consider running it in a subprocess and capturing its output | |
423 | for evaluation and test decision later. Here is an example of how to do it, by |
|
425 | for evaluation and test decision later. Here is an example of how to do it, by | |
424 | relying on the builtin ``_ip`` object that contains the public IPython api as |
|
426 | relying on the builtin ``_ip`` object that contains the public IPython api as | |
425 | defined in :mod:`IPython.ipapi`:: |
|
427 | defined in :mod:`IPython.ipapi`:: | |
426 |
|
428 | |||
427 | def test_obj_del(): |
|
429 | def test_obj_del(): | |
428 | """Test that object's __del__ methods are called on exit.""" |
|
430 | """Test that object's __del__ methods are called on exit.""" | |
429 | test_dir = os.path.dirname(__file__) |
|
431 | test_dir = os.path.dirname(__file__) | |
430 | del_file = os.path.join(test_dir,'obj_del.py') |
|
432 | del_file = os.path.join(test_dir,'obj_del.py') | |
431 | out = _ip.IP.getoutput('ipython %s' % del_file) |
|
433 | out = _ip.IP.getoutput('ipython %s' % del_file) | |
432 | nt.assert_equals(out,'object A deleted') |
|
434 | nt.assert_equals(out,'object A deleted') | |
433 |
|
435 | |||
434 |
|
436 | |||
435 |
|
437 | |||
436 | If a doctest contains input whose output you don't want to verify identically |
|
438 | If a doctest contains input whose output you don't want to verify identically | |
437 | via doctest (random output, an object id, etc), you can mark a docstring with |
|
439 | via doctest (random output, an object id, etc), you can mark a docstring with | |
438 | ``#random``. All of these test will have their code executed but no output |
|
440 | ``#random``. All of these test will have their code executed but no output | |
439 | checking will be done:: |
|
441 | checking will be done:: | |
440 |
|
442 | |||
441 | >>> 1+3 |
|
443 | >>> 1+3 | |
442 | junk goes here... # random |
|
444 | junk goes here... # random | |
443 |
|
445 | |||
444 | >>> 1+2 |
|
446 | >>> 1+2 | |
445 | again, anything goes #random |
|
447 | again, anything goes #random | |
446 | if multiline, the random mark is only needed once. |
|
448 | if multiline, the random mark is only needed once. | |
447 |
|
449 | |||
448 | >>> 1+2 |
|
450 | >>> 1+2 | |
449 | You can also put the random marker at the end: |
|
451 | You can also put the random marker at the end: | |
450 | # random |
|
452 | # random | |
451 |
|
453 | |||
452 | >>> 1+2 |
|
454 | >>> 1+2 | |
453 | # random |
|
455 | # random | |
454 | .. or at the beginning. |
|
456 | .. or at the beginning. | |
455 |
|
457 | |||
456 | In a case where you want an *entire* docstring to be executed but not verified |
|
458 | In a case where you want an *entire* docstring to be executed but not verified | |
457 | (this only serves to check that the code runs without crashing, so it should be |
|
459 | (this only serves to check that the code runs without crashing, so it should be | |
458 | used very sparingly), you can put ``# all-random`` in the docstring. |
|
460 | used very sparingly), you can put ``# all-random`` in the docstring. | |
459 |
|
461 | |||
460 | .. _devel_config: |
|
462 | .. _devel_config: | |
461 |
|
463 | |||
462 | Release checklist |
|
464 | Release checklist | |
463 | ================= |
|
465 | ================= | |
464 |
|
466 | |||
465 | Most of the release process is automated by the :file:`release` script in the |
|
467 | Most of the release process is automated by the :file:`release` script in the | |
466 | :file:`tools` directory. This is just a handy reminder for the release manager. |
|
468 | :file:`tools` directory. This is just a handy reminder for the release manager. | |
467 |
|
469 | |||
468 | #. First, run :file:`build_release`, which does all the file checking and |
|
470 | #. First, run :file:`build_release`, which does all the file checking and | |
469 | building that the real release script will do. This will let you do test |
|
471 | building that the real release script will do. This will let you do test | |
470 | installations, check that the build procedure runs OK, etc. You may want to |
|
472 | installations, check that the build procedure runs OK, etc. You may want to | |
471 | disable a few things like multi-version RPM building while testing, because |
|
473 | disable a few things like multi-version RPM building while testing, because | |
472 | otherwise the build takes really long. |
|
474 | otherwise the build takes really long. | |
473 |
|
475 | |||
474 | #. Run the release script, which makes the tar.gz, eggs and Win32 .exe |
|
476 | #. Run the release script, which makes the tar.gz, eggs and Win32 .exe | |
475 | installer. It posts them to the site and registers the release with PyPI. |
|
477 | installer. It posts them to the site and registers the release with PyPI. | |
476 |
|
478 | |||
477 | #. Updating the website with announcements and links to the updated |
|
479 | #. Updating the website with announcements and links to the updated | |
478 | changes.txt in html form. Remember to put a short note both on the news |
|
480 | changes.txt in html form. Remember to put a short note both on the news | |
479 | page of the site and on Launcphad. |
|
481 | page of the site and on Launcphad. | |
480 |
|
482 | |||
481 | #. Drafting a short release announcement with i) highlights and ii) a link to |
|
483 | #. Drafting a short release announcement with i) highlights and ii) a link to | |
482 | the html changes.txt. |
|
484 | the html changes.txt. | |
483 |
|
485 | |||
484 | #. Make sure that the released version of the docs is live on the site. |
|
486 | #. Make sure that the released version of the docs is live on the site. | |
485 |
|
487 | |||
486 | #. Celebrate! |
|
488 | #. Celebrate! | |
487 |
|
489 | |||
488 | Porting to 3.0 |
|
490 | Porting to 3.0 | |
489 | ============== |
|
491 | ============== | |
490 |
|
492 | |||
491 | There are no definite plans for porting of IPython to python 3. The major |
|
493 | There are no definite plans for porting of IPython to python 3. The major | |
492 | issue is the dependency on twisted framework for the networking/threading |
|
494 | issue is the dependency on twisted framework for the networking/threading | |
493 | stuff. It is possible that it the traditional IPython interactive console |
|
495 | stuff. It is possible that it the traditional IPython interactive console | |
494 | could be ported more easily since it has no such dependency. Here are a few |
|
496 | could be ported more easily since it has no such dependency. Here are a few | |
495 | things that will need to be considered when doing such a port especially |
|
497 | things that will need to be considered when doing such a port especially | |
496 | if we want to have a codebase that works directly on both 2.x and 3.x. |
|
498 | if we want to have a codebase that works directly on both 2.x and 3.x. | |
497 |
|
499 | |||
498 | 1. The syntax for exceptions changed (PEP 3110). The old |
|
500 | 1. The syntax for exceptions changed (PEP 3110). The old | |
499 | `except exc, var` changed to `except exc as var`. At last |
|
501 | `except exc, var` changed to `except exc as var`. At last | |
500 | count there was 78 occurences of this usage in the codebase. This |
|
502 | count there was 78 occurences of this usage in the codebase. This | |
501 | is a particularly problematic issue, because it's not easy to |
|
503 | is a particularly problematic issue, because it's not easy to | |
502 | implement it in a 2.5-compatible way. |
|
504 | implement it in a 2.5-compatible way. | |
503 |
|
505 | |||
504 | Because it is quite difficult to support simultaneously Python 2.5 and 3.x, we |
|
506 | Because it is quite difficult to support simultaneously Python 2.5 and 3.x, we | |
505 | will likely at some point put out a release that requires strictly 2.6 and |
|
507 | will likely at some point put out a release that requires strictly 2.6 and | |
506 | abandons 2.5 compatibility. This will then allow us to port the code to using |
|
508 | abandons 2.5 compatibility. This will then allow us to port the code to using | |
507 | :func:`print` as a function, `except exc as var` syntax, etc. But as of |
|
509 | :func:`print` as a function, `except exc as var` syntax, etc. But as of | |
508 | version 0.11 at least, we will retain Python 2.5 compatibility. |
|
510 | version 0.11 at least, we will retain Python 2.5 compatibility. | |
509 |
|
511 | |||
510 |
|
512 | |||
511 | .. [Bazaar] Bazaar. http://bazaar-vcs.org/ |
|
513 | .. [Bazaar] Bazaar. http://bazaar-vcs.org/ | |
512 | .. [Launchpad] Launchpad. http://www.launchpad.net/ipython |
|
514 | .. [Launchpad] Launchpad. http://www.launchpad.net/ipython | |
513 | .. [reStructuredText] reStructuredText. http://docutils.sourceforge.net/rst.html |
|
515 | .. [reStructuredText] reStructuredText. http://docutils.sourceforge.net/rst.html | |
514 | .. [Sphinx] Sphinx. http://sphinx.pocoo.org/ |
|
516 | .. [Sphinx] Sphinx. http://sphinx.pocoo.org/ | |
515 | .. [Nose] Nose: a discovery based unittest extension. http://code.google.com/p/python-nose/ |
|
517 | .. [Nose] Nose: a discovery based unittest extension. http://code.google.com/p/python-nose/ |
@@ -1,94 +1,84 b'' | |||||
1 | ============================= |
|
1 | ============================= | |
2 | IPython module reorganization |
|
2 | IPython module reorganization | |
3 | ============================= |
|
3 | ============================= | |
4 |
|
4 | |||
5 | Currently, IPython has many top-level modules that serve many different |
|
5 | Currently, IPython has many top-level modules that serve many different | |
6 | purposes. The lack of organization make it very difficult for developers to |
|
6 | purposes. The lack of organization make it very difficult for developers to | |
7 | work on IPython and understand its design. This document contains notes about |
|
7 | work on IPython and understand its design. This document contains notes about | |
8 | how we will reorganize the modules into sub-packages. |
|
8 | how we will reorganize the modules into sub-packages. | |
9 |
|
9 | |||
10 | .. warning:: |
|
10 | .. warning:: | |
11 |
|
11 | |||
12 | This effort will possibly break third party packages that use IPython as |
|
12 | This effort will possibly break third party packages that use IPython as | |
13 | a library or hack on the IPython internals. |
|
13 | a library or hack on the IPython internals. | |
14 |
|
14 | |||
15 | .. warning:: |
|
15 | .. warning:: | |
16 |
|
16 | |||
17 | This effort will result in the removal from IPython of certain modules |
|
17 | This effort will result in the removal from IPython of certain modules | |
18 | that are not used anymore, don't currently work, are unmaintained, etc. |
|
18 | that are not used anymore, don't currently work, are unmaintained, etc. | |
19 |
|
19 | |||
20 |
|
20 | |||
21 | Current subpackges |
|
21 | Current subpackges | |
22 | ================== |
|
22 | ================== | |
23 |
|
23 | |||
24 | IPython currently has the following sub-packages: |
|
24 | IPython currently has the following sub-packages: | |
25 |
|
25 | |||
26 | * :mod:`IPython.config` |
|
26 | * :mod:`IPython.config` | |
27 |
|
27 | |||
28 | * :mod:`IPython.Extensions` |
|
28 | * :mod:`IPython.Extensions` | |
29 |
|
29 | |||
30 | * :mod:`IPython.external` |
|
30 | * :mod:`IPython.external` | |
31 |
|
31 | |||
32 | * :mod:`IPython.frontend` |
|
32 | * :mod:`IPython.frontend` | |
33 |
|
33 | |||
34 | * :mod:`IPython.gui` |
|
34 | * :mod:`IPython.gui` | |
35 |
|
35 | |||
36 | * :mod:`IPython.kernel` |
|
36 | * :mod:`IPython.kernel` | |
37 |
|
37 | |||
38 | * :mod:`IPython.testing` |
|
38 | * :mod:`IPython.testing` | |
39 |
|
39 | |||
40 | * :mod:`IPython.tests` |
|
40 | * :mod:`IPython.tests` | |
41 |
|
41 | |||
42 | * :mod:`IPython.tools` |
|
42 | * :mod:`IPython.tools` | |
43 |
|
43 | |||
44 | * :mod:`IPython.UserConfig` |
|
44 | * :mod:`IPython.UserConfig` | |
45 |
|
45 | |||
46 | New Subpackages to be created |
|
46 | New Subpackages to be created | |
47 | ============================= |
|
47 | ============================= | |
48 |
|
48 | |||
49 | We propose to create the following new sub-packages: |
|
49 | We propose to create the following new sub-packages: | |
50 |
|
50 | |||
51 | * :mod:`IPython.core`. This sub-package will contain the core of the IPython |
|
51 | * :mod:`IPython.core`. This sub-package will contain the core of the IPython | |
52 | interpreter, but none of its extended capabilities. |
|
52 | interpreter, but none of its extended capabilities. | |
53 |
|
53 | |||
54 | * :mod:`IPython.lib`. IPython has many extended capabilities that are not part |
|
54 | * :mod:`IPython.lib`. IPython has many extended capabilities that are not part | |
55 | of the IPython core. These things will go here. |
|
55 | of the IPython core. These things will go here. | |
56 |
|
56 | |||
57 | * :mod:`IPython.utils`. This sub-package will contain anything that might |
|
57 | * :mod:`IPython.utils`. This sub-package will contain anything that might | |
58 | eventually be found in the Python standard library, like things in |
|
58 | eventually be found in the Python standard library, like things in | |
59 | :mod:`genutils`. Each sub-module in this sub-package should contain |
|
59 | :mod:`genutils`. Each sub-module in this sub-package should contain | |
60 | functions and classes that serve a single purpose. |
|
60 | functions and classes that serve a single purpose. | |
61 |
|
61 | |||
62 | * :mod:`IPython.deathrow`. This is for code that is untested and/or rotting |
|
62 | * :mod:`IPython.deathrow`. This is for code that is untested and/or rotting | |
63 | and needs to be removed from IPython. Eventually all this code will either |
|
63 | and needs to be removed from IPython. Eventually all this code will either | |
64 | i) be revived by someone willing to maintain it with tests and docs and |
|
64 | i) be revived by someone willing to maintain it with tests and docs and | |
65 | re-included into IPython or 2) be removed from IPython proper, but put into |
|
65 | re-included into IPython or 2) be removed from IPython proper, but put into | |
66 | a separate top-level (not IPython) package that we keep around. No new code |
|
66 | a separate top-level (not IPython) package that we keep around. No new code | |
67 | will be allowed here. |
|
67 | will be allowed here. | |
68 |
|
68 | |||
69 | * :mod:`IPython.quarantine`. This is for code that doesn't meet IPython's |
|
69 | * :mod:`IPython.quarantine`. This is for code that doesn't meet IPython's | |
70 | standards, but that we plan on keeping. To be moved out of this sub-package |
|
70 | standards, but that we plan on keeping. To be moved out of this sub-package | |
71 | a module needs to have a maintainer, tests and documentation. |
|
71 | a module needs to have a maintainer, tests and documentation. | |
72 |
|
72 | |||
73 | Procedure |
|
73 | Procedure | |
74 | ========= |
|
74 | ========= | |
75 |
|
75 | |||
76 | 1. Move the file to its new location with its new name. |
|
76 | 1. Move the file to its new location with its new name. | |
77 | 2. Rename all import statements to reflect the change. |
|
77 | 2. Rename all import statements to reflect the change. | |
78 | 3. Run PyFlakes on each changes module. |
|
78 | 3. Run PyFlakes on each changes module. | |
79 |
|
|
79 | 4. Add tests/test_imports.py to test it. | |
80 |
|
80 | |||
81 | Status |
|
81 | Status | |
82 | ====== |
|
82 | ====== | |
83 |
|
83 | |||
84 | The new subpackages have been created and the top-level modules have been |
|
84 | This branch was merged into trunk in early August of 2009. | |
85 | moved and renamed. Import tests have been created for all of the moved and |
|
|||
86 | renamed modules. The build infrastructure (setup.py and friends) have been |
|
|||
87 | updated and tested on Mac and Windows. Finally, a compatibility layer has been |
|
|||
88 | added for iplib, ipapi and Shell. The follow things still need to be done:: |
|
|||
89 |
|
||||
90 | * I need to modify iptests to properly skip modules that are no longer top |
|
|||
91 | level modules. |
|
|||
92 |
|
||||
93 | * When running python setup.py sdist, the Sphinx API docs fail to build |
|
|||
94 | because of something going on with IPython.core.fakemodule |
|
@@ -1,81 +1,73 b'' | |||||
1 | .. _roadmap: |
|
1 | .. _roadmap: | |
2 |
|
2 | |||
3 | =================== |
|
3 | =================== | |
4 | Development roadmap |
|
4 | Development roadmap | |
5 | =================== |
|
5 | =================== | |
6 |
|
6 | |||
7 | IPython is an ambitious project that is still under heavy development. However, we want IPython to become useful to as many people as possible, as quickly as possible. To help us accomplish this, we are laying out a roadmap of where we are headed and what needs to happen to get there. Hopefully, this will help the IPython developers figure out the best things to work on for each upcoming release. |
|
7 | IPython is an ambitious project that is still under heavy development. | |
|
8 | However, we want IPython to become useful to as many people as possible, as | |||
|
9 | quickly as possible. To help us accomplish this, we are laying out a roadmap | |||
|
10 | of where we are headed and what needs to happen to get there. Hopefully, this | |||
|
11 | will help the IPython developers figure out the best things to work on for | |||
|
12 | each upcoming release. | |||
8 |
|
13 | |||
9 | Work targeted to particular releases |
|
14 | Work targeted to particular releases | |
10 | ==================================== |
|
15 | ==================================== | |
11 |
|
16 | |||
12 | Release 0.10 |
|
|||
13 | ------------ |
|
|||
14 |
|
||||
15 | * Initial refactor of :command:`ipcluster`. |
|
|||
16 |
|
||||
17 | * Better TextMate integration. |
|
|||
18 |
|
||||
19 | * Merge in the daemon branch. |
|
|||
20 |
|
||||
21 | Release 0.11 |
|
17 | Release 0.11 | |
22 | ------------ |
|
18 | ------------ | |
23 |
|
19 | |||
24 | * Refactor the configuration system and command line options for |
|
20 | * [DONE] Full module and package reorganization. | |
25 | :command:`ipengine` and :command:`ipcontroller`. This will include the |
|
21 | ||
26 | creation of cluster directories that encapsulate all the configuration |
|
22 | * [DONE] Removal of the threaded shells and new implementation of GUI support | |
27 | files, log files and security related files for a particular cluster. |
|
23 | based on ``PyOSInputHook``. | |
28 |
|
24 | |||
29 |
* Refactor |
|
25 | * Refactor the configuration system. | |
30 |
|
26 | |||
31 | * Refactor the daemon stuff to support the new configuration system. |
|
27 | * Prepare to refactor IPython's core by creating a new component and | |
|
28 | application system. | |||
32 |
|
29 | |||
33 | * Merge back in the core of the notebook. |
|
|||
34 |
|
30 | |||
35 | Release 0.12 |
|
31 | Release 0.12 | |
36 | ------------ |
|
32 | ------------ | |
37 |
|
33 | |||
38 | * Fully integrate process startup with the daemons for full process |
|
|||
39 | management. |
|
|||
40 |
|
34 | |||
41 | * Make the capabilites of :command:`ipcluster` available from simple Python |
|
|||
42 | classes. |
|
|||
43 |
|
35 | |||
44 | Major areas of work |
|
36 | Major areas of work | |
45 | =================== |
|
37 | =================== | |
46 |
|
38 | |||
47 | Refactoring the main IPython core |
|
39 | Refactoring the main IPython core | |
48 | --------------------------------- |
|
40 | --------------------------------- | |
49 |
|
41 | |||
50 | Process management for :mod:`IPython.kernel` |
|
42 | Process management for :mod:`IPython.kernel` | |
51 | -------------------------------------------- |
|
43 | -------------------------------------------- | |
52 |
|
44 | |||
53 | Configuration system |
|
45 | Configuration system | |
54 | -------------------- |
|
46 | -------------------- | |
55 |
|
47 | |||
56 | Performance problems |
|
48 | Performance problems | |
57 | -------------------- |
|
49 | -------------------- | |
58 |
|
50 | |||
59 | Currently, we have a number of performance issues that are waiting to bite users: |
|
51 | Currently, we have a number of performance issues that are waiting to bite users: | |
60 |
|
52 | |||
61 | * The controller stores a large amount of state in Python dictionaries. Under |
|
53 | * The controller stores a large amount of state in Python dictionaries. Under | |
62 | heavy usage, these dicts with get very large, causing memory usage problems. |
|
54 | heavy usage, these dicts with get very large, causing memory usage problems. | |
63 | We need to develop more scalable solutions to this problem, such as using a |
|
55 | We need to develop more scalable solutions to this problem, such as using a | |
64 | sqlite database to store this state. This will also help the controller to |
|
56 | sqlite database to store this state. This will also help the controller to | |
65 | be more fault tolerant. |
|
57 | be more fault tolerant. | |
66 |
|
58 | |||
67 | * We currently don't have a good way of handling large objects in the |
|
59 | * We currently don't have a good way of handling large objects in the | |
68 | controller. The biggest problem is that because we don't have any way of |
|
60 | controller. The biggest problem is that because we don't have any way of | |
69 | streaming objects, we get lots of temporary copies in the low-level buffers. |
|
61 | streaming objects, we get lots of temporary copies in the low-level buffers. | |
70 | We need to implement a better serialization approach and true streaming |
|
62 | We need to implement a better serialization approach and true streaming | |
71 | support. |
|
63 | support. | |
72 |
|
64 | |||
73 | * The controller currently unpickles and repickles objects. We need to use the |
|
65 | * The controller currently unpickles and repickles objects. We need to use the | |
74 | [push|pull]_serialized methods instead. |
|
66 | [push|pull]_serialized methods instead. | |
75 |
|
67 | |||
76 | * Currently the controller is a bottleneck. The best approach for this is to |
|
68 | * Currently the controller is a bottleneck. The best approach for this is to | |
77 | separate the controller itself into multiple processes, one for the core |
|
69 | separate the controller itself into multiple processes, one for the core | |
78 | controller and one each for the controller interfaces. |
|
70 | controller and one each for the controller interfaces. | |
79 |
|
71 | |||
80 |
|
72 | |||
81 |
|
73 |
@@ -1,1632 +1,1539 b'' | |||||
1 | ================= |
|
1 | ================= | |
2 | IPython reference |
|
2 | IPython reference | |
3 | ================= |
|
3 | ================= | |
4 |
|
4 | |||
5 | .. _command_line_options: |
|
5 | .. _command_line_options: | |
6 |
|
6 | |||
7 | Command-line usage |
|
7 | Command-line usage | |
8 | ================== |
|
8 | ================== | |
9 |
|
9 | |||
10 | You start IPython with the command:: |
|
10 | You start IPython with the command:: | |
11 |
|
11 | |||
12 | $ ipython [options] files |
|
12 | $ ipython [options] files | |
13 |
|
13 | |||
14 | If invoked with no options, it executes all the files listed in sequence |
|
14 | If invoked with no options, it executes all the files listed in sequence | |
15 | and drops you into the interpreter while still acknowledging any options |
|
15 | and drops you into the interpreter while still acknowledging any options | |
16 | you may have set in your ipythonrc file. This behavior is different from |
|
16 | you may have set in your ipythonrc file. This behavior is different from | |
17 | standard Python, which when called as python -i will only execute one |
|
17 | standard Python, which when called as python -i will only execute one | |
18 | file and ignore your configuration setup. |
|
18 | file and ignore your configuration setup. | |
19 |
|
19 | |||
20 | Please note that some of the configuration options are not available at |
|
20 | Please note that some of the configuration options are not available at | |
21 | the command line, simply because they are not practical here. Look into |
|
21 | the command line, simply because they are not practical here. Look into | |
22 | your ipythonrc configuration file for details on those. This file |
|
22 | your ipythonrc configuration file for details on those. This file | |
23 | typically installed in the $HOME/.ipython directory. For Windows users, |
|
23 | typically installed in the $HOME/.ipython directory. For Windows users, | |
24 | $HOME resolves to C:\\Documents and Settings\\YourUserName in most |
|
24 | $HOME resolves to C:\\Documents and Settings\\YourUserName in most | |
25 | instances. In the rest of this text, we will refer to this directory as |
|
25 | instances. In the rest of this text, we will refer to this directory as | |
26 | IPYTHONDIR. |
|
26 | IPYTHONDIR. | |
27 |
|
27 | |||
28 | .. _Threading options: |
|
|||
29 |
|
28 | |||
30 |
|
29 | |||
31 | Special Threading Options |
|
30 | Special Threading Options | |
32 | ------------------------- |
|
31 | ------------------------- | |
33 |
|
32 | |||
34 | The following special options are ONLY valid at the beginning of the |
|
33 | Previously IPython had command line options for controlling GUI event loop | |
35 | command line, and not later. This is because they control the initial- |
|
34 | integration (-gthread, -qthread, -q4thread, -wthread, -pylab). As of IPython | |
36 | ization of ipython itself, before the normal option-handling mechanism |
|
35 | version 0.11, these have been deprecated. Please see the new ``%gui`` | |
37 | is active. |
|
36 | magic command or :ref:`this section <gui_support>` for details on the new | |
38 |
|
37 | interface. | ||
39 | -gthread, -qthread, -q4thread, -wthread, -pylab: |
|
|||
40 | Only one of these can be given, and it can only be given as |
|
|||
41 | the first option passed to IPython (it will have no effect in |
|
|||
42 | any other position). They provide threading support for the |
|
|||
43 | GTK, Qt (versions 3 and 4) and WXPython toolkits, and for the |
|
|||
44 | matplotlib library. |
|
|||
45 |
|
||||
46 | With any of the first four options, IPython starts running a |
|
|||
47 | separate thread for the graphical toolkit's operation, so that |
|
|||
48 | you can open and control graphical elements from within an |
|
|||
49 | IPython command line, without blocking. All four provide |
|
|||
50 | essentially the same functionality, respectively for GTK, Qt3, |
|
|||
51 | Qt4 and WXWidgets (via their Python interfaces). |
|
|||
52 |
|
||||
53 | Note that with -wthread, you can additionally use the |
|
|||
54 | -wxversion option to request a specific version of wx to be |
|
|||
55 | used. This requires that you have the wxversion Python module |
|
|||
56 | installed, which is part of recent wxPython distributions. |
|
|||
57 |
|
||||
58 | If -pylab is given, IPython loads special support for the mat |
|
|||
59 | plotlib library (http://matplotlib.sourceforge.net), allowing |
|
|||
60 | interactive usage of any of its backends as defined in the |
|
|||
61 | user's ~/.matplotlib/matplotlibrc file. It automatically |
|
|||
62 | activates GTK, Qt or WX threading for IPyhton if the choice of |
|
|||
63 | matplotlib backend requires it. It also modifies the %run |
|
|||
64 | command to correctly execute (without blocking) any |
|
|||
65 | matplotlib-based script which calls show() at the end. |
|
|||
66 |
|
||||
67 | -tk |
|
|||
68 | The -g/q/q4/wthread options, and -pylab (if matplotlib is |
|
|||
69 | configured to use GTK, Qt3, Qt4 or WX), will normally block Tk |
|
|||
70 | graphical interfaces. This means that when either GTK, Qt or WX |
|
|||
71 | threading is active, any attempt to open a Tk GUI will result in a |
|
|||
72 | dead window, and possibly cause the Python interpreter to crash. |
|
|||
73 | An extra option, -tk, is available to address this issue. It can |
|
|||
74 | only be given as a second option after any of the above (-gthread, |
|
|||
75 | -wthread or -pylab). |
|
|||
76 |
|
||||
77 | If -tk is given, IPython will try to coordinate Tk threading |
|
|||
78 | with GTK, Qt or WX. This is however potentially unreliable, and |
|
|||
79 | you will have to test on your platform and Python configuration to |
|
|||
80 | determine whether it works for you. Debian users have reported |
|
|||
81 | success, apparently due to the fact that Debian builds all of Tcl, |
|
|||
82 | Tk, Tkinter and Python with pthreads support. Under other Linux |
|
|||
83 | environments (such as Fedora Core 2/3), this option has caused |
|
|||
84 | random crashes and lockups of the Python interpreter. Under other |
|
|||
85 | operating systems (Mac OSX and Windows), you'll need to try it to |
|
|||
86 | find out, since currently no user reports are available. |
|
|||
87 |
|
||||
88 | There is unfortunately no way for IPython to determine at run time |
|
|||
89 | whether -tk will work reliably or not, so you will need to do some |
|
|||
90 | experiments before relying on it for regular work. |
|
|||
91 |
|
||||
92 |
|
||||
93 |
|
38 | |||
94 | Regular Options |
|
39 | Regular Options | |
95 | --------------- |
|
40 | --------------- | |
96 |
|
41 | |||
97 | After the above threading options have been given, regular options can |
|
42 | After the above threading options have been given, regular options can | |
98 | follow in any order. All options can be abbreviated to their shortest |
|
43 | follow in any order. All options can be abbreviated to their shortest | |
99 | non-ambiguous form and are case-sensitive. One or two dashes can be |
|
44 | non-ambiguous form and are case-sensitive. One or two dashes can be | |
100 | used. Some options have an alternate short form, indicated after a ``|``. |
|
45 | used. Some options have an alternate short form, indicated after a ``|``. | |
101 |
|
46 | |||
102 | Most options can also be set from your ipythonrc configuration file. See |
|
47 | Most options can also be set from your ipythonrc configuration file. See | |
103 | the provided example for more details on what the options do. Options |
|
48 | the provided example for more details on what the options do. Options | |
104 | given at the command line override the values set in the ipythonrc file. |
|
49 | given at the command line override the values set in the ipythonrc file. | |
105 |
|
50 | |||
106 | All options with a [no] prepended can be specified in negated form |
|
51 | All options with a [no] prepended can be specified in negated form | |
107 | (-nooption instead of -option) to turn the feature off. |
|
52 | (-nooption instead of -option) to turn the feature off. | |
108 |
|
53 | |||
109 | -help print a help message and exit. |
|
54 | -help print a help message and exit. | |
110 |
|
55 | |||
111 | -pylab |
|
56 | -pylab | |
112 | this can only be given as the first option passed to IPython |
|
57 | Deprecated. See :ref:`Matplotlib support <matplotlib_support>` | |
113 | (it will have no effect in any other position). It adds |
|
58 | for more details. | |
114 | special support for the matplotlib library |
|
|||
115 | (http://matplotlib.sourceforge.ne), allowing interactive usage |
|
|||
116 | of any of its backends as defined in the user's .matplotlibrc |
|
|||
117 | file. It automatically activates GTK or WX threading for |
|
|||
118 | IPyhton if the choice of matplotlib backend requires it. It |
|
|||
119 | also modifies the %run command to correctly execute (without |
|
|||
120 | blocking) any matplotlib-based script which calls show() at |
|
|||
121 | the end. See `Matplotlib support`_ for more details. |
|
|||
122 |
|
59 | |||
123 | -autocall <val> |
|
60 | -autocall <val> | |
124 | Make IPython automatically call any callable object even if you |
|
61 | Make IPython automatically call any callable object even if you | |
125 | didn't type explicit parentheses. For example, 'str 43' becomes |
|
62 | didn't type explicit parentheses. For example, 'str 43' becomes | |
126 | 'str(43)' automatically. The value can be '0' to disable the feature, |
|
63 | 'str(43)' automatically. The value can be '0' to disable the feature, | |
127 | '1' for smart autocall, where it is not applied if there are no more |
|
64 | '1' for smart autocall, where it is not applied if there are no more | |
128 | arguments on the line, and '2' for full autocall, where all callable |
|
65 | arguments on the line, and '2' for full autocall, where all callable | |
129 | objects are automatically called (even if no arguments are |
|
66 | objects are automatically called (even if no arguments are | |
130 | present). The default is '1'. |
|
67 | present). The default is '1'. | |
131 |
|
68 | |||
132 | -[no]autoindent |
|
69 | -[no]autoindent | |
133 | Turn automatic indentation on/off. |
|
70 | Turn automatic indentation on/off. | |
134 |
|
71 | |||
135 | -[no]automagic |
|
72 | -[no]automagic | |
136 | make magic commands automatic (without needing their first character |
|
73 | make magic commands automatic (without needing their first character | |
137 | to be %). Type %magic at the IPython prompt for more information. |
|
74 | to be %). Type %magic at the IPython prompt for more information. | |
138 |
|
75 | |||
139 | -[no]autoedit_syntax |
|
76 | -[no]autoedit_syntax | |
140 | When a syntax error occurs after editing a file, automatically |
|
77 | When a syntax error occurs after editing a file, automatically | |
141 | open the file to the trouble causing line for convenient |
|
78 | open the file to the trouble causing line for convenient | |
142 | fixing. |
|
79 | fixing. | |
143 |
|
80 | |||
144 | -[no]banner Print the initial information banner (default on). |
|
81 | -[no]banner Print the initial information banner (default on). | |
145 |
|
82 | |||
146 | -c <command> |
|
83 | -c <command> | |
147 | execute the given command string. This is similar to the -c |
|
84 | execute the given command string. This is similar to the -c | |
148 | option in the normal Python interpreter. |
|
85 | option in the normal Python interpreter. | |
149 |
|
86 | |||
150 | -cache_size, cs <n> |
|
87 | -cache_size, cs <n> | |
151 | size of the output cache (maximum number of entries to hold in |
|
88 | size of the output cache (maximum number of entries to hold in | |
152 | memory). The default is 1000, you can change it permanently in your |
|
89 | memory). The default is 1000, you can change it permanently in your | |
153 | config file. Setting it to 0 completely disables the caching system, |
|
90 | config file. Setting it to 0 completely disables the caching system, | |
154 | and the minimum value accepted is 20 (if you provide a value less than |
|
91 | and the minimum value accepted is 20 (if you provide a value less than | |
155 | 20, it is reset to 0 and a warning is issued) This limit is defined |
|
92 | 20, it is reset to 0 and a warning is issued) This limit is defined | |
156 | because otherwise you'll spend more time re-flushing a too small cache |
|
93 | because otherwise you'll spend more time re-flushing a too small cache | |
157 | than working. |
|
94 | than working. | |
158 |
|
95 | |||
159 | -classic, cl |
|
96 | -classic, cl | |
160 | Gives IPython a similar feel to the classic Python |
|
97 | Gives IPython a similar feel to the classic Python | |
161 | prompt. |
|
98 | prompt. | |
162 |
|
99 | |||
163 | -colors <scheme> |
|
100 | -colors <scheme> | |
164 | Color scheme for prompts and exception reporting. Currently |
|
101 | Color scheme for prompts and exception reporting. Currently | |
165 | implemented: NoColor, Linux and LightBG. |
|
102 | implemented: NoColor, Linux and LightBG. | |
166 |
|
103 | |||
167 | -[no]color_info |
|
104 | -[no]color_info | |
168 | IPython can display information about objects via a set of functions, |
|
105 | IPython can display information about objects via a set of functions, | |
169 | and optionally can use colors for this, syntax highlighting source |
|
106 | and optionally can use colors for this, syntax highlighting source | |
170 | code and various other elements. However, because this information is |
|
107 | code and various other elements. However, because this information is | |
171 | passed through a pager (like 'less') and many pagers get confused with |
|
108 | passed through a pager (like 'less') and many pagers get confused with | |
172 | color codes, this option is off by default. You can test it and turn |
|
109 | color codes, this option is off by default. You can test it and turn | |
173 | it on permanently in your ipythonrc file if it works for you. As a |
|
110 | it on permanently in your ipythonrc file if it works for you. As a | |
174 | reference, the 'less' pager supplied with Mandrake 8.2 works ok, but |
|
111 | reference, the 'less' pager supplied with Mandrake 8.2 works ok, but | |
175 | that in RedHat 7.2 doesn't. |
|
112 | that in RedHat 7.2 doesn't. | |
176 |
|
113 | |||
177 | Test it and turn it on permanently if it works with your |
|
114 | Test it and turn it on permanently if it works with your | |
178 | system. The magic function %color_info allows you to toggle this |
|
115 | system. The magic function %color_info allows you to toggle this | |
179 | interactively for testing. |
|
116 | interactively for testing. | |
180 |
|
117 | |||
181 | -[no]debug |
|
118 | -[no]debug | |
182 | Show information about the loading process. Very useful to pin down |
|
119 | Show information about the loading process. Very useful to pin down | |
183 | problems with your configuration files or to get details about |
|
120 | problems with your configuration files or to get details about | |
184 | session restores. |
|
121 | session restores. | |
185 |
|
122 | |||
186 | -[no]deep_reload: |
|
123 | -[no]deep_reload: | |
187 | IPython can use the deep_reload module which reloads changes in |
|
124 | IPython can use the deep_reload module which reloads changes in | |
188 | modules recursively (it replaces the reload() function, so you don't |
|
125 | modules recursively (it replaces the reload() function, so you don't | |
189 | need to change anything to use it). deep_reload() forces a full |
|
126 | need to change anything to use it). deep_reload() forces a full | |
190 | reload of modules whose code may have changed, which the default |
|
127 | reload of modules whose code may have changed, which the default | |
191 | reload() function does not. |
|
128 | reload() function does not. | |
192 |
|
129 | |||
193 | When deep_reload is off, IPython will use the normal reload(), |
|
130 | When deep_reload is off, IPython will use the normal reload(), | |
194 | but deep_reload will still be available as dreload(). This |
|
131 | but deep_reload will still be available as dreload(). This | |
195 | feature is off by default [which means that you have both |
|
132 | feature is off by default [which means that you have both | |
196 | normal reload() and dreload()]. |
|
133 | normal reload() and dreload()]. | |
197 |
|
134 | |||
198 | -editor <name> |
|
135 | -editor <name> | |
199 | Which editor to use with the %edit command. By default, |
|
136 | Which editor to use with the %edit command. By default, | |
200 | IPython will honor your EDITOR environment variable (if not |
|
137 | IPython will honor your EDITOR environment variable (if not | |
201 | set, vi is the Unix default and notepad the Windows one). |
|
138 | set, vi is the Unix default and notepad the Windows one). | |
202 | Since this editor is invoked on the fly by IPython and is |
|
139 | Since this editor is invoked on the fly by IPython and is | |
203 | meant for editing small code snippets, you may want to use a |
|
140 | meant for editing small code snippets, you may want to use a | |
204 | small, lightweight editor here (in case your default EDITOR is |
|
141 | small, lightweight editor here (in case your default EDITOR is | |
205 | something like Emacs). |
|
142 | something like Emacs). | |
206 |
|
143 | |||
207 | -ipythondir <name> |
|
144 | -ipythondir <name> | |
208 | name of your IPython configuration directory IPYTHONDIR. This |
|
145 | name of your IPython configuration directory IPYTHONDIR. This | |
209 | can also be specified through the environment variable |
|
146 | can also be specified through the environment variable | |
210 | IPYTHONDIR. |
|
147 | IPYTHONDIR. | |
211 |
|
148 | |||
212 | -log, l |
|
149 | -log, l | |
213 | generate a log file of all input. The file is named |
|
150 | generate a log file of all input. The file is named | |
214 | ipython_log.py in your current directory (which prevents logs |
|
151 | ipython_log.py in your current directory (which prevents logs | |
215 | from multiple IPython sessions from trampling each other). You |
|
152 | from multiple IPython sessions from trampling each other). You | |
216 | can use this to later restore a session by loading your |
|
153 | can use this to later restore a session by loading your | |
217 | logfile as a file to be executed with option -logplay (see |
|
154 | logfile as a file to be executed with option -logplay (see | |
218 | below). |
|
155 | below). | |
219 |
|
156 | |||
220 | -logfile, lf <name> specify the name of your logfile. |
|
157 | -logfile, lf <name> specify the name of your logfile. | |
221 |
|
158 | |||
222 | -logplay, lp <name> |
|
159 | -logplay, lp <name> | |
223 |
|
160 | |||
224 | you can replay a previous log. For restoring a session as close as |
|
161 | you can replay a previous log. For restoring a session as close as | |
225 | possible to the state you left it in, use this option (don't just run |
|
162 | possible to the state you left it in, use this option (don't just run | |
226 | the logfile). With -logplay, IPython will try to reconstruct the |
|
163 | the logfile). With -logplay, IPython will try to reconstruct the | |
227 | previous working environment in full, not just execute the commands in |
|
164 | previous working environment in full, not just execute the commands in | |
228 | the logfile. |
|
165 | the logfile. | |
229 |
|
166 | |||
230 | When a session is restored, logging is automatically turned on |
|
167 | When a session is restored, logging is automatically turned on | |
231 | again with the name of the logfile it was invoked with (it is |
|
168 | again with the name of the logfile it was invoked with (it is | |
232 | read from the log header). So once you've turned logging on for |
|
169 | read from the log header). So once you've turned logging on for | |
233 | a session, you can quit IPython and reload it as many times as |
|
170 | a session, you can quit IPython and reload it as many times as | |
234 | you want and it will continue to log its history and restore |
|
171 | you want and it will continue to log its history and restore | |
235 | from the beginning every time. |
|
172 | from the beginning every time. | |
236 |
|
173 | |||
237 | Caveats: there are limitations in this option. The history |
|
174 | Caveats: there are limitations in this option. The history | |
238 | variables _i*,_* and _dh don't get restored properly. In the |
|
175 | variables _i*,_* and _dh don't get restored properly. In the | |
239 | future we will try to implement full session saving by writing |
|
176 | future we will try to implement full session saving by writing | |
240 | and retrieving a 'snapshot' of the memory state of IPython. But |
|
177 | and retrieving a 'snapshot' of the memory state of IPython. But | |
241 | our first attempts failed because of inherent limitations of |
|
178 | our first attempts failed because of inherent limitations of | |
242 | Python's Pickle module, so this may have to wait. |
|
179 | Python's Pickle module, so this may have to wait. | |
243 |
|
180 | |||
244 | -[no]messages |
|
181 | -[no]messages | |
245 | Print messages which IPython collects about its startup |
|
182 | Print messages which IPython collects about its startup | |
246 | process (default on). |
|
183 | process (default on). | |
247 |
|
184 | |||
248 | -[no]pdb |
|
185 | -[no]pdb | |
249 | Automatically call the pdb debugger after every uncaught |
|
186 | Automatically call the pdb debugger after every uncaught | |
250 | exception. If you are used to debugging using pdb, this puts |
|
187 | exception. If you are used to debugging using pdb, this puts | |
251 | you automatically inside of it after any call (either in |
|
188 | you automatically inside of it after any call (either in | |
252 | IPython or in code called by it) which triggers an exception |
|
189 | IPython or in code called by it) which triggers an exception | |
253 | which goes uncaught. |
|
190 | which goes uncaught. | |
254 |
|
191 | |||
255 | -pydb |
|
192 | -pydb | |
256 | Makes IPython use the third party "pydb" package as debugger, |
|
193 | Makes IPython use the third party "pydb" package as debugger, | |
257 | instead of pdb. Requires that pydb is installed. |
|
194 | instead of pdb. Requires that pydb is installed. | |
258 |
|
195 | |||
259 | -[no]pprint |
|
196 | -[no]pprint | |
260 | ipython can optionally use the pprint (pretty printer) module |
|
197 | ipython can optionally use the pprint (pretty printer) module | |
261 | for displaying results. pprint tends to give a nicer display |
|
198 | for displaying results. pprint tends to give a nicer display | |
262 | of nested data structures. If you like it, you can turn it on |
|
199 | of nested data structures. If you like it, you can turn it on | |
263 | permanently in your config file (default off). |
|
200 | permanently in your config file (default off). | |
264 |
|
201 | |||
265 | -profile, p <name> |
|
202 | -profile, p <name> | |
266 |
|
203 | |||
267 | assume that your config file is ipythonrc-<name> or |
|
204 | assume that your config file is ipythonrc-<name> or | |
268 | ipy_profile_<name>.py (looks in current dir first, then in |
|
205 | ipy_profile_<name>.py (looks in current dir first, then in | |
269 | IPYTHONDIR). This is a quick way to keep and load multiple |
|
206 | IPYTHONDIR). This is a quick way to keep and load multiple | |
270 | config files for different tasks, especially if you use the |
|
207 | config files for different tasks, especially if you use the | |
271 | include option of config files. You can keep a basic |
|
208 | include option of config files. You can keep a basic | |
272 | IPYTHONDIR/ipythonrc file and then have other 'profiles' which |
|
209 | IPYTHONDIR/ipythonrc file and then have other 'profiles' which | |
273 | include this one and load extra things for particular |
|
210 | include this one and load extra things for particular | |
274 | tasks. For example: |
|
211 | tasks. For example: | |
275 |
|
212 | |||
276 | 1. $HOME/.ipython/ipythonrc : load basic things you always want. |
|
213 | 1. $HOME/.ipython/ipythonrc : load basic things you always want. | |
277 | 2. $HOME/.ipython/ipythonrc-math : load (1) and basic math-related modules. |
|
214 | 2. $HOME/.ipython/ipythonrc-math : load (1) and basic math-related modules. | |
278 | 3. $HOME/.ipython/ipythonrc-numeric : load (1) and Numeric and plotting modules. |
|
215 | 3. $HOME/.ipython/ipythonrc-numeric : load (1) and Numeric and plotting modules. | |
279 |
|
216 | |||
280 | Since it is possible to create an endless loop by having |
|
217 | Since it is possible to create an endless loop by having | |
281 | circular file inclusions, IPython will stop if it reaches 15 |
|
218 | circular file inclusions, IPython will stop if it reaches 15 | |
282 | recursive inclusions. |
|
219 | recursive inclusions. | |
283 |
|
220 | |||
284 | -prompt_in1, pi1 <string> |
|
221 | -prompt_in1, pi1 <string> | |
285 |
|
222 | |||
286 | Specify the string used for input prompts. Note that if you are using |
|
223 | Specify the string used for input prompts. Note that if you are using | |
287 | numbered prompts, the number is represented with a '\#' in the |
|
224 | numbered prompts, the number is represented with a '\#' in the | |
288 | string. Don't forget to quote strings with spaces embedded in |
|
225 | string. Don't forget to quote strings with spaces embedded in | |
289 | them. Default: 'In [\#]:'. The :ref:`prompts section <prompts>` |
|
226 | them. Default: 'In [\#]:'. The :ref:`prompts section <prompts>` | |
290 | discusses in detail all the available escapes to customize your |
|
227 | discusses in detail all the available escapes to customize your | |
291 | prompts. |
|
228 | prompts. | |
292 |
|
229 | |||
293 | -prompt_in2, pi2 <string> |
|
230 | -prompt_in2, pi2 <string> | |
294 | Similar to the previous option, but used for the continuation |
|
231 | Similar to the previous option, but used for the continuation | |
295 | prompts. The special sequence '\D' is similar to '\#', but |
|
232 | prompts. The special sequence '\D' is similar to '\#', but | |
296 | with all digits replaced dots (so you can have your |
|
233 | with all digits replaced dots (so you can have your | |
297 | continuation prompt aligned with your input prompt). Default: |
|
234 | continuation prompt aligned with your input prompt). Default: | |
298 | ' .\D.:' (note three spaces at the start for alignment with |
|
235 | ' .\D.:' (note three spaces at the start for alignment with | |
299 | 'In [\#]'). |
|
236 | 'In [\#]'). | |
300 |
|
237 | |||
301 | -prompt_out,po <string> |
|
238 | -prompt_out,po <string> | |
302 | String used for output prompts, also uses numbers like |
|
239 | String used for output prompts, also uses numbers like | |
303 | prompt_in1. Default: 'Out[\#]:' |
|
240 | prompt_in1. Default: 'Out[\#]:' | |
304 |
|
241 | |||
305 | -quick start in bare bones mode (no config file loaded). |
|
242 | -quick start in bare bones mode (no config file loaded). | |
306 |
|
243 | |||
307 | -rcfile <name> |
|
244 | -rcfile <name> | |
308 | name of your IPython resource configuration file. Normally |
|
245 | name of your IPython resource configuration file. Normally | |
309 | IPython loads ipythonrc (from current directory) or |
|
246 | IPython loads ipythonrc (from current directory) or | |
310 | IPYTHONDIR/ipythonrc. |
|
247 | IPYTHONDIR/ipythonrc. | |
311 |
|
248 | |||
312 | If the loading of your config file fails, IPython starts with |
|
249 | If the loading of your config file fails, IPython starts with | |
313 | a bare bones configuration (no modules loaded at all). |
|
250 | a bare bones configuration (no modules loaded at all). | |
314 |
|
251 | |||
315 | -[no]readline |
|
252 | -[no]readline | |
316 | use the readline library, which is needed to support name |
|
253 | use the readline library, which is needed to support name | |
317 | completion and command history, among other things. It is |
|
254 | completion and command history, among other things. It is | |
318 | enabled by default, but may cause problems for users of |
|
255 | enabled by default, but may cause problems for users of | |
319 | X/Emacs in Python comint or shell buffers. |
|
256 | X/Emacs in Python comint or shell buffers. | |
320 |
|
257 | |||
321 | Note that X/Emacs 'eterm' buffers (opened with M-x term) support |
|
258 | Note that X/Emacs 'eterm' buffers (opened with M-x term) support | |
322 | IPython's readline and syntax coloring fine, only 'emacs' (M-x |
|
259 | IPython's readline and syntax coloring fine, only 'emacs' (M-x | |
323 | shell and C-c !) buffers do not. |
|
260 | shell and C-c !) buffers do not. | |
324 |
|
261 | |||
325 | -screen_length, sl <n> |
|
262 | -screen_length, sl <n> | |
326 | number of lines of your screen. This is used to control |
|
263 | number of lines of your screen. This is used to control | |
327 | printing of very long strings. Strings longer than this number |
|
264 | printing of very long strings. Strings longer than this number | |
328 | of lines will be sent through a pager instead of directly |
|
265 | of lines will be sent through a pager instead of directly | |
329 | printed. |
|
266 | printed. | |
330 |
|
267 | |||
331 | The default value for this is 0, which means IPython will |
|
268 | The default value for this is 0, which means IPython will | |
332 | auto-detect your screen size every time it needs to print certain |
|
269 | auto-detect your screen size every time it needs to print certain | |
333 | potentially long strings (this doesn't change the behavior of the |
|
270 | potentially long strings (this doesn't change the behavior of the | |
334 | 'print' keyword, it's only triggered internally). If for some |
|
271 | 'print' keyword, it's only triggered internally). If for some | |
335 | reason this isn't working well (it needs curses support), specify |
|
272 | reason this isn't working well (it needs curses support), specify | |
336 | it yourself. Otherwise don't change the default. |
|
273 | it yourself. Otherwise don't change the default. | |
337 |
|
274 | |||
338 | -separate_in, si <string> |
|
275 | -separate_in, si <string> | |
339 |
|
276 | |||
340 | separator before input prompts. |
|
277 | separator before input prompts. | |
341 | Default: '\n' |
|
278 | Default: '\n' | |
342 |
|
279 | |||
343 | -separate_out, so <string> |
|
280 | -separate_out, so <string> | |
344 | separator before output prompts. |
|
281 | separator before output prompts. | |
345 | Default: nothing. |
|
282 | Default: nothing. | |
346 |
|
283 | |||
347 | -separate_out2, so2 |
|
284 | -separate_out2, so2 | |
348 | separator after output prompts. |
|
285 | separator after output prompts. | |
349 | Default: nothing. |
|
286 | Default: nothing. | |
350 | For these three options, use the value 0 to specify no separator. |
|
287 | For these three options, use the value 0 to specify no separator. | |
351 |
|
288 | |||
352 | -nosep |
|
289 | -nosep | |
353 | shorthand for '-SeparateIn 0 -SeparateOut 0 -SeparateOut2 |
|
290 | shorthand for '-SeparateIn 0 -SeparateOut 0 -SeparateOut2 | |
354 | 0'. Simply removes all input/output separators. |
|
291 | 0'. Simply removes all input/output separators. | |
355 |
|
292 | |||
356 | -upgrade |
|
293 | -upgrade | |
357 | allows you to upgrade your IPYTHONDIR configuration when you |
|
294 | allows you to upgrade your IPYTHONDIR configuration when you | |
358 | install a new version of IPython. Since new versions may |
|
295 | install a new version of IPython. Since new versions may | |
359 | include new command line options or example files, this copies |
|
296 | include new command line options or example files, this copies | |
360 | updated ipythonrc-type files. However, it backs up (with a |
|
297 | updated ipythonrc-type files. However, it backs up (with a | |
361 | .old extension) all files which it overwrites so that you can |
|
298 | .old extension) all files which it overwrites so that you can | |
362 | merge back any customizations you might have in your personal |
|
299 | merge back any customizations you might have in your personal | |
363 | files. Note that you should probably use %upgrade instead, |
|
300 | files. Note that you should probably use %upgrade instead, | |
364 | it's a safer alternative. |
|
301 | it's a safer alternative. | |
365 |
|
302 | |||
366 |
|
303 | |||
367 | -Version print version information and exit. |
|
304 | -Version print version information and exit. | |
368 |
|
305 | |||
369 | -wxversion <string> |
|
306 | -wxversion <string> | |
370 | Select a specific version of wxPython (used in conjunction |
|
307 | Deprecated. | |
371 | with -wthread). Requires the wxversion module, part of recent |
|
|||
372 | wxPython distributions |
|
|||
373 |
|
308 | |||
374 | -xmode <modename> |
|
309 | -xmode <modename> | |
375 |
|
310 | |||
376 | Mode for exception reporting. |
|
311 | Mode for exception reporting. | |
377 |
|
312 | |||
378 | Valid modes: Plain, Context and Verbose. |
|
313 | Valid modes: Plain, Context and Verbose. | |
379 |
|
314 | |||
380 | * Plain: similar to python's normal traceback printing. |
|
315 | * Plain: similar to python's normal traceback printing. | |
381 | * Context: prints 5 lines of context source code around each |
|
316 | * Context: prints 5 lines of context source code around each | |
382 | line in the traceback. |
|
317 | line in the traceback. | |
383 | * Verbose: similar to Context, but additionally prints the |
|
318 | * Verbose: similar to Context, but additionally prints the | |
384 | variables currently visible where the exception happened |
|
319 | variables currently visible where the exception happened | |
385 | (shortening their strings if too long). This can potentially be |
|
320 | (shortening their strings if too long). This can potentially be | |
386 | very slow, if you happen to have a huge data structure whose |
|
321 | very slow, if you happen to have a huge data structure whose | |
387 | string representation is complex to compute. Your computer may |
|
322 | string representation is complex to compute. Your computer may | |
388 | appear to freeze for a while with cpu usage at 100%. If this |
|
323 | appear to freeze for a while with cpu usage at 100%. If this | |
389 | occurs, you can cancel the traceback with Ctrl-C (maybe hitting it |
|
324 | occurs, you can cancel the traceback with Ctrl-C (maybe hitting it | |
390 | more than once). |
|
325 | more than once). | |
391 |
|
326 | |||
392 | Interactive use |
|
327 | Interactive use | |
393 | =============== |
|
328 | =============== | |
394 |
|
329 | |||
395 | Warning: IPython relies on the existence of a global variable called |
|
330 | Warning: IPython relies on the existence of a global variable called | |
396 | _ip which controls the shell itself. If you redefine _ip to anything, |
|
331 | _ip which controls the shell itself. If you redefine _ip to anything, | |
397 | bizarre behavior will quickly occur. |
|
332 | bizarre behavior will quickly occur. | |
398 |
|
333 | |||
399 | Other than the above warning, IPython is meant to work as a drop-in |
|
334 | Other than the above warning, IPython is meant to work as a drop-in | |
400 | replacement for the standard interactive interpreter. As such, any code |
|
335 | replacement for the standard interactive interpreter. As such, any code | |
401 | which is valid python should execute normally under IPython (cases where |
|
336 | which is valid python should execute normally under IPython (cases where | |
402 | this is not true should be reported as bugs). It does, however, offer |
|
337 | this is not true should be reported as bugs). It does, however, offer | |
403 | many features which are not available at a standard python prompt. What |
|
338 | many features which are not available at a standard python prompt. What | |
404 | follows is a list of these. |
|
339 | follows is a list of these. | |
405 |
|
340 | |||
406 |
|
341 | |||
407 | Caution for Windows users |
|
342 | Caution for Windows users | |
408 | ------------------------- |
|
343 | ------------------------- | |
409 |
|
344 | |||
410 | Windows, unfortunately, uses the '\' character as a path |
|
345 | Windows, unfortunately, uses the '\' character as a path | |
411 | separator. This is a terrible choice, because '\' also represents the |
|
346 | separator. This is a terrible choice, because '\' also represents the | |
412 | escape character in most modern programming languages, including |
|
347 | escape character in most modern programming languages, including | |
413 | Python. For this reason, using '/' character is recommended if you |
|
348 | Python. For this reason, using '/' character is recommended if you | |
414 | have problems with ``\``. However, in Windows commands '/' flags |
|
349 | have problems with ``\``. However, in Windows commands '/' flags | |
415 | options, so you can not use it for the root directory. This means that |
|
350 | options, so you can not use it for the root directory. This means that | |
416 | paths beginning at the root must be typed in a contrived manner like: |
|
351 | paths beginning at the root must be typed in a contrived manner like: | |
417 | ``%copy \opt/foo/bar.txt \tmp`` |
|
352 | ``%copy \opt/foo/bar.txt \tmp`` | |
418 |
|
353 | |||
419 | .. _magic: |
|
354 | .. _magic: | |
420 |
|
355 | |||
421 | Magic command system |
|
356 | Magic command system | |
422 | -------------------- |
|
357 | -------------------- | |
423 |
|
358 | |||
424 | IPython will treat any line whose first character is a % as a special |
|
359 | IPython will treat any line whose first character is a % as a special | |
425 | call to a 'magic' function. These allow you to control the behavior of |
|
360 | call to a 'magic' function. These allow you to control the behavior of | |
426 | IPython itself, plus a lot of system-type features. They are all |
|
361 | IPython itself, plus a lot of system-type features. They are all | |
427 | prefixed with a % character, but parameters are given without |
|
362 | prefixed with a % character, but parameters are given without | |
428 | parentheses or quotes. |
|
363 | parentheses or quotes. | |
429 |
|
364 | |||
430 | Example: typing '%cd mydir' (without the quotes) changes you working |
|
365 | Example: typing '%cd mydir' (without the quotes) changes you working | |
431 | directory to 'mydir', if it exists. |
|
366 | directory to 'mydir', if it exists. | |
432 |
|
367 | |||
433 | If you have 'automagic' enabled (in your ipythonrc file, via the command |
|
368 | If you have 'automagic' enabled (in your ipythonrc file, via the command | |
434 | line option -automagic or with the %automagic function), you don't need |
|
369 | line option -automagic or with the %automagic function), you don't need | |
435 | to type in the % explicitly. IPython will scan its internal list of |
|
370 | to type in the % explicitly. IPython will scan its internal list of | |
436 | magic functions and call one if it exists. With automagic on you can |
|
371 | magic functions and call one if it exists. With automagic on you can | |
437 | then just type 'cd mydir' to go to directory 'mydir'. The automagic |
|
372 | then just type 'cd mydir' to go to directory 'mydir'. The automagic | |
438 | system has the lowest possible precedence in name searches, so defining |
|
373 | system has the lowest possible precedence in name searches, so defining | |
439 | an identifier with the same name as an existing magic function will |
|
374 | an identifier with the same name as an existing magic function will | |
440 | shadow it for automagic use. You can still access the shadowed magic |
|
375 | shadow it for automagic use. You can still access the shadowed magic | |
441 | function by explicitly using the % character at the beginning of the line. |
|
376 | function by explicitly using the % character at the beginning of the line. | |
442 |
|
377 | |||
443 | An example (with automagic on) should clarify all this:: |
|
378 | An example (with automagic on) should clarify all this:: | |
444 |
|
379 | |||
445 | In [1]: cd ipython # %cd is called by automagic |
|
380 | In [1]: cd ipython # %cd is called by automagic | |
446 |
|
381 | |||
447 | /home/fperez/ipython |
|
382 | /home/fperez/ipython | |
448 |
|
383 | |||
449 | In [2]: cd=1 # now cd is just a variable |
|
384 | In [2]: cd=1 # now cd is just a variable | |
450 |
|
385 | |||
451 | In [3]: cd .. # and doesn't work as a function anymore |
|
386 | In [3]: cd .. # and doesn't work as a function anymore | |
452 |
|
387 | |||
453 | ------------------------------ |
|
388 | ------------------------------ | |
454 |
|
389 | |||
455 | File "<console>", line 1 |
|
390 | File "<console>", line 1 | |
456 |
|
391 | |||
457 | cd .. |
|
392 | cd .. | |
458 |
|
393 | |||
459 | ^ |
|
394 | ^ | |
460 |
|
395 | |||
461 | SyntaxError: invalid syntax |
|
396 | SyntaxError: invalid syntax | |
462 |
|
397 | |||
463 | In [4]: %cd .. # but %cd always works |
|
398 | In [4]: %cd .. # but %cd always works | |
464 |
|
399 | |||
465 | /home/fperez |
|
400 | /home/fperez | |
466 |
|
401 | |||
467 | In [5]: del cd # if you remove the cd variable |
|
402 | In [5]: del cd # if you remove the cd variable | |
468 |
|
403 | |||
469 | In [6]: cd ipython # automagic can work again |
|
404 | In [6]: cd ipython # automagic can work again | |
470 |
|
405 | |||
471 | /home/fperez/ipython |
|
406 | /home/fperez/ipython | |
472 |
|
407 | |||
473 | You can define your own magic functions to extend the system. The |
|
408 | You can define your own magic functions to extend the system. The | |
474 | following example defines a new magic command, %impall:: |
|
409 | following example defines a new magic command, %impall:: | |
475 |
|
410 | |||
476 | import IPython.ipapi |
|
411 | import IPython.ipapi | |
477 |
|
412 | |||
478 | ip = IPython.ipapi.get() |
|
413 | ip = IPython.ipapi.get() | |
479 |
|
414 | |||
480 | def doimp(self, arg): |
|
415 | def doimp(self, arg): | |
481 |
|
416 | |||
482 | ip = self.api |
|
417 | ip = self.api | |
483 |
|
418 | |||
484 | ip.ex("import %s; reload(%s); from %s import *" % ( |
|
419 | ip.ex("import %s; reload(%s); from %s import *" % ( | |
485 |
|
420 | |||
486 | arg,arg,arg) |
|
421 | arg,arg,arg) | |
487 |
|
422 | |||
488 | ) |
|
423 | ) | |
489 |
|
424 | |||
490 | ip.expose_magic('impall', doimp) |
|
425 | ip.expose_magic('impall', doimp) | |
491 |
|
426 | |||
492 | You can also define your own aliased names for magic functions. In your |
|
427 | You can also define your own aliased names for magic functions. In your | |
493 | ipythonrc file, placing a line like:: |
|
428 | ipythonrc file, placing a line like:: | |
494 |
|
429 | |||
495 | execute __IP.magic_cl = __IP.magic_clear |
|
430 | execute __IP.magic_cl = __IP.magic_clear | |
496 |
|
431 | |||
497 | will define %cl as a new name for %clear. |
|
432 | will define %cl as a new name for %clear. | |
498 |
|
433 | |||
499 | Type %magic for more information, including a list of all available |
|
434 | Type %magic for more information, including a list of all available | |
500 | magic functions at any time and their docstrings. You can also type |
|
435 | magic functions at any time and their docstrings. You can also type | |
501 | %magic_function_name? (see sec. 6.4 <#sec:dyn-object-info> for |
|
436 | %magic_function_name? (see sec. 6.4 <#sec:dyn-object-info> for | |
502 | information on the '?' system) to get information about any particular |
|
437 | information on the '?' system) to get information about any particular | |
503 | magic function you are interested in. |
|
438 | magic function you are interested in. | |
504 |
|
439 | |||
505 | The API documentation for the :mod:`IPython.Magic` module contains the full |
|
440 | The API documentation for the :mod:`IPython.Magic` module contains the full | |
506 | docstrings of all currently available magic commands. |
|
441 | docstrings of all currently available magic commands. | |
507 |
|
442 | |||
508 |
|
443 | |||
509 | Access to the standard Python help |
|
444 | Access to the standard Python help | |
510 | ---------------------------------- |
|
445 | ---------------------------------- | |
511 |
|
446 | |||
512 | As of Python 2.1, a help system is available with access to object docstrings |
|
447 | As of Python 2.1, a help system is available with access to object docstrings | |
513 | and the Python manuals. Simply type 'help' (no quotes) to access it. You can |
|
448 | and the Python manuals. Simply type 'help' (no quotes) to access it. You can | |
514 | also type help(object) to obtain information about a given object, and |
|
449 | also type help(object) to obtain information about a given object, and | |
515 | help('keyword') for information on a keyword. As noted :ref:`here |
|
450 | help('keyword') for information on a keyword. As noted :ref:`here | |
516 | <accessing_help>`, you need to properly configure your environment variable |
|
451 | <accessing_help>`, you need to properly configure your environment variable | |
517 | PYTHONDOCS for this feature to work correctly. |
|
452 | PYTHONDOCS for this feature to work correctly. | |
518 |
|
453 | |||
519 | .. _dynamic_object_info: |
|
454 | .. _dynamic_object_info: | |
520 |
|
455 | |||
521 | Dynamic object information |
|
456 | Dynamic object information | |
522 | -------------------------- |
|
457 | -------------------------- | |
523 |
|
458 | |||
524 | Typing ?word or word? prints detailed information about an object. If |
|
459 | Typing ?word or word? prints detailed information about an object. If | |
525 | certain strings in the object are too long (docstrings, code, etc.) they |
|
460 | certain strings in the object are too long (docstrings, code, etc.) they | |
526 | get snipped in the center for brevity. This system gives access variable |
|
461 | get snipped in the center for brevity. This system gives access variable | |
527 | types and values, full source code for any object (if available), |
|
462 | types and values, full source code for any object (if available), | |
528 | function prototypes and other useful information. |
|
463 | function prototypes and other useful information. | |
529 |
|
464 | |||
530 | Typing ??word or word?? gives access to the full information without |
|
465 | Typing ??word or word?? gives access to the full information without | |
531 | snipping long strings. Long strings are sent to the screen through the |
|
466 | snipping long strings. Long strings are sent to the screen through the | |
532 | less pager if longer than the screen and printed otherwise. On systems |
|
467 | less pager if longer than the screen and printed otherwise. On systems | |
533 | lacking the less command, IPython uses a very basic internal pager. |
|
468 | lacking the less command, IPython uses a very basic internal pager. | |
534 |
|
469 | |||
535 | The following magic functions are particularly useful for gathering |
|
470 | The following magic functions are particularly useful for gathering | |
536 | information about your working environment. You can get more details by |
|
471 | information about your working environment. You can get more details by | |
537 | typing %magic or querying them individually (use %function_name? with or |
|
472 | typing %magic or querying them individually (use %function_name? with or | |
538 | without the %), this is just a summary: |
|
473 | without the %), this is just a summary: | |
539 |
|
474 | |||
540 | * **%pdoc <object>**: Print (or run through a pager if too long) the |
|
475 | * **%pdoc <object>**: Print (or run through a pager if too long) the | |
541 | docstring for an object. If the given object is a class, it will |
|
476 | docstring for an object. If the given object is a class, it will | |
542 | print both the class and the constructor docstrings. |
|
477 | print both the class and the constructor docstrings. | |
543 | * **%pdef <object>**: Print the definition header for any callable |
|
478 | * **%pdef <object>**: Print the definition header for any callable | |
544 | object. If the object is a class, print the constructor information. |
|
479 | object. If the object is a class, print the constructor information. | |
545 | * **%psource <object>**: Print (or run through a pager if too long) |
|
480 | * **%psource <object>**: Print (or run through a pager if too long) | |
546 | the source code for an object. |
|
481 | the source code for an object. | |
547 | * **%pfile <object>**: Show the entire source file where an object was |
|
482 | * **%pfile <object>**: Show the entire source file where an object was | |
548 | defined via a pager, opening it at the line where the object |
|
483 | defined via a pager, opening it at the line where the object | |
549 | definition begins. |
|
484 | definition begins. | |
550 | * **%who/%whos**: These functions give information about identifiers |
|
485 | * **%who/%whos**: These functions give information about identifiers | |
551 | you have defined interactively (not things you loaded or defined |
|
486 | you have defined interactively (not things you loaded or defined | |
552 | in your configuration files). %who just prints a list of |
|
487 | in your configuration files). %who just prints a list of | |
553 | identifiers and %whos prints a table with some basic details about |
|
488 | identifiers and %whos prints a table with some basic details about | |
554 | each identifier. |
|
489 | each identifier. | |
555 |
|
490 | |||
556 | Note that the dynamic object information functions (?/??, %pdoc, %pfile, |
|
491 | Note that the dynamic object information functions (?/??, %pdoc, %pfile, | |
557 | %pdef, %psource) give you access to documentation even on things which |
|
492 | %pdef, %psource) give you access to documentation even on things which | |
558 | are not really defined as separate identifiers. Try for example typing |
|
493 | are not really defined as separate identifiers. Try for example typing | |
559 | {}.get? or after doing import os, type os.path.abspath??. |
|
494 | {}.get? or after doing import os, type os.path.abspath??. | |
560 |
|
495 | |||
561 |
|
496 | |||
562 | .. _readline: |
|
497 | .. _readline: | |
563 |
|
498 | |||
564 | Readline-based features |
|
499 | Readline-based features | |
565 | ----------------------- |
|
500 | ----------------------- | |
566 |
|
501 | |||
567 | These features require the GNU readline library, so they won't work if |
|
502 | These features require the GNU readline library, so they won't work if | |
568 | your Python installation lacks readline support. We will first describe |
|
503 | your Python installation lacks readline support. We will first describe | |
569 | the default behavior IPython uses, and then how to change it to suit |
|
504 | the default behavior IPython uses, and then how to change it to suit | |
570 | your preferences. |
|
505 | your preferences. | |
571 |
|
506 | |||
572 |
|
507 | |||
573 | Command line completion |
|
508 | Command line completion | |
574 | +++++++++++++++++++++++ |
|
509 | +++++++++++++++++++++++ | |
575 |
|
510 | |||
576 | At any time, hitting TAB will complete any available python commands or |
|
511 | At any time, hitting TAB will complete any available python commands or | |
577 | variable names, and show you a list of the possible completions if |
|
512 | variable names, and show you a list of the possible completions if | |
578 | there's no unambiguous one. It will also complete filenames in the |
|
513 | there's no unambiguous one. It will also complete filenames in the | |
579 | current directory if no python names match what you've typed so far. |
|
514 | current directory if no python names match what you've typed so far. | |
580 |
|
515 | |||
581 |
|
516 | |||
582 | Search command history |
|
517 | Search command history | |
583 | ++++++++++++++++++++++ |
|
518 | ++++++++++++++++++++++ | |
584 |
|
519 | |||
585 | IPython provides two ways for searching through previous input and thus |
|
520 | IPython provides two ways for searching through previous input and thus | |
586 | reduce the need for repetitive typing: |
|
521 | reduce the need for repetitive typing: | |
587 |
|
522 | |||
588 | 1. Start typing, and then use Ctrl-p (previous,up) and Ctrl-n |
|
523 | 1. Start typing, and then use Ctrl-p (previous,up) and Ctrl-n | |
589 | (next,down) to search through only the history items that match |
|
524 | (next,down) to search through only the history items that match | |
590 | what you've typed so far. If you use Ctrl-p/Ctrl-n at a blank |
|
525 | what you've typed so far. If you use Ctrl-p/Ctrl-n at a blank | |
591 | prompt, they just behave like normal arrow keys. |
|
526 | prompt, they just behave like normal arrow keys. | |
592 | 2. Hit Ctrl-r: opens a search prompt. Begin typing and the system |
|
527 | 2. Hit Ctrl-r: opens a search prompt. Begin typing and the system | |
593 | searches your history for lines that contain what you've typed so |
|
528 | searches your history for lines that contain what you've typed so | |
594 | far, completing as much as it can. |
|
529 | far, completing as much as it can. | |
595 |
|
530 | |||
596 |
|
531 | |||
597 | Persistent command history across sessions |
|
532 | Persistent command history across sessions | |
598 | ++++++++++++++++++++++++++++++++++++++++++ |
|
533 | ++++++++++++++++++++++++++++++++++++++++++ | |
599 |
|
534 | |||
600 | IPython will save your input history when it leaves and reload it next |
|
535 | IPython will save your input history when it leaves and reload it next | |
601 | time you restart it. By default, the history file is named |
|
536 | time you restart it. By default, the history file is named | |
602 | $IPYTHONDIR/history, but if you've loaded a named profile, |
|
537 | $IPYTHONDIR/history, but if you've loaded a named profile, | |
603 | '-PROFILE_NAME' is appended to the name. This allows you to keep |
|
538 | '-PROFILE_NAME' is appended to the name. This allows you to keep | |
604 | separate histories related to various tasks: commands related to |
|
539 | separate histories related to various tasks: commands related to | |
605 | numerical work will not be clobbered by a system shell history, for |
|
540 | numerical work will not be clobbered by a system shell history, for | |
606 | example. |
|
541 | example. | |
607 |
|
542 | |||
608 |
|
543 | |||
609 | Autoindent |
|
544 | Autoindent | |
610 | ++++++++++ |
|
545 | ++++++++++ | |
611 |
|
546 | |||
612 | IPython can recognize lines ending in ':' and indent the next line, |
|
547 | IPython can recognize lines ending in ':' and indent the next line, | |
613 | while also un-indenting automatically after 'raise' or 'return'. |
|
548 | while also un-indenting automatically after 'raise' or 'return'. | |
614 |
|
549 | |||
615 | This feature uses the readline library, so it will honor your ~/.inputrc |
|
550 | This feature uses the readline library, so it will honor your ~/.inputrc | |
616 | configuration (or whatever file your INPUTRC variable points to). Adding |
|
551 | configuration (or whatever file your INPUTRC variable points to). Adding | |
617 | the following lines to your .inputrc file can make indenting/unindenting |
|
552 | the following lines to your .inputrc file can make indenting/unindenting | |
618 | more convenient (M-i indents, M-u unindents):: |
|
553 | more convenient (M-i indents, M-u unindents):: | |
619 |
|
554 | |||
620 | $if Python |
|
555 | $if Python | |
621 | "\M-i": " " |
|
556 | "\M-i": " " | |
622 | "\M-u": "\d\d\d\d" |
|
557 | "\M-u": "\d\d\d\d" | |
623 | $endif |
|
558 | $endif | |
624 |
|
559 | |||
625 | Note that there are 4 spaces between the quote marks after "M-i" above. |
|
560 | Note that there are 4 spaces between the quote marks after "M-i" above. | |
626 |
|
561 | |||
627 | Warning: this feature is ON by default, but it can cause problems with |
|
562 | Warning: this feature is ON by default, but it can cause problems with | |
628 | the pasting of multi-line indented code (the pasted code gets |
|
563 | the pasting of multi-line indented code (the pasted code gets | |
629 | re-indented on each line). A magic function %autoindent allows you to |
|
564 | re-indented on each line). A magic function %autoindent allows you to | |
630 | toggle it on/off at runtime. You can also disable it permanently on in |
|
565 | toggle it on/off at runtime. You can also disable it permanently on in | |
631 | your ipythonrc file (set autoindent 0). |
|
566 | your ipythonrc file (set autoindent 0). | |
632 |
|
567 | |||
633 |
|
568 | |||
634 | Customizing readline behavior |
|
569 | Customizing readline behavior | |
635 | +++++++++++++++++++++++++++++ |
|
570 | +++++++++++++++++++++++++++++ | |
636 |
|
571 | |||
637 | All these features are based on the GNU readline library, which has an |
|
572 | All these features are based on the GNU readline library, which has an | |
638 | extremely customizable interface. Normally, readline is configured via a |
|
573 | extremely customizable interface. Normally, readline is configured via a | |
639 | file which defines the behavior of the library; the details of the |
|
574 | file which defines the behavior of the library; the details of the | |
640 | syntax for this can be found in the readline documentation available |
|
575 | syntax for this can be found in the readline documentation available | |
641 | with your system or on the Internet. IPython doesn't read this file (if |
|
576 | with your system or on the Internet. IPython doesn't read this file (if | |
642 | it exists) directly, but it does support passing to readline valid |
|
577 | it exists) directly, but it does support passing to readline valid | |
643 | options via a simple interface. In brief, you can customize readline by |
|
578 | options via a simple interface. In brief, you can customize readline by | |
644 | setting the following options in your ipythonrc configuration file (note |
|
579 | setting the following options in your ipythonrc configuration file (note | |
645 | that these options can not be specified at the command line): |
|
580 | that these options can not be specified at the command line): | |
646 |
|
581 | |||
647 | * **readline_parse_and_bind**: this option can appear as many times as |
|
582 | * **readline_parse_and_bind**: this option can appear as many times as | |
648 | you want, each time defining a string to be executed via a |
|
583 | you want, each time defining a string to be executed via a | |
649 | readline.parse_and_bind() command. The syntax for valid commands |
|
584 | readline.parse_and_bind() command. The syntax for valid commands | |
650 | of this kind can be found by reading the documentation for the GNU |
|
585 | of this kind can be found by reading the documentation for the GNU | |
651 | readline library, as these commands are of the kind which readline |
|
586 | readline library, as these commands are of the kind which readline | |
652 | accepts in its configuration file. |
|
587 | accepts in its configuration file. | |
653 | * **readline_remove_delims**: a string of characters to be removed |
|
588 | * **readline_remove_delims**: a string of characters to be removed | |
654 | from the default word-delimiters list used by readline, so that |
|
589 | from the default word-delimiters list used by readline, so that | |
655 | completions may be performed on strings which contain them. Do not |
|
590 | completions may be performed on strings which contain them. Do not | |
656 | change the default value unless you know what you're doing. |
|
591 | change the default value unless you know what you're doing. | |
657 | * **readline_omit__names**: when tab-completion is enabled, hitting |
|
592 | * **readline_omit__names**: when tab-completion is enabled, hitting | |
658 | <tab> after a '.' in a name will complete all attributes of an |
|
593 | <tab> after a '.' in a name will complete all attributes of an | |
659 | object, including all the special methods whose names include |
|
594 | object, including all the special methods whose names include | |
660 | double underscores (like __getitem__ or __class__). If you'd |
|
595 | double underscores (like __getitem__ or __class__). If you'd | |
661 | rather not see these names by default, you can set this option to |
|
596 | rather not see these names by default, you can set this option to | |
662 | 1. Note that even when this option is set, you can still see those |
|
597 | 1. Note that even when this option is set, you can still see those | |
663 | names by explicitly typing a _ after the period and hitting <tab>: |
|
598 | names by explicitly typing a _ after the period and hitting <tab>: | |
664 | 'name._<tab>' will always complete attribute names starting with '_'. |
|
599 | 'name._<tab>' will always complete attribute names starting with '_'. | |
665 |
|
600 | |||
666 | This option is off by default so that new users see all |
|
601 | This option is off by default so that new users see all | |
667 | attributes of any objects they are dealing with. |
|
602 | attributes of any objects they are dealing with. | |
668 |
|
603 | |||
669 | You will find the default values along with a corresponding detailed |
|
604 | You will find the default values along with a corresponding detailed | |
670 | explanation in your ipythonrc file. |
|
605 | explanation in your ipythonrc file. | |
671 |
|
606 | |||
672 |
|
607 | |||
673 | Session logging and restoring |
|
608 | Session logging and restoring | |
674 | ----------------------------- |
|
609 | ----------------------------- | |
675 |
|
610 | |||
676 | You can log all input from a session either by starting IPython with the |
|
611 | You can log all input from a session either by starting IPython with the | |
677 | command line switches -log or -logfile (see :ref:`here <command_line_options>`) |
|
612 | command line switches -log or -logfile (see :ref:`here <command_line_options>`) | |
678 | or by activating the logging at any moment with the magic function %logstart. |
|
613 | or by activating the logging at any moment with the magic function %logstart. | |
679 |
|
614 | |||
680 | Log files can later be reloaded with the -logplay option and IPython |
|
615 | Log files can later be reloaded with the -logplay option and IPython | |
681 | will attempt to 'replay' the log by executing all the lines in it, thus |
|
616 | will attempt to 'replay' the log by executing all the lines in it, thus | |
682 | restoring the state of a previous session. This feature is not quite |
|
617 | restoring the state of a previous session. This feature is not quite | |
683 | perfect, but can still be useful in many cases. |
|
618 | perfect, but can still be useful in many cases. | |
684 |
|
619 | |||
685 | The log files can also be used as a way to have a permanent record of |
|
620 | The log files can also be used as a way to have a permanent record of | |
686 | any code you wrote while experimenting. Log files are regular text files |
|
621 | any code you wrote while experimenting. Log files are regular text files | |
687 | which you can later open in your favorite text editor to extract code or |
|
622 | which you can later open in your favorite text editor to extract code or | |
688 | to 'clean them up' before using them to replay a session. |
|
623 | to 'clean them up' before using them to replay a session. | |
689 |
|
624 | |||
690 | The %logstart function for activating logging in mid-session is used as |
|
625 | The %logstart function for activating logging in mid-session is used as | |
691 | follows: |
|
626 | follows: | |
692 |
|
627 | |||
693 | %logstart [log_name [log_mode]] |
|
628 | %logstart [log_name [log_mode]] | |
694 |
|
629 | |||
695 | If no name is given, it defaults to a file named 'log' in your |
|
630 | If no name is given, it defaults to a file named 'log' in your | |
696 | IPYTHONDIR directory, in 'rotate' mode (see below). |
|
631 | IPYTHONDIR directory, in 'rotate' mode (see below). | |
697 |
|
632 | |||
698 | '%logstart name' saves to file 'name' in 'backup' mode. It saves your |
|
633 | '%logstart name' saves to file 'name' in 'backup' mode. It saves your | |
699 | history up to that point and then continues logging. |
|
634 | history up to that point and then continues logging. | |
700 |
|
635 | |||
701 | %logstart takes a second optional parameter: logging mode. This can be |
|
636 | %logstart takes a second optional parameter: logging mode. This can be | |
702 | one of (note that the modes are given unquoted): |
|
637 | one of (note that the modes are given unquoted): | |
703 |
|
638 | |||
704 | * [over:] overwrite existing log_name. |
|
639 | * [over:] overwrite existing log_name. | |
705 | * [backup:] rename (if exists) to log_name~ and start log_name. |
|
640 | * [backup:] rename (if exists) to log_name~ and start log_name. | |
706 | * [append:] well, that says it. |
|
641 | * [append:] well, that says it. | |
707 | * [rotate:] create rotating logs log_name.1~, log_name.2~, etc. |
|
642 | * [rotate:] create rotating logs log_name.1~, log_name.2~, etc. | |
708 |
|
643 | |||
709 | The %logoff and %logon functions allow you to temporarily stop and |
|
644 | The %logoff and %logon functions allow you to temporarily stop and | |
710 | resume logging to a file which had previously been started with |
|
645 | resume logging to a file which had previously been started with | |
711 | %logstart. They will fail (with an explanation) if you try to use them |
|
646 | %logstart. They will fail (with an explanation) if you try to use them | |
712 | before logging has been started. |
|
647 | before logging has been started. | |
713 |
|
648 | |||
714 | .. _system_shell_access: |
|
649 | .. _system_shell_access: | |
715 |
|
650 | |||
716 | System shell access |
|
651 | System shell access | |
717 | ------------------- |
|
652 | ------------------- | |
718 |
|
653 | |||
719 | Any input line beginning with a ! character is passed verbatim (minus |
|
654 | Any input line beginning with a ! character is passed verbatim (minus | |
720 | the !, of course) to the underlying operating system. For example, |
|
655 | the !, of course) to the underlying operating system. For example, | |
721 | typing !ls will run 'ls' in the current directory. |
|
656 | typing !ls will run 'ls' in the current directory. | |
722 |
|
657 | |||
723 | Manual capture of command output |
|
658 | Manual capture of command output | |
724 | -------------------------------- |
|
659 | -------------------------------- | |
725 |
|
660 | |||
726 | If the input line begins with two exclamation marks, !!, the command is |
|
661 | If the input line begins with two exclamation marks, !!, the command is | |
727 | executed but its output is captured and returned as a python list, split |
|
662 | executed but its output is captured and returned as a python list, split | |
728 | on newlines. Any output sent by the subprocess to standard error is |
|
663 | on newlines. Any output sent by the subprocess to standard error is | |
729 | printed separately, so that the resulting list only captures standard |
|
664 | printed separately, so that the resulting list only captures standard | |
730 | output. The !! syntax is a shorthand for the %sx magic command. |
|
665 | output. The !! syntax is a shorthand for the %sx magic command. | |
731 |
|
666 | |||
732 | Finally, the %sc magic (short for 'shell capture') is similar to %sx, |
|
667 | Finally, the %sc magic (short for 'shell capture') is similar to %sx, | |
733 | but allowing more fine-grained control of the capture details, and |
|
668 | but allowing more fine-grained control of the capture details, and | |
734 | storing the result directly into a named variable. The direct use of |
|
669 | storing the result directly into a named variable. The direct use of | |
735 | %sc is now deprecated, and you should ise the ``var = !cmd`` syntax |
|
670 | %sc is now deprecated, and you should ise the ``var = !cmd`` syntax | |
736 | instead. |
|
671 | instead. | |
737 |
|
672 | |||
738 | IPython also allows you to expand the value of python variables when |
|
673 | IPython also allows you to expand the value of python variables when | |
739 | making system calls. Any python variable or expression which you prepend |
|
674 | making system calls. Any python variable or expression which you prepend | |
740 | with $ will get expanded before the system call is made:: |
|
675 | with $ will get expanded before the system call is made:: | |
741 |
|
676 | |||
742 | In [1]: pyvar='Hello world' |
|
677 | In [1]: pyvar='Hello world' | |
743 | In [2]: !echo "A python variable: $pyvar" |
|
678 | In [2]: !echo "A python variable: $pyvar" | |
744 | A python variable: Hello world |
|
679 | A python variable: Hello world | |
745 |
|
680 | |||
746 | If you want the shell to actually see a literal $, you need to type it |
|
681 | If you want the shell to actually see a literal $, you need to type it | |
747 | twice:: |
|
682 | twice:: | |
748 |
|
683 | |||
749 | In [3]: !echo "A system variable: $$HOME" |
|
684 | In [3]: !echo "A system variable: $$HOME" | |
750 | A system variable: /home/fperez |
|
685 | A system variable: /home/fperez | |
751 |
|
686 | |||
752 | You can pass arbitrary expressions, though you'll need to delimit them |
|
687 | You can pass arbitrary expressions, though you'll need to delimit them | |
753 | with {} if there is ambiguity as to the extent of the expression:: |
|
688 | with {} if there is ambiguity as to the extent of the expression:: | |
754 |
|
689 | |||
755 | In [5]: x=10 |
|
690 | In [5]: x=10 | |
756 | In [6]: y=20 |
|
691 | In [6]: y=20 | |
757 | In [13]: !echo $x+y |
|
692 | In [13]: !echo $x+y | |
758 | 10+y |
|
693 | 10+y | |
759 | In [7]: !echo ${x+y} |
|
694 | In [7]: !echo ${x+y} | |
760 | 30 |
|
695 | 30 | |
761 |
|
696 | |||
762 | Even object attributes can be expanded:: |
|
697 | Even object attributes can be expanded:: | |
763 |
|
698 | |||
764 | In [12]: !echo $sys.argv |
|
699 | In [12]: !echo $sys.argv | |
765 | [/home/fperez/usr/bin/ipython] |
|
700 | [/home/fperez/usr/bin/ipython] | |
766 |
|
701 | |||
767 |
|
702 | |||
768 | System command aliases |
|
703 | System command aliases | |
769 | ---------------------- |
|
704 | ---------------------- | |
770 |
|
705 | |||
771 | The %alias magic function and the alias option in the ipythonrc |
|
706 | The %alias magic function and the alias option in the ipythonrc | |
772 | configuration file allow you to define magic functions which are in fact |
|
707 | configuration file allow you to define magic functions which are in fact | |
773 | system shell commands. These aliases can have parameters. |
|
708 | system shell commands. These aliases can have parameters. | |
774 |
|
709 | |||
775 | '%alias alias_name cmd' defines 'alias_name' as an alias for 'cmd' |
|
710 | '%alias alias_name cmd' defines 'alias_name' as an alias for 'cmd' | |
776 |
|
711 | |||
777 | Then, typing '%alias_name params' will execute the system command 'cmd |
|
712 | Then, typing '%alias_name params' will execute the system command 'cmd | |
778 | params' (from your underlying operating system). |
|
713 | params' (from your underlying operating system). | |
779 |
|
714 | |||
780 | You can also define aliases with parameters using %s specifiers (one per |
|
715 | You can also define aliases with parameters using %s specifiers (one per | |
781 | parameter). The following example defines the %parts function as an |
|
716 | parameter). The following example defines the %parts function as an | |
782 | alias to the command 'echo first %s second %s' where each %s will be |
|
717 | alias to the command 'echo first %s second %s' where each %s will be | |
783 | replaced by a positional parameter to the call to %parts:: |
|
718 | replaced by a positional parameter to the call to %parts:: | |
784 |
|
719 | |||
785 | In [1]: alias parts echo first %s second %s |
|
720 | In [1]: alias parts echo first %s second %s | |
786 | In [2]: %parts A B |
|
721 | In [2]: %parts A B | |
787 | first A second B |
|
722 | first A second B | |
788 | In [3]: %parts A |
|
723 | In [3]: %parts A | |
789 | Incorrect number of arguments: 2 expected. |
|
724 | Incorrect number of arguments: 2 expected. | |
790 | parts is an alias to: 'echo first %s second %s' |
|
725 | parts is an alias to: 'echo first %s second %s' | |
791 |
|
726 | |||
792 | If called with no parameters, %alias prints the table of currently |
|
727 | If called with no parameters, %alias prints the table of currently | |
793 | defined aliases. |
|
728 | defined aliases. | |
794 |
|
729 | |||
795 | The %rehash/rehashx magics allow you to load your entire $PATH as |
|
730 | The %rehash/rehashx magics allow you to load your entire $PATH as | |
796 | ipython aliases. See their respective docstrings (or sec. 6.2 |
|
731 | ipython aliases. See their respective docstrings (or sec. 6.2 | |
797 | <#sec:magic> for further details). |
|
732 | <#sec:magic> for further details). | |
798 |
|
733 | |||
799 |
|
734 | |||
800 | .. _dreload: |
|
735 | .. _dreload: | |
801 |
|
736 | |||
802 | Recursive reload |
|
737 | Recursive reload | |
803 | ---------------- |
|
738 | ---------------- | |
804 |
|
739 | |||
805 | The dreload function does a recursive reload of a module: changes made |
|
740 | The dreload function does a recursive reload of a module: changes made | |
806 | to the module since you imported will actually be available without |
|
741 | to the module since you imported will actually be available without | |
807 | having to exit. |
|
742 | having to exit. | |
808 |
|
743 | |||
809 |
|
744 | |||
810 | Verbose and colored exception traceback printouts |
|
745 | Verbose and colored exception traceback printouts | |
811 | ------------------------------------------------- |
|
746 | ------------------------------------------------- | |
812 |
|
747 | |||
813 | IPython provides the option to see very detailed exception tracebacks, |
|
748 | IPython provides the option to see very detailed exception tracebacks, | |
814 | which can be especially useful when debugging large programs. You can |
|
749 | which can be especially useful when debugging large programs. You can | |
815 | run any Python file with the %run function to benefit from these |
|
750 | run any Python file with the %run function to benefit from these | |
816 | detailed tracebacks. Furthermore, both normal and verbose tracebacks can |
|
751 | detailed tracebacks. Furthermore, both normal and verbose tracebacks can | |
817 | be colored (if your terminal supports it) which makes them much easier |
|
752 | be colored (if your terminal supports it) which makes them much easier | |
818 | to parse visually. |
|
753 | to parse visually. | |
819 |
|
754 | |||
820 | See the magic xmode and colors functions for details (just type %magic). |
|
755 | See the magic xmode and colors functions for details (just type %magic). | |
821 |
|
756 | |||
822 | These features are basically a terminal version of Ka-Ping Yee's cgitb |
|
757 | These features are basically a terminal version of Ka-Ping Yee's cgitb | |
823 | module, now part of the standard Python library. |
|
758 | module, now part of the standard Python library. | |
824 |
|
759 | |||
825 |
|
760 | |||
826 | .. _input_caching: |
|
761 | .. _input_caching: | |
827 |
|
762 | |||
828 | Input caching system |
|
763 | Input caching system | |
829 | -------------------- |
|
764 | -------------------- | |
830 |
|
765 | |||
831 | IPython offers numbered prompts (In/Out) with input and output caching |
|
766 | IPython offers numbered prompts (In/Out) with input and output caching | |
832 | (also referred to as 'input history'). All input is saved and can be |
|
767 | (also referred to as 'input history'). All input is saved and can be | |
833 | retrieved as variables (besides the usual arrow key recall), in |
|
768 | retrieved as variables (besides the usual arrow key recall), in | |
834 | addition to the %rep magic command that brings a history entry |
|
769 | addition to the %rep magic command that brings a history entry | |
835 | up for editing on the next command line. |
|
770 | up for editing on the next command line. | |
836 |
|
771 | |||
837 | The following GLOBAL variables always exist (so don't overwrite them!): |
|
772 | The following GLOBAL variables always exist (so don't overwrite them!): | |
838 | _i: stores previous input. _ii: next previous. _iii: next-next previous. |
|
773 | _i: stores previous input. _ii: next previous. _iii: next-next previous. | |
839 | _ih : a list of all input _ih[n] is the input from line n and this list |
|
774 | _ih : a list of all input _ih[n] is the input from line n and this list | |
840 | is aliased to the global variable In. If you overwrite In with a |
|
775 | is aliased to the global variable In. If you overwrite In with a | |
841 | variable of your own, you can remake the assignment to the internal list |
|
776 | variable of your own, you can remake the assignment to the internal list | |
842 | with a simple 'In=_ih'. |
|
777 | with a simple 'In=_ih'. | |
843 |
|
778 | |||
844 | Additionally, global variables named _i<n> are dynamically created (<n> |
|
779 | Additionally, global variables named _i<n> are dynamically created (<n> | |
845 | being the prompt counter), such that |
|
780 | being the prompt counter), such that | |
846 | _i<n> == _ih[<n>] == In[<n>]. |
|
781 | _i<n> == _ih[<n>] == In[<n>]. | |
847 |
|
782 | |||
848 | For example, what you typed at prompt 14 is available as _i14, _ih[14] |
|
783 | For example, what you typed at prompt 14 is available as _i14, _ih[14] | |
849 | and In[14]. |
|
784 | and In[14]. | |
850 |
|
785 | |||
851 | This allows you to easily cut and paste multi line interactive prompts |
|
786 | This allows you to easily cut and paste multi line interactive prompts | |
852 | by printing them out: they print like a clean string, without prompt |
|
787 | by printing them out: they print like a clean string, without prompt | |
853 | characters. You can also manipulate them like regular variables (they |
|
788 | characters. You can also manipulate them like regular variables (they | |
854 | are strings), modify or exec them (typing 'exec _i9' will re-execute the |
|
789 | are strings), modify or exec them (typing 'exec _i9' will re-execute the | |
855 | contents of input prompt 9, 'exec In[9:14]+In[18]' will re-execute lines |
|
790 | contents of input prompt 9, 'exec In[9:14]+In[18]' will re-execute lines | |
856 | 9 through 13 and line 18). |
|
791 | 9 through 13 and line 18). | |
857 |
|
792 | |||
858 | You can also re-execute multiple lines of input easily by using the |
|
793 | You can also re-execute multiple lines of input easily by using the | |
859 | magic %macro function (which automates the process and allows |
|
794 | magic %macro function (which automates the process and allows | |
860 | re-execution without having to type 'exec' every time). The macro system |
|
795 | re-execution without having to type 'exec' every time). The macro system | |
861 | also allows you to re-execute previous lines which include magic |
|
796 | also allows you to re-execute previous lines which include magic | |
862 | function calls (which require special processing). Type %macro? or see |
|
797 | function calls (which require special processing). Type %macro? or see | |
863 | sec. 6.2 <#sec:magic> for more details on the macro system. |
|
798 | sec. 6.2 <#sec:magic> for more details on the macro system. | |
864 |
|
799 | |||
865 | A history function %hist allows you to see any part of your input |
|
800 | A history function %hist allows you to see any part of your input | |
866 | history by printing a range of the _i variables. |
|
801 | history by printing a range of the _i variables. | |
867 |
|
802 | |||
868 | You can also search ('grep') through your history by typing |
|
803 | You can also search ('grep') through your history by typing | |
869 | '%hist -g somestring'. This also searches through the so called *shadow history*, |
|
804 | '%hist -g somestring'. This also searches through the so called *shadow history*, | |
870 | which remembers all the commands (apart from multiline code blocks) |
|
805 | which remembers all the commands (apart from multiline code blocks) | |
871 | you have ever entered. Handy for searching for svn/bzr URL's, IP adrresses |
|
806 | you have ever entered. Handy for searching for svn/bzr URL's, IP adrresses | |
872 | etc. You can bring shadow history entries listed by '%hist -g' up for editing |
|
807 | etc. You can bring shadow history entries listed by '%hist -g' up for editing | |
873 | (or re-execution by just pressing ENTER) with %rep command. Shadow history |
|
808 | (or re-execution by just pressing ENTER) with %rep command. Shadow history | |
874 | entries are not available as _iNUMBER variables, and they are identified by |
|
809 | entries are not available as _iNUMBER variables, and they are identified by | |
875 | the '0' prefix in %hist -g output. That is, history entry 12 is a normal |
|
810 | the '0' prefix in %hist -g output. That is, history entry 12 is a normal | |
876 | history entry, but 0231 is a shadow history entry. |
|
811 | history entry, but 0231 is a shadow history entry. | |
877 |
|
812 | |||
878 | Shadow history was added because the readline history is inherently very |
|
813 | Shadow history was added because the readline history is inherently very | |
879 | unsafe - if you have multiple IPython sessions open, the last session |
|
814 | unsafe - if you have multiple IPython sessions open, the last session | |
880 | to close will overwrite the history of previountly closed session. Likewise, |
|
815 | to close will overwrite the history of previountly closed session. Likewise, | |
881 | if a crash occurs, history is never saved, whereas shadow history entries |
|
816 | if a crash occurs, history is never saved, whereas shadow history entries | |
882 | are added after entering every command (so a command executed |
|
817 | are added after entering every command (so a command executed | |
883 | in another IPython session is immediately available in other IPython |
|
818 | in another IPython session is immediately available in other IPython | |
884 | sessions that are open). |
|
819 | sessions that are open). | |
885 |
|
820 | |||
886 | To conserve space, a command can exist in shadow history only once - it doesn't |
|
821 | To conserve space, a command can exist in shadow history only once - it doesn't | |
887 | make sense to store a common line like "cd .." a thousand times. The idea is |
|
822 | make sense to store a common line like "cd .." a thousand times. The idea is | |
888 | mainly to provide a reliable place where valuable, hard-to-remember commands can |
|
823 | mainly to provide a reliable place where valuable, hard-to-remember commands can | |
889 | always be retrieved, as opposed to providing an exact sequence of commands |
|
824 | always be retrieved, as opposed to providing an exact sequence of commands | |
890 | you have entered in actual order. |
|
825 | you have entered in actual order. | |
891 |
|
826 | |||
892 | Because shadow history has all the commands you have ever executed, |
|
827 | Because shadow history has all the commands you have ever executed, | |
893 | time taken by %hist -g will increase oven time. If it ever starts to take |
|
828 | time taken by %hist -g will increase oven time. If it ever starts to take | |
894 | too long (or it ends up containing sensitive information like passwords), |
|
829 | too long (or it ends up containing sensitive information like passwords), | |
895 | clear the shadow history by `%clear shadow_nuke`. |
|
830 | clear the shadow history by `%clear shadow_nuke`. | |
896 |
|
831 | |||
897 | Time taken to add entries to shadow history should be negligible, but |
|
832 | Time taken to add entries to shadow history should be negligible, but | |
898 | in any case, if you start noticing performance degradation after using |
|
833 | in any case, if you start noticing performance degradation after using | |
899 | IPython for a long time (or running a script that floods the shadow history!), |
|
834 | IPython for a long time (or running a script that floods the shadow history!), | |
900 | you can 'compress' the shadow history by executing |
|
835 | you can 'compress' the shadow history by executing | |
901 | `%clear shadow_compress`. In practice, this should never be necessary |
|
836 | `%clear shadow_compress`. In practice, this should never be necessary | |
902 | in normal use. |
|
837 | in normal use. | |
903 |
|
838 | |||
904 | .. _output_caching: |
|
839 | .. _output_caching: | |
905 |
|
840 | |||
906 | Output caching system |
|
841 | Output caching system | |
907 | --------------------- |
|
842 | --------------------- | |
908 |
|
843 | |||
909 | For output that is returned from actions, a system similar to the input |
|
844 | For output that is returned from actions, a system similar to the input | |
910 | cache exists but using _ instead of _i. Only actions that produce a |
|
845 | cache exists but using _ instead of _i. Only actions that produce a | |
911 | result (NOT assignments, for example) are cached. If you are familiar |
|
846 | result (NOT assignments, for example) are cached. If you are familiar | |
912 | with Mathematica, IPython's _ variables behave exactly like |
|
847 | with Mathematica, IPython's _ variables behave exactly like | |
913 | Mathematica's % variables. |
|
848 | Mathematica's % variables. | |
914 |
|
849 | |||
915 | The following GLOBAL variables always exist (so don't overwrite them!): |
|
850 | The following GLOBAL variables always exist (so don't overwrite them!): | |
916 |
|
851 | |||
917 | * [_] (a single underscore) : stores previous output, like Python's |
|
852 | * [_] (a single underscore) : stores previous output, like Python's | |
918 | default interpreter. |
|
853 | default interpreter. | |
919 | * [__] (two underscores): next previous. |
|
854 | * [__] (two underscores): next previous. | |
920 | * [___] (three underscores): next-next previous. |
|
855 | * [___] (three underscores): next-next previous. | |
921 |
|
856 | |||
922 | Additionally, global variables named _<n> are dynamically created (<n> |
|
857 | Additionally, global variables named _<n> are dynamically created (<n> | |
923 | being the prompt counter), such that the result of output <n> is always |
|
858 | being the prompt counter), such that the result of output <n> is always | |
924 | available as _<n> (don't use the angle brackets, just the number, e.g. |
|
859 | available as _<n> (don't use the angle brackets, just the number, e.g. | |
925 | _21). |
|
860 | _21). | |
926 |
|
861 | |||
927 | These global variables are all stored in a global dictionary (not a |
|
862 | These global variables are all stored in a global dictionary (not a | |
928 | list, since it only has entries for lines which returned a result) |
|
863 | list, since it only has entries for lines which returned a result) | |
929 | available under the names _oh and Out (similar to _ih and In). So the |
|
864 | available under the names _oh and Out (similar to _ih and In). So the | |
930 | output from line 12 can be obtained as _12, Out[12] or _oh[12]. If you |
|
865 | output from line 12 can be obtained as _12, Out[12] or _oh[12]. If you | |
931 | accidentally overwrite the Out variable you can recover it by typing |
|
866 | accidentally overwrite the Out variable you can recover it by typing | |
932 | 'Out=_oh' at the prompt. |
|
867 | 'Out=_oh' at the prompt. | |
933 |
|
868 | |||
934 | This system obviously can potentially put heavy memory demands on your |
|
869 | This system obviously can potentially put heavy memory demands on your | |
935 | system, since it prevents Python's garbage collector from removing any |
|
870 | system, since it prevents Python's garbage collector from removing any | |
936 | previously computed results. You can control how many results are kept |
|
871 | previously computed results. You can control how many results are kept | |
937 | in memory with the option (at the command line or in your ipythonrc |
|
872 | in memory with the option (at the command line or in your ipythonrc | |
938 | file) cache_size. If you set it to 0, the whole system is completely |
|
873 | file) cache_size. If you set it to 0, the whole system is completely | |
939 | disabled and the prompts revert to the classic '>>>' of normal Python. |
|
874 | disabled and the prompts revert to the classic '>>>' of normal Python. | |
940 |
|
875 | |||
941 |
|
876 | |||
942 | Directory history |
|
877 | Directory history | |
943 | ----------------- |
|
878 | ----------------- | |
944 |
|
879 | |||
945 | Your history of visited directories is kept in the global list _dh, and |
|
880 | Your history of visited directories is kept in the global list _dh, and | |
946 | the magic %cd command can be used to go to any entry in that list. The |
|
881 | the magic %cd command can be used to go to any entry in that list. The | |
947 | %dhist command allows you to view this history. Do ``cd -<TAB`` to |
|
882 | %dhist command allows you to view this history. Do ``cd -<TAB`` to | |
948 | conventiently view the directory history. |
|
883 | conventiently view the directory history. | |
949 |
|
884 | |||
950 |
|
885 | |||
951 | Automatic parentheses and quotes |
|
886 | Automatic parentheses and quotes | |
952 | -------------------------------- |
|
887 | -------------------------------- | |
953 |
|
888 | |||
954 | These features were adapted from Nathan Gray's LazyPython. They are |
|
889 | These features were adapted from Nathan Gray's LazyPython. They are | |
955 | meant to allow less typing for common situations. |
|
890 | meant to allow less typing for common situations. | |
956 |
|
891 | |||
957 |
|
892 | |||
958 | Automatic parentheses |
|
893 | Automatic parentheses | |
959 | --------------------- |
|
894 | --------------------- | |
960 |
|
895 | |||
961 | Callable objects (i.e. functions, methods, etc) can be invoked like this |
|
896 | Callable objects (i.e. functions, methods, etc) can be invoked like this | |
962 | (notice the commas between the arguments):: |
|
897 | (notice the commas between the arguments):: | |
963 |
|
898 | |||
964 | >>> callable_ob arg1, arg2, arg3 |
|
899 | >>> callable_ob arg1, arg2, arg3 | |
965 |
|
900 | |||
966 | and the input will be translated to this:: |
|
901 | and the input will be translated to this:: | |
967 |
|
902 | |||
968 | -> callable_ob(arg1, arg2, arg3) |
|
903 | -> callable_ob(arg1, arg2, arg3) | |
969 |
|
904 | |||
970 | You can force automatic parentheses by using '/' as the first character |
|
905 | You can force automatic parentheses by using '/' as the first character | |
971 | of a line. For example:: |
|
906 | of a line. For example:: | |
972 |
|
907 | |||
973 | >>> /globals # becomes 'globals()' |
|
908 | >>> /globals # becomes 'globals()' | |
974 |
|
909 | |||
975 | Note that the '/' MUST be the first character on the line! This won't work:: |
|
910 | Note that the '/' MUST be the first character on the line! This won't work:: | |
976 |
|
911 | |||
977 | >>> print /globals # syntax error |
|
912 | >>> print /globals # syntax error | |
978 |
|
913 | |||
979 | In most cases the automatic algorithm should work, so you should rarely |
|
914 | In most cases the automatic algorithm should work, so you should rarely | |
980 | need to explicitly invoke /. One notable exception is if you are trying |
|
915 | need to explicitly invoke /. One notable exception is if you are trying | |
981 | to call a function with a list of tuples as arguments (the parenthesis |
|
916 | to call a function with a list of tuples as arguments (the parenthesis | |
982 | will confuse IPython):: |
|
917 | will confuse IPython):: | |
983 |
|
918 | |||
984 | In [1]: zip (1,2,3),(4,5,6) # won't work |
|
919 | In [1]: zip (1,2,3),(4,5,6) # won't work | |
985 |
|
920 | |||
986 | but this will work:: |
|
921 | but this will work:: | |
987 |
|
922 | |||
988 | In [2]: /zip (1,2,3),(4,5,6) |
|
923 | In [2]: /zip (1,2,3),(4,5,6) | |
989 | ---> zip ((1,2,3),(4,5,6)) |
|
924 | ---> zip ((1,2,3),(4,5,6)) | |
990 | Out[2]= [(1, 4), (2, 5), (3, 6)] |
|
925 | Out[2]= [(1, 4), (2, 5), (3, 6)] | |
991 |
|
926 | |||
992 | IPython tells you that it has altered your command line by displaying |
|
927 | IPython tells you that it has altered your command line by displaying | |
993 | the new command line preceded by ->. e.g.:: |
|
928 | the new command line preceded by ->. e.g.:: | |
994 |
|
929 | |||
995 | In [18]: callable list |
|
930 | In [18]: callable list | |
996 | ----> callable (list) |
|
931 | ----> callable (list) | |
997 |
|
932 | |||
998 |
|
933 | |||
999 | Automatic quoting |
|
934 | Automatic quoting | |
1000 | ----------------- |
|
935 | ----------------- | |
1001 |
|
936 | |||
1002 | You can force automatic quoting of a function's arguments by using ',' |
|
937 | You can force automatic quoting of a function's arguments by using ',' | |
1003 | or ';' as the first character of a line. For example:: |
|
938 | or ';' as the first character of a line. For example:: | |
1004 |
|
939 | |||
1005 | >>> ,my_function /home/me # becomes my_function("/home/me") |
|
940 | >>> ,my_function /home/me # becomes my_function("/home/me") | |
1006 |
|
941 | |||
1007 | If you use ';' instead, the whole argument is quoted as a single string |
|
942 | If you use ';' instead, the whole argument is quoted as a single string | |
1008 | (while ',' splits on whitespace):: |
|
943 | (while ',' splits on whitespace):: | |
1009 |
|
944 | |||
1010 | >>> ,my_function a b c # becomes my_function("a","b","c") |
|
945 | >>> ,my_function a b c # becomes my_function("a","b","c") | |
1011 |
|
946 | |||
1012 | >>> ;my_function a b c # becomes my_function("a b c") |
|
947 | >>> ;my_function a b c # becomes my_function("a b c") | |
1013 |
|
948 | |||
1014 | Note that the ',' or ';' MUST be the first character on the line! This |
|
949 | Note that the ',' or ';' MUST be the first character on the line! This | |
1015 | won't work:: |
|
950 | won't work:: | |
1016 |
|
951 | |||
1017 | >>> x = ,my_function /home/me # syntax error |
|
952 | >>> x = ,my_function /home/me # syntax error | |
1018 |
|
953 | |||
1019 | IPython as your default Python environment |
|
954 | IPython as your default Python environment | |
1020 | ========================================== |
|
955 | ========================================== | |
1021 |
|
956 | |||
1022 | Python honors the environment variable PYTHONSTARTUP and will execute at |
|
957 | Python honors the environment variable PYTHONSTARTUP and will execute at | |
1023 | startup the file referenced by this variable. If you put at the end of |
|
958 | startup the file referenced by this variable. If you put at the end of | |
1024 | this file the following two lines of code:: |
|
959 | this file the following two lines of code:: | |
1025 |
|
960 | |||
1026 | import IPython |
|
961 | import IPython | |
1027 | IPython.Shell.IPShell().mainloop(sys_exit=1) |
|
962 | IPython.Shell.IPShell().mainloop(sys_exit=1) | |
1028 |
|
963 | |||
1029 | then IPython will be your working environment anytime you start Python. |
|
964 | then IPython will be your working environment anytime you start Python. | |
1030 | The sys_exit=1 is needed to have IPython issue a call to sys.exit() when |
|
965 | The sys_exit=1 is needed to have IPython issue a call to sys.exit() when | |
1031 | it finishes, otherwise you'll be back at the normal Python '>>>' |
|
966 | it finishes, otherwise you'll be back at the normal Python '>>>' | |
1032 | prompt. |
|
967 | prompt. | |
1033 |
|
968 | |||
1034 | This is probably useful to developers who manage multiple Python |
|
969 | This is probably useful to developers who manage multiple Python | |
1035 | versions and don't want to have correspondingly multiple IPython |
|
970 | versions and don't want to have correspondingly multiple IPython | |
1036 | versions. Note that in this mode, there is no way to pass IPython any |
|
971 | versions. Note that in this mode, there is no way to pass IPython any | |
1037 | command-line options, as those are trapped first by Python itself. |
|
972 | command-line options, as those are trapped first by Python itself. | |
1038 |
|
973 | |||
1039 | .. _Embedding: |
|
974 | .. _Embedding: | |
1040 |
|
975 | |||
1041 | Embedding IPython |
|
976 | Embedding IPython | |
1042 | ================= |
|
977 | ================= | |
1043 |
|
978 | |||
1044 | It is possible to start an IPython instance inside your own Python |
|
979 | It is possible to start an IPython instance inside your own Python | |
1045 | programs. This allows you to evaluate dynamically the state of your |
|
980 | programs. This allows you to evaluate dynamically the state of your | |
1046 | code, operate with your variables, analyze them, etc. Note however that |
|
981 | code, operate with your variables, analyze them, etc. Note however that | |
1047 | any changes you make to values while in the shell do not propagate back |
|
982 | any changes you make to values while in the shell do not propagate back | |
1048 | to the running code, so it is safe to modify your values because you |
|
983 | to the running code, so it is safe to modify your values because you | |
1049 | won't break your code in bizarre ways by doing so. |
|
984 | won't break your code in bizarre ways by doing so. | |
1050 |
|
985 | |||
1051 | This feature allows you to easily have a fully functional python |
|
986 | This feature allows you to easily have a fully functional python | |
1052 | environment for doing object introspection anywhere in your code with a |
|
987 | environment for doing object introspection anywhere in your code with a | |
1053 | simple function call. In some cases a simple print statement is enough, |
|
988 | simple function call. In some cases a simple print statement is enough, | |
1054 | but if you need to do more detailed analysis of a code fragment this |
|
989 | but if you need to do more detailed analysis of a code fragment this | |
1055 | feature can be very valuable. |
|
990 | feature can be very valuable. | |
1056 |
|
991 | |||
1057 | It can also be useful in scientific computing situations where it is |
|
992 | It can also be useful in scientific computing situations where it is | |
1058 | common to need to do some automatic, computationally intensive part and |
|
993 | common to need to do some automatic, computationally intensive part and | |
1059 | then stop to look at data, plots, etc. |
|
994 | then stop to look at data, plots, etc. | |
1060 | Opening an IPython instance will give you full access to your data and |
|
995 | Opening an IPython instance will give you full access to your data and | |
1061 | functions, and you can resume program execution once you are done with |
|
996 | functions, and you can resume program execution once you are done with | |
1062 | the interactive part (perhaps to stop again later, as many times as |
|
997 | the interactive part (perhaps to stop again later, as many times as | |
1063 | needed). |
|
998 | needed). | |
1064 |
|
999 | |||
1065 | The following code snippet is the bare minimum you need to include in |
|
1000 | The following code snippet is the bare minimum you need to include in | |
1066 | your Python programs for this to work (detailed examples follow later):: |
|
1001 | your Python programs for this to work (detailed examples follow later):: | |
1067 |
|
1002 | |||
1068 | from IPython.Shell import IPShellEmbed |
|
1003 | from IPython.Shell import IPShellEmbed | |
1069 |
|
1004 | |||
1070 | ipshell = IPShellEmbed() |
|
1005 | ipshell = IPShellEmbed() | |
1071 |
|
1006 | |||
1072 | ipshell() # this call anywhere in your program will start IPython |
|
1007 | ipshell() # this call anywhere in your program will start IPython | |
1073 |
|
1008 | |||
1074 | You can run embedded instances even in code which is itself being run at |
|
1009 | You can run embedded instances even in code which is itself being run at | |
1075 | the IPython interactive prompt with '%run <filename>'. Since it's easy |
|
1010 | the IPython interactive prompt with '%run <filename>'. Since it's easy | |
1076 | to get lost as to where you are (in your top-level IPython or in your |
|
1011 | to get lost as to where you are (in your top-level IPython or in your | |
1077 | embedded one), it's a good idea in such cases to set the in/out prompts |
|
1012 | embedded one), it's a good idea in such cases to set the in/out prompts | |
1078 | to something different for the embedded instances. The code examples |
|
1013 | to something different for the embedded instances. The code examples | |
1079 | below illustrate this. |
|
1014 | below illustrate this. | |
1080 |
|
1015 | |||
1081 | You can also have multiple IPython instances in your program and open |
|
1016 | You can also have multiple IPython instances in your program and open | |
1082 | them separately, for example with different options for data |
|
1017 | them separately, for example with different options for data | |
1083 | presentation. If you close and open the same instance multiple times, |
|
1018 | presentation. If you close and open the same instance multiple times, | |
1084 | its prompt counters simply continue from each execution to the next. |
|
1019 | its prompt counters simply continue from each execution to the next. | |
1085 |
|
1020 | |||
1086 | Please look at the docstrings in the Shell.py module for more details on |
|
1021 | Please look at the docstrings in the Shell.py module for more details on | |
1087 | the use of this system. |
|
1022 | the use of this system. | |
1088 |
|
1023 | |||
1089 | The following sample file illustrating how to use the embedding |
|
1024 | The following sample file illustrating how to use the embedding | |
1090 | functionality is provided in the examples directory as example-embed.py. |
|
1025 | functionality is provided in the examples directory as example-embed.py. | |
1091 | It should be fairly self-explanatory:: |
|
1026 | It should be fairly self-explanatory:: | |
1092 |
|
1027 | |||
1093 |
|
1028 | |||
1094 | #!/usr/bin/env python |
|
1029 | #!/usr/bin/env python | |
1095 |
|
1030 | |||
1096 | """An example of how to embed an IPython shell into a running program. |
|
1031 | """An example of how to embed an IPython shell into a running program. | |
1097 |
|
1032 | |||
1098 | Please see the documentation in the IPython.Shell module for more details. |
|
1033 | Please see the documentation in the IPython.Shell module for more details. | |
1099 |
|
1034 | |||
1100 | The accompanying file example-embed-short.py has quick code fragments for |
|
1035 | The accompanying file example-embed-short.py has quick code fragments for | |
1101 | embedding which you can cut and paste in your code once you understand how |
|
1036 | embedding which you can cut and paste in your code once you understand how | |
1102 | things work. |
|
1037 | things work. | |
1103 |
|
1038 | |||
1104 | The code in this file is deliberately extra-verbose, meant for learning.""" |
|
1039 | The code in this file is deliberately extra-verbose, meant for learning.""" | |
1105 |
|
1040 | |||
1106 | # The basics to get you going: |
|
1041 | # The basics to get you going: | |
1107 |
|
1042 | |||
1108 | # IPython sets the __IPYTHON__ variable so you can know if you have nested |
|
1043 | # IPython sets the __IPYTHON__ variable so you can know if you have nested | |
1109 | # copies running. |
|
1044 | # copies running. | |
1110 |
|
1045 | |||
1111 | # Try running this code both at the command line and from inside IPython (with |
|
1046 | # Try running this code both at the command line and from inside IPython (with | |
1112 | # %run example-embed.py) |
|
1047 | # %run example-embed.py) | |
1113 | try: |
|
1048 | try: | |
1114 | __IPYTHON__ |
|
1049 | __IPYTHON__ | |
1115 | except NameError: |
|
1050 | except NameError: | |
1116 | nested = 0 |
|
1051 | nested = 0 | |
1117 | args = [''] |
|
1052 | args = [''] | |
1118 | else: |
|
1053 | else: | |
1119 | print "Running nested copies of IPython." |
|
1054 | print "Running nested copies of IPython." | |
1120 | print "The prompts for the nested copy have been modified" |
|
1055 | print "The prompts for the nested copy have been modified" | |
1121 | nested = 1 |
|
1056 | nested = 1 | |
1122 | # what the embedded instance will see as sys.argv: |
|
1057 | # what the embedded instance will see as sys.argv: | |
1123 | args = ['-pi1','In <\\#>: ','-pi2',' .\\D.: ', |
|
1058 | args = ['-pi1','In <\\#>: ','-pi2',' .\\D.: ', | |
1124 | '-po','Out<\\#>: ','-nosep'] |
|
1059 | '-po','Out<\\#>: ','-nosep'] | |
1125 |
|
1060 | |||
1126 | # First import the embeddable shell class |
|
1061 | # First import the embeddable shell class | |
1127 | from IPython.Shell import IPShellEmbed |
|
1062 | from IPython.Shell import IPShellEmbed | |
1128 |
|
1063 | |||
1129 | # Now create an instance of the embeddable shell. The first argument is a |
|
1064 | # Now create an instance of the embeddable shell. The first argument is a | |
1130 | # string with options exactly as you would type them if you were starting |
|
1065 | # string with options exactly as you would type them if you were starting | |
1131 | # IPython at the system command line. Any parameters you want to define for |
|
1066 | # IPython at the system command line. Any parameters you want to define for | |
1132 | # configuration can thus be specified here. |
|
1067 | # configuration can thus be specified here. | |
1133 | ipshell = IPShellEmbed(args, |
|
1068 | ipshell = IPShellEmbed(args, | |
1134 | banner = 'Dropping into IPython', |
|
1069 | banner = 'Dropping into IPython', | |
1135 | exit_msg = 'Leaving Interpreter, back to program.') |
|
1070 | exit_msg = 'Leaving Interpreter, back to program.') | |
1136 |
|
1071 | |||
1137 | # Make a second instance, you can have as many as you want. |
|
1072 | # Make a second instance, you can have as many as you want. | |
1138 | if nested: |
|
1073 | if nested: | |
1139 | args[1] = 'In2<\\#>' |
|
1074 | args[1] = 'In2<\\#>' | |
1140 | else: |
|
1075 | else: | |
1141 | args = ['-pi1','In2<\\#>: ','-pi2',' .\\D.: ', |
|
1076 | args = ['-pi1','In2<\\#>: ','-pi2',' .\\D.: ', | |
1142 | '-po','Out<\\#>: ','-nosep'] |
|
1077 | '-po','Out<\\#>: ','-nosep'] | |
1143 | ipshell2 = IPShellEmbed(args,banner = 'Second IPython instance.') |
|
1078 | ipshell2 = IPShellEmbed(args,banner = 'Second IPython instance.') | |
1144 |
|
1079 | |||
1145 | print '\nHello. This is printed from the main controller program.\n' |
|
1080 | print '\nHello. This is printed from the main controller program.\n' | |
1146 |
|
1081 | |||
1147 | # You can then call ipshell() anywhere you need it (with an optional |
|
1082 | # You can then call ipshell() anywhere you need it (with an optional | |
1148 | # message): |
|
1083 | # message): | |
1149 | ipshell('***Called from top level. ' |
|
1084 | ipshell('***Called from top level. ' | |
1150 | 'Hit Ctrl-D to exit interpreter and continue program.\n' |
|
1085 | 'Hit Ctrl-D to exit interpreter and continue program.\n' | |
1151 | 'Note that if you use %kill_embedded, you can fully deactivate\n' |
|
1086 | 'Note that if you use %kill_embedded, you can fully deactivate\n' | |
1152 | 'This embedded instance so it will never turn on again') |
|
1087 | 'This embedded instance so it will never turn on again') | |
1153 |
|
1088 | |||
1154 | print '\nBack in caller program, moving along...\n' |
|
1089 | print '\nBack in caller program, moving along...\n' | |
1155 |
|
1090 | |||
1156 | #--------------------------------------------------------------------------- |
|
1091 | #--------------------------------------------------------------------------- | |
1157 | # More details: |
|
1092 | # More details: | |
1158 |
|
1093 | |||
1159 | # IPShellEmbed instances don't print the standard system banner and |
|
1094 | # IPShellEmbed instances don't print the standard system banner and | |
1160 | # messages. The IPython banner (which actually may contain initialization |
|
1095 | # messages. The IPython banner (which actually may contain initialization | |
1161 | # messages) is available as <instance>.IP.BANNER in case you want it. |
|
1096 | # messages) is available as <instance>.IP.BANNER in case you want it. | |
1162 |
|
1097 | |||
1163 | # IPShellEmbed instances print the following information everytime they |
|
1098 | # IPShellEmbed instances print the following information everytime they | |
1164 | # start: |
|
1099 | # start: | |
1165 |
|
1100 | |||
1166 | # - A global startup banner. |
|
1101 | # - A global startup banner. | |
1167 |
|
1102 | |||
1168 | # - A call-specific header string, which you can use to indicate where in the |
|
1103 | # - A call-specific header string, which you can use to indicate where in the | |
1169 | # execution flow the shell is starting. |
|
1104 | # execution flow the shell is starting. | |
1170 |
|
1105 | |||
1171 | # They also print an exit message every time they exit. |
|
1106 | # They also print an exit message every time they exit. | |
1172 |
|
1107 | |||
1173 | # Both the startup banner and the exit message default to None, and can be set |
|
1108 | # Both the startup banner and the exit message default to None, and can be set | |
1174 | # either at the instance constructor or at any other time with the |
|
1109 | # either at the instance constructor or at any other time with the | |
1175 | # set_banner() and set_exit_msg() methods. |
|
1110 | # set_banner() and set_exit_msg() methods. | |
1176 |
|
1111 | |||
1177 | # The shell instance can be also put in 'dummy' mode globally or on a per-call |
|
1112 | # The shell instance can be also put in 'dummy' mode globally or on a per-call | |
1178 | # basis. This gives you fine control for debugging without having to change |
|
1113 | # basis. This gives you fine control for debugging without having to change | |
1179 | # code all over the place. |
|
1114 | # code all over the place. | |
1180 |
|
1115 | |||
1181 | # The code below illustrates all this. |
|
1116 | # The code below illustrates all this. | |
1182 |
|
1117 | |||
1183 |
|
1118 | |||
1184 | # This is how the global banner and exit_msg can be reset at any point |
|
1119 | # This is how the global banner and exit_msg can be reset at any point | |
1185 | ipshell.set_banner('Entering interpreter - New Banner') |
|
1120 | ipshell.set_banner('Entering interpreter - New Banner') | |
1186 | ipshell.set_exit_msg('Leaving interpreter - New exit_msg') |
|
1121 | ipshell.set_exit_msg('Leaving interpreter - New exit_msg') | |
1187 |
|
1122 | |||
1188 | def foo(m): |
|
1123 | def foo(m): | |
1189 | s = 'spam' |
|
1124 | s = 'spam' | |
1190 | ipshell('***In foo(). Try @whos, or print s or m:') |
|
1125 | ipshell('***In foo(). Try @whos, or print s or m:') | |
1191 | print 'foo says m = ',m |
|
1126 | print 'foo says m = ',m | |
1192 |
|
1127 | |||
1193 | def bar(n): |
|
1128 | def bar(n): | |
1194 | s = 'eggs' |
|
1129 | s = 'eggs' | |
1195 | ipshell('***In bar(). Try @whos, or print s or n:') |
|
1130 | ipshell('***In bar(). Try @whos, or print s or n:') | |
1196 | print 'bar says n = ',n |
|
1131 | print 'bar says n = ',n | |
1197 |
|
1132 | |||
1198 | # Some calls to the above functions which will trigger IPython: |
|
1133 | # Some calls to the above functions which will trigger IPython: | |
1199 | print 'Main program calling foo("eggs")\n' |
|
1134 | print 'Main program calling foo("eggs")\n' | |
1200 | foo('eggs') |
|
1135 | foo('eggs') | |
1201 |
|
1136 | |||
1202 | # The shell can be put in 'dummy' mode where calls to it silently return. This |
|
1137 | # The shell can be put in 'dummy' mode where calls to it silently return. This | |
1203 | # allows you, for example, to globally turn off debugging for a program with a |
|
1138 | # allows you, for example, to globally turn off debugging for a program with a | |
1204 | # single call. |
|
1139 | # single call. | |
1205 | ipshell.set_dummy_mode(1) |
|
1140 | ipshell.set_dummy_mode(1) | |
1206 | print '\nTrying to call IPython which is now "dummy":' |
|
1141 | print '\nTrying to call IPython which is now "dummy":' | |
1207 | ipshell() |
|
1142 | ipshell() | |
1208 | print 'Nothing happened...' |
|
1143 | print 'Nothing happened...' | |
1209 | # The global 'dummy' mode can still be overridden for a single call |
|
1144 | # The global 'dummy' mode can still be overridden for a single call | |
1210 | print '\nOverriding dummy mode manually:' |
|
1145 | print '\nOverriding dummy mode manually:' | |
1211 | ipshell(dummy=0) |
|
1146 | ipshell(dummy=0) | |
1212 |
|
1147 | |||
1213 | # Reactivate the IPython shell |
|
1148 | # Reactivate the IPython shell | |
1214 | ipshell.set_dummy_mode(0) |
|
1149 | ipshell.set_dummy_mode(0) | |
1215 |
|
1150 | |||
1216 | print 'You can even have multiple embedded instances:' |
|
1151 | print 'You can even have multiple embedded instances:' | |
1217 | ipshell2() |
|
1152 | ipshell2() | |
1218 |
|
1153 | |||
1219 | print '\nMain program calling bar("spam")\n' |
|
1154 | print '\nMain program calling bar("spam")\n' | |
1220 | bar('spam') |
|
1155 | bar('spam') | |
1221 |
|
1156 | |||
1222 | print 'Main program finished. Bye!' |
|
1157 | print 'Main program finished. Bye!' | |
1223 |
|
1158 | |||
1224 | #********************** End of file <example-embed.py> *********************** |
|
1159 | #********************** End of file <example-embed.py> *********************** | |
1225 |
|
1160 | |||
1226 | Once you understand how the system functions, you can use the following |
|
1161 | Once you understand how the system functions, you can use the following | |
1227 | code fragments in your programs which are ready for cut and paste:: |
|
1162 | code fragments in your programs which are ready for cut and paste:: | |
1228 |
|
1163 | |||
1229 |
|
1164 | |||
1230 | """Quick code snippets for embedding IPython into other programs. |
|
1165 | """Quick code snippets for embedding IPython into other programs. | |
1231 |
|
1166 | |||
1232 | See example-embed.py for full details, this file has the bare minimum code for |
|
1167 | See example-embed.py for full details, this file has the bare minimum code for | |
1233 | cut and paste use once you understand how to use the system.""" |
|
1168 | cut and paste use once you understand how to use the system.""" | |
1234 |
|
1169 | |||
1235 | #--------------------------------------------------------------------------- |
|
1170 | #--------------------------------------------------------------------------- | |
1236 | # This code loads IPython but modifies a few things if it detects it's running |
|
1171 | # This code loads IPython but modifies a few things if it detects it's running | |
1237 | # embedded in another IPython session (helps avoid confusion) |
|
1172 | # embedded in another IPython session (helps avoid confusion) | |
1238 |
|
1173 | |||
1239 | try: |
|
1174 | try: | |
1240 | __IPYTHON__ |
|
1175 | __IPYTHON__ | |
1241 | except NameError: |
|
1176 | except NameError: | |
1242 | argv = [''] |
|
1177 | argv = [''] | |
1243 | banner = exit_msg = '' |
|
1178 | banner = exit_msg = '' | |
1244 | else: |
|
1179 | else: | |
1245 | # Command-line options for IPython (a list like sys.argv) |
|
1180 | # Command-line options for IPython (a list like sys.argv) | |
1246 | argv = ['-pi1','In <\\#>:','-pi2',' .\\D.:','-po','Out<\\#>:'] |
|
1181 | argv = ['-pi1','In <\\#>:','-pi2',' .\\D.:','-po','Out<\\#>:'] | |
1247 | banner = '*** Nested interpreter ***' |
|
1182 | banner = '*** Nested interpreter ***' | |
1248 | exit_msg = '*** Back in main IPython ***' |
|
1183 | exit_msg = '*** Back in main IPython ***' | |
1249 |
|
1184 | |||
1250 | # First import the embeddable shell class |
|
1185 | # First import the embeddable shell class | |
1251 | from IPython.Shell import IPShellEmbed |
|
1186 | from IPython.Shell import IPShellEmbed | |
1252 | # Now create the IPython shell instance. Put ipshell() anywhere in your code |
|
1187 | # Now create the IPython shell instance. Put ipshell() anywhere in your code | |
1253 | # where you want it to open. |
|
1188 | # where you want it to open. | |
1254 | ipshell = IPShellEmbed(argv,banner=banner,exit_msg=exit_msg) |
|
1189 | ipshell = IPShellEmbed(argv,banner=banner,exit_msg=exit_msg) | |
1255 |
|
1190 | |||
1256 | #--------------------------------------------------------------------------- |
|
1191 | #--------------------------------------------------------------------------- | |
1257 | # This code will load an embeddable IPython shell always with no changes for |
|
1192 | # This code will load an embeddable IPython shell always with no changes for | |
1258 | # nested embededings. |
|
1193 | # nested embededings. | |
1259 |
|
1194 | |||
1260 | from IPython.Shell import IPShellEmbed |
|
1195 | from IPython.Shell import IPShellEmbed | |
1261 | ipshell = IPShellEmbed() |
|
1196 | ipshell = IPShellEmbed() | |
1262 | # Now ipshell() will open IPython anywhere in the code. |
|
1197 | # Now ipshell() will open IPython anywhere in the code. | |
1263 |
|
1198 | |||
1264 | #--------------------------------------------------------------------------- |
|
1199 | #--------------------------------------------------------------------------- | |
1265 | # This code loads an embeddable shell only if NOT running inside |
|
1200 | # This code loads an embeddable shell only if NOT running inside | |
1266 | # IPython. Inside IPython, the embeddable shell variable ipshell is just a |
|
1201 | # IPython. Inside IPython, the embeddable shell variable ipshell is just a | |
1267 | # dummy function. |
|
1202 | # dummy function. | |
1268 |
|
1203 | |||
1269 | try: |
|
1204 | try: | |
1270 | __IPYTHON__ |
|
1205 | __IPYTHON__ | |
1271 | except NameError: |
|
1206 | except NameError: | |
1272 | from IPython.Shell import IPShellEmbed |
|
1207 | from IPython.Shell import IPShellEmbed | |
1273 | ipshell = IPShellEmbed() |
|
1208 | ipshell = IPShellEmbed() | |
1274 | # Now ipshell() will open IPython anywhere in the code |
|
1209 | # Now ipshell() will open IPython anywhere in the code | |
1275 | else: |
|
1210 | else: | |
1276 | # Define a dummy ipshell() so the same code doesn't crash inside an |
|
1211 | # Define a dummy ipshell() so the same code doesn't crash inside an | |
1277 | # interactive IPython |
|
1212 | # interactive IPython | |
1278 | def ipshell(): pass |
|
1213 | def ipshell(): pass | |
1279 |
|
1214 | |||
1280 | #******************* End of file <example-embed-short.py> ******************** |
|
1215 | #******************* End of file <example-embed-short.py> ******************** | |
1281 |
|
1216 | |||
1282 | Using the Python debugger (pdb) |
|
1217 | Using the Python debugger (pdb) | |
1283 | =============================== |
|
1218 | =============================== | |
1284 |
|
1219 | |||
1285 | Running entire programs via pdb |
|
1220 | Running entire programs via pdb | |
1286 | ------------------------------- |
|
1221 | ------------------------------- | |
1287 |
|
1222 | |||
1288 | pdb, the Python debugger, is a powerful interactive debugger which |
|
1223 | pdb, the Python debugger, is a powerful interactive debugger which | |
1289 | allows you to step through code, set breakpoints, watch variables, |
|
1224 | allows you to step through code, set breakpoints, watch variables, | |
1290 | etc. IPython makes it very easy to start any script under the control |
|
1225 | etc. IPython makes it very easy to start any script under the control | |
1291 | of pdb, regardless of whether you have wrapped it into a 'main()' |
|
1226 | of pdb, regardless of whether you have wrapped it into a 'main()' | |
1292 | function or not. For this, simply type '%run -d myscript' at an |
|
1227 | function or not. For this, simply type '%run -d myscript' at an | |
1293 | IPython prompt. See the %run command's documentation (via '%run?' or |
|
1228 | IPython prompt. See the %run command's documentation (via '%run?' or | |
1294 | in Sec. magic_ for more details, including how to control where pdb |
|
1229 | in Sec. magic_ for more details, including how to control where pdb | |
1295 | will stop execution first. |
|
1230 | will stop execution first. | |
1296 |
|
1231 | |||
1297 | For more information on the use of the pdb debugger, read the included |
|
1232 | For more information on the use of the pdb debugger, read the included | |
1298 | pdb.doc file (part of the standard Python distribution). On a stock |
|
1233 | pdb.doc file (part of the standard Python distribution). On a stock | |
1299 | Linux system it is located at /usr/lib/python2.3/pdb.doc, but the |
|
1234 | Linux system it is located at /usr/lib/python2.3/pdb.doc, but the | |
1300 | easiest way to read it is by using the help() function of the pdb module |
|
1235 | easiest way to read it is by using the help() function of the pdb module | |
1301 | as follows (in an IPython prompt): |
|
1236 | as follows (in an IPython prompt): | |
1302 |
|
1237 | |||
1303 | In [1]: import pdb |
|
1238 | In [1]: import pdb | |
1304 | In [2]: pdb.help() |
|
1239 | In [2]: pdb.help() | |
1305 |
|
1240 | |||
1306 | This will load the pdb.doc document in a file viewer for you automatically. |
|
1241 | This will load the pdb.doc document in a file viewer for you automatically. | |
1307 |
|
1242 | |||
1308 |
|
1243 | |||
1309 | Automatic invocation of pdb on exceptions |
|
1244 | Automatic invocation of pdb on exceptions | |
1310 | ----------------------------------------- |
|
1245 | ----------------------------------------- | |
1311 |
|
1246 | |||
1312 | IPython, if started with the -pdb option (or if the option is set in |
|
1247 | IPython, if started with the -pdb option (or if the option is set in | |
1313 | your rc file) can call the Python pdb debugger every time your code |
|
1248 | your rc file) can call the Python pdb debugger every time your code | |
1314 | triggers an uncaught exception. This feature |
|
1249 | triggers an uncaught exception. This feature | |
1315 | can also be toggled at any time with the %pdb magic command. This can be |
|
1250 | can also be toggled at any time with the %pdb magic command. This can be | |
1316 | extremely useful in order to find the origin of subtle bugs, because pdb |
|
1251 | extremely useful in order to find the origin of subtle bugs, because pdb | |
1317 | opens up at the point in your code which triggered the exception, and |
|
1252 | opens up at the point in your code which triggered the exception, and | |
1318 | while your program is at this point 'dead', all the data is still |
|
1253 | while your program is at this point 'dead', all the data is still | |
1319 | available and you can walk up and down the stack frame and understand |
|
1254 | available and you can walk up and down the stack frame and understand | |
1320 | the origin of the problem. |
|
1255 | the origin of the problem. | |
1321 |
|
1256 | |||
1322 | Furthermore, you can use these debugging facilities both with the |
|
1257 | Furthermore, you can use these debugging facilities both with the | |
1323 | embedded IPython mode and without IPython at all. For an embedded shell |
|
1258 | embedded IPython mode and without IPython at all. For an embedded shell | |
1324 | (see sec. Embedding_), simply call the constructor with |
|
1259 | (see sec. Embedding_), simply call the constructor with | |
1325 | '-pdb' in the argument string and automatically pdb will be called if an |
|
1260 | '-pdb' in the argument string and automatically pdb will be called if an | |
1326 | uncaught exception is triggered by your code. |
|
1261 | uncaught exception is triggered by your code. | |
1327 |
|
1262 | |||
1328 | For stand-alone use of the feature in your programs which do not use |
|
1263 | For stand-alone use of the feature in your programs which do not use | |
1329 | IPython at all, put the following lines toward the top of your 'main' |
|
1264 | IPython at all, put the following lines toward the top of your 'main' | |
1330 | routine:: |
|
1265 | routine:: | |
1331 |
|
1266 | |||
1332 | import sys |
|
1267 | import sys | |
1333 | from IPython.core import ultratb |
|
1268 | from IPython.core import ultratb | |
1334 | sys.excepthook = ultratb.FormattedTB(mode='Verbose', |
|
1269 | sys.excepthook = ultratb.FormattedTB(mode='Verbose', | |
1335 | color_scheme='Linux', call_pdb=1) |
|
1270 | color_scheme='Linux', call_pdb=1) | |
1336 |
|
1271 | |||
1337 | The mode keyword can be either 'Verbose' or 'Plain', giving either very |
|
1272 | The mode keyword can be either 'Verbose' or 'Plain', giving either very | |
1338 | detailed or normal tracebacks respectively. The color_scheme keyword can |
|
1273 | detailed or normal tracebacks respectively. The color_scheme keyword can | |
1339 | be one of 'NoColor', 'Linux' (default) or 'LightBG'. These are the same |
|
1274 | be one of 'NoColor', 'Linux' (default) or 'LightBG'. These are the same | |
1340 | options which can be set in IPython with -colors and -xmode. |
|
1275 | options which can be set in IPython with -colors and -xmode. | |
1341 |
|
1276 | |||
1342 | This will give any of your programs detailed, colored tracebacks with |
|
1277 | This will give any of your programs detailed, colored tracebacks with | |
1343 | automatic invocation of pdb. |
|
1278 | automatic invocation of pdb. | |
1344 |
|
1279 | |||
1345 |
|
1280 | |||
1346 | Extensions for syntax processing |
|
1281 | Extensions for syntax processing | |
1347 | ================================ |
|
1282 | ================================ | |
1348 |
|
1283 | |||
1349 | This isn't for the faint of heart, because the potential for breaking |
|
1284 | This isn't for the faint of heart, because the potential for breaking | |
1350 | things is quite high. But it can be a very powerful and useful feature. |
|
1285 | things is quite high. But it can be a very powerful and useful feature. | |
1351 | In a nutshell, you can redefine the way IPython processes the user input |
|
1286 | In a nutshell, you can redefine the way IPython processes the user input | |
1352 | line to accept new, special extensions to the syntax without needing to |
|
1287 | line to accept new, special extensions to the syntax without needing to | |
1353 | change any of IPython's own code. |
|
1288 | change any of IPython's own code. | |
1354 |
|
1289 | |||
1355 | In the IPython/extensions directory you will find some examples |
|
1290 | In the IPython/extensions directory you will find some examples | |
1356 | supplied, which we will briefly describe now. These can be used 'as is' |
|
1291 | supplied, which we will briefly describe now. These can be used 'as is' | |
1357 | (and both provide very useful functionality), or you can use them as a |
|
1292 | (and both provide very useful functionality), or you can use them as a | |
1358 | starting point for writing your own extensions. |
|
1293 | starting point for writing your own extensions. | |
1359 |
|
1294 | |||
1360 |
|
1295 | |||
1361 | Pasting of code starting with '>>> ' or '... ' |
|
1296 | Pasting of code starting with '>>> ' or '... ' | |
1362 | ---------------------------------------------- |
|
1297 | ---------------------------------------------- | |
1363 |
|
1298 | |||
1364 | In the python tutorial it is common to find code examples which have |
|
1299 | In the python tutorial it is common to find code examples which have | |
1365 | been taken from real python sessions. The problem with those is that all |
|
1300 | been taken from real python sessions. The problem with those is that all | |
1366 | the lines begin with either '>>> ' or '... ', which makes it impossible |
|
1301 | the lines begin with either '>>> ' or '... ', which makes it impossible | |
1367 | to paste them all at once. One must instead do a line by line manual |
|
1302 | to paste them all at once. One must instead do a line by line manual | |
1368 | copying, carefully removing the leading extraneous characters. |
|
1303 | copying, carefully removing the leading extraneous characters. | |
1369 |
|
1304 | |||
1370 | This extension identifies those starting characters and removes them |
|
1305 | This extension identifies those starting characters and removes them | |
1371 | from the input automatically, so that one can paste multi-line examples |
|
1306 | from the input automatically, so that one can paste multi-line examples | |
1372 | directly into IPython, saving a lot of time. Please look at the file |
|
1307 | directly into IPython, saving a lot of time. Please look at the file | |
1373 | InterpreterPasteInput.py in the IPython/extensions directory for details |
|
1308 | InterpreterPasteInput.py in the IPython/extensions directory for details | |
1374 | on how this is done. |
|
1309 | on how this is done. | |
1375 |
|
1310 | |||
1376 | IPython comes with a special profile enabling this feature, called |
|
1311 | IPython comes with a special profile enabling this feature, called | |
1377 | tutorial. Simply start IPython via 'ipython -p tutorial' and the feature |
|
1312 | tutorial. Simply start IPython via 'ipython -p tutorial' and the feature | |
1378 | will be available. In a normal IPython session you can activate the |
|
1313 | will be available. In a normal IPython session you can activate the | |
1379 | feature by importing the corresponding module with: |
|
1314 | feature by importing the corresponding module with: | |
1380 | In [1]: import IPython.extensions.InterpreterPasteInput |
|
1315 | In [1]: import IPython.extensions.InterpreterPasteInput | |
1381 |
|
1316 | |||
1382 | The following is a 'screenshot' of how things work when this extension |
|
1317 | The following is a 'screenshot' of how things work when this extension | |
1383 | is on, copying an example from the standard tutorial:: |
|
1318 | is on, copying an example from the standard tutorial:: | |
1384 |
|
1319 | |||
1385 | IPython profile: tutorial |
|
1320 | IPython profile: tutorial | |
1386 |
|
1321 | |||
1387 | *** Pasting of code with ">>>" or "..." has been enabled. |
|
1322 | *** Pasting of code with ">>>" or "..." has been enabled. | |
1388 |
|
1323 | |||
1389 | In [1]: >>> def fib2(n): # return Fibonacci series up to n |
|
1324 | In [1]: >>> def fib2(n): # return Fibonacci series up to n | |
1390 | ...: ... """Return a list containing the Fibonacci series up to |
|
1325 | ...: ... """Return a list containing the Fibonacci series up to | |
1391 | n.""" |
|
1326 | n.""" | |
1392 | ...: ... result = [] |
|
1327 | ...: ... result = [] | |
1393 | ...: ... a, b = 0, 1 |
|
1328 | ...: ... a, b = 0, 1 | |
1394 | ...: ... while b < n: |
|
1329 | ...: ... while b < n: | |
1395 | ...: ... result.append(b) # see below |
|
1330 | ...: ... result.append(b) # see below | |
1396 | ...: ... a, b = b, a+b |
|
1331 | ...: ... a, b = b, a+b | |
1397 | ...: ... return result |
|
1332 | ...: ... return result | |
1398 | ...: |
|
1333 | ...: | |
1399 |
|
1334 | |||
1400 | In [2]: fib2(10) |
|
1335 | In [2]: fib2(10) | |
1401 | Out[2]: [1, 1, 2, 3, 5, 8] |
|
1336 | Out[2]: [1, 1, 2, 3, 5, 8] | |
1402 |
|
1337 | |||
1403 | Note that as currently written, this extension does not recognize |
|
1338 | Note that as currently written, this extension does not recognize | |
1404 | IPython's prompts for pasting. Those are more complicated, since the |
|
1339 | IPython's prompts for pasting. Those are more complicated, since the | |
1405 | user can change them very easily, they involve numbers and can vary in |
|
1340 | user can change them very easily, they involve numbers and can vary in | |
1406 | length. One could however extract all the relevant information from the |
|
1341 | length. One could however extract all the relevant information from the | |
1407 | IPython instance and build an appropriate regular expression. This is |
|
1342 | IPython instance and build an appropriate regular expression. This is | |
1408 | left as an exercise for the reader. |
|
1343 | left as an exercise for the reader. | |
1409 |
|
1344 | |||
1410 |
|
1345 | |||
1411 | Input of physical quantities with units |
|
1346 | Input of physical quantities with units | |
1412 | --------------------------------------- |
|
1347 | --------------------------------------- | |
1413 |
|
1348 | |||
1414 | The module PhysicalQInput allows a simplified form of input for physical |
|
1349 | The module PhysicalQInput allows a simplified form of input for physical | |
1415 | quantities with units. This file is meant to be used in conjunction with |
|
1350 | quantities with units. This file is meant to be used in conjunction with | |
1416 | the PhysicalQInteractive module (in the same directory) and |
|
1351 | the PhysicalQInteractive module (in the same directory) and | |
1417 | Physics.PhysicalQuantities from Konrad Hinsen's ScientificPython |
|
1352 | Physics.PhysicalQuantities from Konrad Hinsen's ScientificPython | |
1418 | (http://dirac.cnrs-orleans.fr/ScientificPython/). |
|
1353 | (http://dirac.cnrs-orleans.fr/ScientificPython/). | |
1419 |
|
1354 | |||
1420 | The Physics.PhysicalQuantities module defines PhysicalQuantity objects, |
|
1355 | The Physics.PhysicalQuantities module defines PhysicalQuantity objects, | |
1421 | but these must be declared as instances of a class. For example, to |
|
1356 | but these must be declared as instances of a class. For example, to | |
1422 | define v as a velocity of 3 m/s, normally you would write:: |
|
1357 | define v as a velocity of 3 m/s, normally you would write:: | |
1423 |
|
1358 | |||
1424 | In [1]: v = PhysicalQuantity(3,'m/s') |
|
1359 | In [1]: v = PhysicalQuantity(3,'m/s') | |
1425 |
|
1360 | |||
1426 | Using the PhysicalQ_Input extension this can be input instead as: |
|
1361 | Using the PhysicalQ_Input extension this can be input instead as: | |
1427 | In [1]: v = 3 m/s |
|
1362 | In [1]: v = 3 m/s | |
1428 | which is much more convenient for interactive use (even though it is |
|
1363 | which is much more convenient for interactive use (even though it is | |
1429 | blatantly invalid Python syntax). |
|
1364 | blatantly invalid Python syntax). | |
1430 |
|
1365 | |||
1431 | The physics profile supplied with IPython (enabled via 'ipython -p |
|
1366 | The physics profile supplied with IPython (enabled via 'ipython -p | |
1432 | physics') uses these extensions, which you can also activate with: |
|
1367 | physics') uses these extensions, which you can also activate with: | |
1433 |
|
1368 | |||
1434 | from math import * # math MUST be imported BEFORE PhysicalQInteractive |
|
1369 | from math import * # math MUST be imported BEFORE PhysicalQInteractive | |
1435 | from IPython.extensions.PhysicalQInteractive import * |
|
1370 | from IPython.extensions.PhysicalQInteractive import * | |
1436 | import IPython.extensions.PhysicalQInput |
|
1371 | import IPython.extensions.PhysicalQInput | |
1437 |
|
1372 | |||
|
1373 | .. _gui_support: | |||
1438 |
|
1374 | |||
1439 | Threading support |
|
1375 | GUI event loop support support | |
1440 | ================= |
|
1376 | ============================== | |
|
1377 | ||||
|
1378 | .. versionadded:: 0.11 | |||
|
1379 | The ``%gui`` magic and :mod:`IPython.lib.inputhook`. | |||
|
1380 | ||||
|
1381 | IPython has excellent support for working interactively with Graphical User | |||
|
1382 | Interface (GUI) toolkits, such as wxPython, PyQt4, PyGTK and Tk. This is | |||
|
1383 | implemented using Python's builtin ``PyOSInputHook`` hook. This implementation | |||
|
1384 | is extremely robust compared to our previous threaded based version. The | |||
|
1385 | advantages of | |||
|
1386 | ||||
|
1387 | * GUIs can be enabled and disabled dynamically at runtime. | |||
|
1388 | * The active GUI can be switched dynamically at runtime. | |||
|
1389 | * In some cases, multiple GUIs can run simultaneously with no problems. | |||
|
1390 | * There is a developer API in :mod:`IPython.lib.inputhook` for customizing | |||
|
1391 | all of these things. | |||
|
1392 | ||||
|
1393 | For users, enabling GUI event loop integration is simple. You simple use the | |||
|
1394 | ``%gui`` magic as follows:: | |||
|
1395 | ||||
|
1396 | %gui [-a] [GUINAME] | |||
|
1397 | ||||
|
1398 | With no arguments, ``%gui`` removes all GUI support. Valid ``GUINAME`` | |||
|
1399 | arguments are ``wx``, ``qt4``, ``gtk`` and ``tk``. The ``-a`` option will | |||
|
1400 | create and return a running application object for the selected GUI toolkit. | |||
|
1401 | ||||
|
1402 | This to use wxPython interactively and create a running :class:`wx.App` | |||
|
1403 | object, do:: | |||
|
1404 | ||||
|
1405 | %gui -a wx | |||
|
1406 | ||||
|
1407 | For information on IPython's Matplotlib integration (and the ``pylab`` mode) | |||
|
1408 | see :ref:`this section <matplotlib_support>`. | |||
1441 |
|
1409 | |||
1442 | WARNING: The threading support is still somewhat experimental, and it |
|
1410 | For developers that want to use IPython's GUI event loop integration in | |
1443 | has only seen reasonable testing under Linux. Threaded code is |
|
1411 | the form of a library, the capabilities are exposed in library form | |
1444 | particularly tricky to debug, and it tends to show extremely |
|
1412 | in the :mod:`IPython.lib.inputhook`. Interested developers should see the | |
1445 | platform-dependent behavior. Since I only have access to Linux machines, |
|
1413 | module docstrings for more information. | |
1446 | I will have to rely on user's experiences and assistance for this area |
|
1414 | ||
1447 | of IPython to improve under other platforms. |
|
1415 | .. _matplotlib_support: | |
1448 |
|
1416 | |||
1449 | IPython, via the -gthread , -qthread, -q4thread and -wthread options |
|
1417 | Plotting with matplotlib | |
1450 | (described in Sec. `Threading options`_), can run in |
|
1418 | ======================== | |
1451 | multithreaded mode to support pyGTK, Qt3, Qt4 and WXPython applications |
|
1419 | ||
1452 | respectively. These GUI toolkits need to control the python main loop of |
|
1420 | ||
1453 | execution, so under a normal Python interpreter, starting a pyGTK, Qt3, |
|
1421 | `Matplotlib`_ provides high quality 2D and | |
1454 | Qt4 or WXPython application will immediately freeze the shell. |
|
1422 | 3D plotting for Python. Matplotlib can produce plots on screen using a variety | |
1455 |
|
1423 | of GUI toolkits, including Tk, PyGTK, PyQt4 and wxPython. It also provides a | ||
1456 | IPython, with one of these options (you can only use one at a time), |
|
1424 | number of commands useful for scientific computing, all with a syntax | |
1457 | separates the graphical loop and IPython's code execution run into |
|
1425 | compatible with that of the popular Matlab program. | |
1458 | different threads. This allows you to test interactively (with %run, for |
|
1426 | ||
1459 | example) your GUI code without blocking. |
|
1427 | Many IPython users have come to rely on IPython's ``-pylab`` mode which | |
1460 |
|
1428 | automates the integration of Matplotlib with IPython. We are still in the | ||
1461 | A nice mini-tutorial on using IPython along with the Qt Designer |
|
1429 | process of working with the Matplotlib developers to finalize the new pylab | |
1462 | application is available at the SciPy wiki: |
|
1430 | API, but for now you can use Matplotlib interactively using the following | |
1463 | http://www.scipy.org/Cookbook/Matplotlib/Qt_with_IPython_and_Designer. |
|
1431 | commands:: | |
1464 |
|
1432 | |||
1465 |
|
1433 | %gui -a wx | ||
1466 | Tk issues |
|
1434 | import matplotlib | |
1467 | --------- |
|
1435 | matplotlib.use('wxagg') | |
1468 |
|
1436 | from matplotlib import pylab | ||
1469 | As indicated in Sec. `Threading options`_, a special -tk option is |
|
1437 | pylab.interactive(True) | |
1470 | provided to try and allow Tk graphical applications to coexist |
|
1438 | ||
1471 | interactively with WX, Qt or GTK ones. Whether this works at all, |
|
1439 | All of this will soon be automated as Matplotlib beings to include | |
1472 | however, is very platform and configuration dependent. Please |
|
1440 | new logic that uses our new GUI support. | |
1473 | experiment with simple test cases before committing to using this |
|
|||
1474 | combination of Tk and GTK/Qt/WX threading in a production environment. |
|
|||
1475 |
|
||||
1476 |
|
||||
1477 | I/O pitfalls |
|
|||
1478 | ------------ |
|
|||
1479 |
|
||||
1480 | Be mindful that the Python interpreter switches between threads every |
|
|||
1481 | $N$ bytecodes, where the default value as of Python 2.3 is $N=100.$ This |
|
|||
1482 | value can be read by using the sys.getcheckinterval() function, and it |
|
|||
1483 | can be reset via sys.setcheckinterval(N). This switching of threads can |
|
|||
1484 | cause subtly confusing effects if one of your threads is doing file I/O. |
|
|||
1485 | In text mode, most systems only flush file buffers when they encounter a |
|
|||
1486 | '\n'. An instruction as simple as:: |
|
|||
1487 |
|
||||
1488 | print >> filehandle, ''hello world'' |
|
|||
1489 |
|
||||
1490 | actually consists of several bytecodes, so it is possible that the |
|
|||
1491 | newline does not reach your file before the next thread switch. |
|
|||
1492 | Similarly, if you are writing to a file in binary mode, the file won't |
|
|||
1493 | be flushed until the buffer fills, and your other thread may see |
|
|||
1494 | apparently truncated files. |
|
|||
1495 |
|
||||
1496 | For this reason, if you are using IPython's thread support and have (for |
|
|||
1497 | example) a GUI application which will read data generated by files |
|
|||
1498 | written to from the IPython thread, the safest approach is to open all |
|
|||
1499 | of your files in unbuffered mode (the third argument to the file/open |
|
|||
1500 | function is the buffering value):: |
|
|||
1501 |
|
||||
1502 | filehandle = open(filename,mode,0) |
|
|||
1503 |
|
||||
1504 | This is obviously a brute force way of avoiding race conditions with the |
|
|||
1505 | file buffering. If you want to do it cleanly, and you have a resource |
|
|||
1506 | which is being shared by the interactive IPython loop and your GUI |
|
|||
1507 | thread, you should really handle it with thread locking and |
|
|||
1508 | syncrhonization properties. The Python documentation discusses these. |
|
|||
1509 |
|
1441 | |||
1510 | .. _interactive_demos: |
|
1442 | .. _interactive_demos: | |
1511 |
|
1443 | |||
1512 | Interactive demos with IPython |
|
1444 | Interactive demos with IPython | |
1513 | ============================== |
|
1445 | ============================== | |
1514 |
|
1446 | |||
1515 | IPython ships with a basic system for running scripts interactively in |
|
1447 | IPython ships with a basic system for running scripts interactively in | |
1516 | sections, useful when presenting code to audiences. A few tags embedded |
|
1448 | sections, useful when presenting code to audiences. A few tags embedded | |
1517 | in comments (so that the script remains valid Python code) divide a file |
|
1449 | in comments (so that the script remains valid Python code) divide a file | |
1518 | into separate blocks, and the demo can be run one block at a time, with |
|
1450 | into separate blocks, and the demo can be run one block at a time, with | |
1519 | IPython printing (with syntax highlighting) the block before executing |
|
1451 | IPython printing (with syntax highlighting) the block before executing | |
1520 | it, and returning to the interactive prompt after each block. The |
|
1452 | it, and returning to the interactive prompt after each block. The | |
1521 | interactive namespace is updated after each block is run with the |
|
1453 | interactive namespace is updated after each block is run with the | |
1522 | contents of the demo's namespace. |
|
1454 | contents of the demo's namespace. | |
1523 |
|
1455 | |||
1524 | This allows you to show a piece of code, run it and then execute |
|
1456 | This allows you to show a piece of code, run it and then execute | |
1525 | interactively commands based on the variables just created. Once you |
|
1457 | interactively commands based on the variables just created. Once you | |
1526 | want to continue, you simply execute the next block of the demo. The |
|
1458 | want to continue, you simply execute the next block of the demo. The | |
1527 | following listing shows the markup necessary for dividing a script into |
|
1459 | following listing shows the markup necessary for dividing a script into | |
1528 | sections for execution as a demo:: |
|
1460 | sections for execution as a demo:: | |
1529 |
|
1461 | |||
1530 |
|
1462 | |||
1531 | """A simple interactive demo to illustrate the use of IPython's Demo class. |
|
1463 | """A simple interactive demo to illustrate the use of IPython's Demo class. | |
1532 |
|
1464 | |||
1533 | Any python script can be run as a demo, but that does little more than showing |
|
1465 | Any python script can be run as a demo, but that does little more than showing | |
1534 | it on-screen, syntax-highlighted in one shot. If you add a little simple |
|
1466 | it on-screen, syntax-highlighted in one shot. If you add a little simple | |
1535 | markup, you can stop at specified intervals and return to the ipython prompt, |
|
1467 | markup, you can stop at specified intervals and return to the ipython prompt, | |
1536 | resuming execution later. |
|
1468 | resuming execution later. | |
1537 | """ |
|
1469 | """ | |
1538 |
|
1470 | |||
1539 | print 'Hello, welcome to an interactive IPython demo.' |
|
1471 | print 'Hello, welcome to an interactive IPython demo.' | |
1540 | print 'Executing this block should require confirmation before proceeding,' |
|
1472 | print 'Executing this block should require confirmation before proceeding,' | |
1541 | print 'unless auto_all has been set to true in the demo object' |
|
1473 | print 'unless auto_all has been set to true in the demo object' | |
1542 |
|
1474 | |||
1543 | # The mark below defines a block boundary, which is a point where IPython will |
|
1475 | # The mark below defines a block boundary, which is a point where IPython will | |
1544 | # stop execution and return to the interactive prompt. |
|
1476 | # stop execution and return to the interactive prompt. | |
1545 | # Note that in actual interactive execution, |
|
1477 | # Note that in actual interactive execution, | |
1546 | # <demo> --- stop --- |
|
1478 | # <demo> --- stop --- | |
1547 |
|
1479 | |||
1548 | x = 1 |
|
1480 | x = 1 | |
1549 | y = 2 |
|
1481 | y = 2 | |
1550 |
|
1482 | |||
1551 | # <demo> --- stop --- |
|
1483 | # <demo> --- stop --- | |
1552 |
|
1484 | |||
1553 | # the mark below makes this block as silent |
|
1485 | # the mark below makes this block as silent | |
1554 | # <demo> silent |
|
1486 | # <demo> silent | |
1555 |
|
1487 | |||
1556 | print 'This is a silent block, which gets executed but not printed.' |
|
1488 | print 'This is a silent block, which gets executed but not printed.' | |
1557 |
|
1489 | |||
1558 | # <demo> --- stop --- |
|
1490 | # <demo> --- stop --- | |
1559 | # <demo> auto |
|
1491 | # <demo> auto | |
1560 | print 'This is an automatic block.' |
|
1492 | print 'This is an automatic block.' | |
1561 | print 'It is executed without asking for confirmation, but printed.' |
|
1493 | print 'It is executed without asking for confirmation, but printed.' | |
1562 | z = x+y |
|
1494 | z = x+y | |
1563 |
|
1495 | |||
1564 | print 'z=',x |
|
1496 | print 'z=',x | |
1565 |
|
1497 | |||
1566 | # <demo> --- stop --- |
|
1498 | # <demo> --- stop --- | |
1567 | # This is just another normal block. |
|
1499 | # This is just another normal block. | |
1568 | print 'z is now:', z |
|
1500 | print 'z is now:', z | |
1569 |
|
1501 | |||
1570 | print 'bye!' |
|
1502 | print 'bye!' | |
1571 |
|
1503 | |||
1572 | In order to run a file as a demo, you must first make a Demo object out |
|
1504 | In order to run a file as a demo, you must first make a Demo object out | |
1573 | of it. If the file is named myscript.py, the following code will make a |
|
1505 | of it. If the file is named myscript.py, the following code will make a | |
1574 | demo:: |
|
1506 | demo:: | |
1575 |
|
1507 | |||
1576 | from IPython.demo import Demo |
|
1508 | from IPython.demo import Demo | |
1577 |
|
1509 | |||
1578 | mydemo = Demo('myscript.py') |
|
1510 | mydemo = Demo('myscript.py') | |
1579 |
|
1511 | |||
1580 | This creates the mydemo object, whose blocks you run one at a time by |
|
1512 | This creates the mydemo object, whose blocks you run one at a time by | |
1581 | simply calling the object with no arguments. If you have autocall active |
|
1513 | simply calling the object with no arguments. If you have autocall active | |
1582 | in IPython (the default), all you need to do is type:: |
|
1514 | in IPython (the default), all you need to do is type:: | |
1583 |
|
1515 | |||
1584 | mydemo |
|
1516 | mydemo | |
1585 |
|
1517 | |||
1586 | and IPython will call it, executing each block. Demo objects can be |
|
1518 | and IPython will call it, executing each block. Demo objects can be | |
1587 | restarted, you can move forward or back skipping blocks, re-execute the |
|
1519 | restarted, you can move forward or back skipping blocks, re-execute the | |
1588 | last block, etc. Simply use the Tab key on a demo object to see its |
|
1520 | last block, etc. Simply use the Tab key on a demo object to see its | |
1589 | methods, and call '?' on them to see their docstrings for more usage |
|
1521 | methods, and call '?' on them to see their docstrings for more usage | |
1590 | details. In addition, the demo module itself contains a comprehensive |
|
1522 | details. In addition, the demo module itself contains a comprehensive | |
1591 | docstring, which you can access via:: |
|
1523 | docstring, which you can access via:: | |
1592 |
|
1524 | |||
1593 | from IPython import demo |
|
1525 | from IPython import demo | |
1594 |
|
1526 | |||
1595 | demo? |
|
1527 | demo? | |
1596 |
|
1528 | |||
1597 | Limitations: It is important to note that these demos are limited to |
|
1529 | Limitations: It is important to note that these demos are limited to | |
1598 | fairly simple uses. In particular, you can not put division marks in |
|
1530 | fairly simple uses. In particular, you can not put division marks in | |
1599 | indented code (loops, if statements, function definitions, etc.) |
|
1531 | indented code (loops, if statements, function definitions, etc.) | |
1600 | Supporting something like this would basically require tracking the |
|
1532 | Supporting something like this would basically require tracking the | |
1601 | internal execution state of the Python interpreter, so only top-level |
|
1533 | internal execution state of the Python interpreter, so only top-level | |
1602 | divisions are allowed. If you want to be able to open an IPython |
|
1534 | divisions are allowed. If you want to be able to open an IPython | |
1603 | instance at an arbitrary point in a program, you can use IPython's |
|
1535 | instance at an arbitrary point in a program, you can use IPython's | |
1604 | embedding facilities, described in detail in Sec. 9 |
|
1536 | embedding facilities, described in detail in Sec. 9 | |
1605 |
|
1537 | |||
|
1538 | .. [Matplotlib] Matplotlib. http://matplotlib.sourceforge.net | |||
1606 |
|
1539 | |||
1607 | .. _Matplotlib support: |
|
|||
1608 |
|
||||
1609 | Plotting with matplotlib |
|
|||
1610 | ======================== |
|
|||
1611 |
|
||||
1612 | The matplotlib library (http://matplotlib.sourceforge.net |
|
|||
1613 | http://matplotlib.sourceforge.net) provides high quality 2D plotting for |
|
|||
1614 | Python. Matplotlib can produce plots on screen using a variety of GUI |
|
|||
1615 | toolkits, including Tk, GTK and WXPython. It also provides a number of |
|
|||
1616 | commands useful for scientific computing, all with a syntax compatible |
|
|||
1617 | with that of the popular Matlab program. |
|
|||
1618 |
|
||||
1619 | IPython accepts the special option -pylab (see :ref:`here |
|
|||
1620 | <command_line_options>`). This configures it to support matplotlib, honoring |
|
|||
1621 | the settings in the .matplotlibrc file. IPython will detect the user's choice |
|
|||
1622 | of matplotlib GUI backend, and automatically select the proper threading model |
|
|||
1623 | to prevent blocking. It also sets matplotlib in interactive mode and modifies |
|
|||
1624 | %run slightly, so that any matplotlib-based script can be executed using %run |
|
|||
1625 | and the final show() command does not block the interactive shell. |
|
|||
1626 |
|
||||
1627 | The -pylab option must be given first in order for IPython to configure its |
|
|||
1628 | threading mode. However, you can still issue other options afterwards. This |
|
|||
1629 | allows you to have a matplotlib-based environment customized with additional |
|
|||
1630 | modules using the standard IPython profile mechanism (see :ref:`here |
|
|||
1631 | <profiles>`): ``ipython -pylab -p myprofile`` will load the profile defined in |
|
|||
1632 | ipythonrc-myprofile after configuring matplotlib. |
|
@@ -1,16 +1,15 b'' | |||||
1 | .. _parallel_index: |
|
1 | .. _parallel_index: | |
2 |
|
2 | |||
3 | ==================================== |
|
3 | ==================================== | |
4 | Using IPython for parallel computing |
|
4 | Using IPython for parallel computing | |
5 | ==================================== |
|
5 | ==================================== | |
6 |
|
6 | |||
7 | .. toctree:: |
|
7 | .. toctree:: | |
8 | :maxdepth: 2 |
|
8 | :maxdepth: 2 | |
9 |
|
9 | |||
10 | parallel_intro.txt |
|
10 | parallel_intro.txt | |
11 | parallel_process.txt |
|
11 | parallel_process.txt | |
12 | parallel_multiengine.txt |
|
12 | parallel_multiengine.txt | |
13 | parallel_task.txt |
|
13 | parallel_task.txt | |
14 | parallel_mpi.txt |
|
14 | parallel_mpi.txt | |
15 | parallel_security.txt |
|
15 | parallel_security.txt | |
16 | visionhpc.txt |
|
@@ -1,205 +1,237 b'' | |||||
1 | .. _ip1par: |
|
1 | .. _ip1par: | |
2 |
|
2 | |||
3 | ============================ |
|
3 | ============================ | |
4 | Overview and getting started |
|
4 | Overview and getting started | |
5 | ============================ |
|
5 | ============================ | |
6 |
|
6 | |||
7 | Introduction |
|
7 | Introduction | |
8 | ============ |
|
8 | ============ | |
9 |
|
9 | |||
10 | This section gives an overview of IPython's sophisticated and powerful |
|
10 | This section gives an overview of IPython's sophisticated and powerful | |
11 | architecture for parallel and distributed computing. This architecture |
|
11 | architecture for parallel and distributed computing. This architecture | |
12 | abstracts out parallelism in a very general way, which enables IPython to |
|
12 | abstracts out parallelism in a very general way, which enables IPython to | |
13 | support many different styles of parallelism including: |
|
13 | support many different styles of parallelism including: | |
14 |
|
14 | |||
15 | * Single program, multiple data (SPMD) parallelism. |
|
15 | * Single program, multiple data (SPMD) parallelism. | |
16 | * Multiple program, multiple data (MPMD) parallelism. |
|
16 | * Multiple program, multiple data (MPMD) parallelism. | |
17 | * Message passing using MPI. |
|
17 | * Message passing using MPI. | |
18 | * Task farming. |
|
18 | * Task farming. | |
19 | * Data parallel. |
|
19 | * Data parallel. | |
20 | * Combinations of these approaches. |
|
20 | * Combinations of these approaches. | |
21 | * Custom user defined approaches. |
|
21 | * Custom user defined approaches. | |
22 |
|
22 | |||
23 | Most importantly, IPython enables all types of parallel applications to |
|
23 | Most importantly, IPython enables all types of parallel applications to | |
24 | be developed, executed, debugged and monitored *interactively*. Hence, |
|
24 | be developed, executed, debugged and monitored *interactively*. Hence, | |
25 | the ``I`` in IPython. The following are some example usage cases for IPython: |
|
25 | the ``I`` in IPython. The following are some example usage cases for IPython: | |
26 |
|
26 | |||
27 | * Quickly parallelize algorithms that are embarrassingly parallel |
|
27 | * Quickly parallelize algorithms that are embarrassingly parallel | |
28 | using a number of simple approaches. Many simple things can be |
|
28 | using a number of simple approaches. Many simple things can be | |
29 | parallelized interactively in one or two lines of code. |
|
29 | parallelized interactively in one or two lines of code. | |
30 |
|
30 | |||
31 | * Steer traditional MPI applications on a supercomputer from an |
|
31 | * Steer traditional MPI applications on a supercomputer from an | |
32 | IPython session on your laptop. |
|
32 | IPython session on your laptop. | |
33 |
|
33 | |||
34 | * Analyze and visualize large datasets (that could be remote and/or |
|
34 | * Analyze and visualize large datasets (that could be remote and/or | |
35 | distributed) interactively using IPython and tools like |
|
35 | distributed) interactively using IPython and tools like | |
36 | matplotlib/TVTK. |
|
36 | matplotlib/TVTK. | |
37 |
|
37 | |||
38 | * Develop, test and debug new parallel algorithms |
|
38 | * Develop, test and debug new parallel algorithms | |
39 | (that may use MPI) interactively. |
|
39 | (that may use MPI) interactively. | |
40 |
|
40 | |||
41 | * Tie together multiple MPI jobs running on different systems into |
|
41 | * Tie together multiple MPI jobs running on different systems into | |
42 | one giant distributed and parallel system. |
|
42 | one giant distributed and parallel system. | |
43 |
|
43 | |||
44 | * Start a parallel job on your cluster and then have a remote |
|
44 | * Start a parallel job on your cluster and then have a remote | |
45 | collaborator connect to it and pull back data into their |
|
45 | collaborator connect to it and pull back data into their | |
46 | local IPython session for plotting and analysis. |
|
46 | local IPython session for plotting and analysis. | |
47 |
|
47 | |||
48 | * Run a set of tasks on a set of CPUs using dynamic load balancing. |
|
48 | * Run a set of tasks on a set of CPUs using dynamic load balancing. | |
49 |
|
49 | |||
50 | Architecture overview |
|
50 | Architecture overview | |
51 | ===================== |
|
51 | ===================== | |
52 |
|
52 | |||
53 | The IPython architecture consists of three components: |
|
53 | The IPython architecture consists of three components: | |
54 |
|
54 | |||
55 | * The IPython engine. |
|
55 | * The IPython engine. | |
56 | * The IPython controller. |
|
56 | * The IPython controller. | |
57 | * Various controller clients. |
|
57 | * Various controller clients. | |
58 |
|
58 | |||
59 | These components live in the :mod:`IPython.kernel` package and are |
|
59 | These components live in the :mod:`IPython.kernel` package and are | |
60 | installed with IPython. They do, however, have additional dependencies |
|
60 | installed with IPython. They do, however, have additional dependencies | |
61 | that must be installed. For more information, see our |
|
61 | that must be installed. For more information, see our | |
62 | :ref:`installation documentation <install_index>`. |
|
62 | :ref:`installation documentation <install_index>`. | |
63 |
|
63 | |||
64 | IPython engine |
|
64 | IPython engine | |
65 | --------------- |
|
65 | --------------- | |
66 |
|
66 | |||
67 | The IPython engine is a Python instance that takes Python commands over a |
|
67 | The IPython engine is a Python instance that takes Python commands over a | |
68 | network connection. Eventually, the IPython engine will be a full IPython |
|
68 | network connection. Eventually, the IPython engine will be a full IPython | |
69 | interpreter, but for now, it is a regular Python interpreter. The engine |
|
69 | interpreter, but for now, it is a regular Python interpreter. The engine | |
70 | can also handle incoming and outgoing Python objects sent over a network |
|
70 | can also handle incoming and outgoing Python objects sent over a network | |
71 | connection. When multiple engines are started, parallel and distributed |
|
71 | connection. When multiple engines are started, parallel and distributed | |
72 | computing becomes possible. An important feature of an IPython engine is |
|
72 | computing becomes possible. An important feature of an IPython engine is | |
73 | that it blocks while user code is being executed. Read on for how the |
|
73 | that it blocks while user code is being executed. Read on for how the | |
74 | IPython controller solves this problem to expose a clean asynchronous API |
|
74 | IPython controller solves this problem to expose a clean asynchronous API | |
75 | to the user. |
|
75 | to the user. | |
76 |
|
76 | |||
77 | IPython controller |
|
77 | IPython controller | |
78 | ------------------ |
|
78 | ------------------ | |
79 |
|
79 | |||
80 | The IPython controller provides an interface for working with a set of |
|
80 | The IPython controller provides an interface for working with a set of | |
81 | engines. At an general level, the controller is a process to which |
|
81 | engines. At an general level, the controller is a process to which | |
82 | IPython engines can connect. For each connected engine, the controller |
|
82 | IPython engines can connect. For each connected engine, the controller | |
83 | manages a queue. All actions that can be performed on the engine go |
|
83 | manages a queue. All actions that can be performed on the engine go | |
84 | through this queue. While the engines themselves block when user code is |
|
84 | through this queue. While the engines themselves block when user code is | |
85 | run, the controller hides that from the user to provide a fully |
|
85 | run, the controller hides that from the user to provide a fully | |
86 | asynchronous interface to a set of engines. |
|
86 | asynchronous interface to a set of engines. | |
87 |
|
87 | |||
88 | .. note:: |
|
88 | .. note:: | |
89 |
|
89 | |||
90 | Because the controller listens on a network port for engines to |
|
90 | Because the controller listens on a network port for engines to | |
91 | connect to it, it must be started *before* any engines are started. |
|
91 | connect to it, it must be started *before* any engines are started. | |
92 |
|
92 | |||
93 | The controller also provides a single point of contact for users who wish |
|
93 | The controller also provides a single point of contact for users who wish to | |
94 |
|
|
94 | utilize the engines connected to the controller. There are different ways of | |
95 | ways of working with a controller. In IPython these ways correspond to different interfaces that the controller is adapted to. Currently we have two default interfaces to the controller: |
|
95 | working with a controller. In IPython these ways correspond to different | |
|
96 | interfaces that the controller is adapted to. Currently we have two default | |||
|
97 | interfaces to the controller: | |||
96 |
|
98 | |||
97 | * The MultiEngine interface, which provides the simplest possible way of |
|
99 | * The MultiEngine interface, which provides the simplest possible way of | |
98 | working with engines interactively. |
|
100 | working with engines interactively. | |
99 |
* The Task interface, which pr |
|
101 | * The Task interface, which presents the engines as a load balanced | |
100 | task farming system. |
|
102 | task farming system. | |
101 |
|
103 | |||
102 | Advanced users can easily add new custom interfaces to enable other |
|
104 | Advanced users can easily add new custom interfaces to enable other | |
103 | styles of parallelism. |
|
105 | styles of parallelism. | |
104 |
|
106 | |||
105 | .. note:: |
|
107 | .. note:: | |
106 |
|
108 | |||
107 | A single controller and set of engines can be accessed |
|
109 | A single controller and set of engines can be accessed | |
108 | through multiple interfaces simultaneously. This opens the |
|
110 | through multiple interfaces simultaneously. This opens the | |
109 | door for lots of interesting things. |
|
111 | door for lots of interesting things. | |
110 |
|
112 | |||
111 | Controller clients |
|
113 | Controller clients | |
112 | ------------------ |
|
114 | ------------------ | |
113 |
|
115 | |||
114 | For each controller interface, there is a corresponding client. These |
|
116 | For each controller interface, there is a corresponding client. These | |
115 | clients allow users to interact with a set of engines through the |
|
117 | clients allow users to interact with a set of engines through the | |
116 | interface. Here are the two default clients: |
|
118 | interface. Here are the two default clients: | |
117 |
|
119 | |||
118 | * The :class:`MultiEngineClient` class. |
|
120 | * The :class:`MultiEngineClient` class. | |
119 | * The :class:`TaskClient` class. |
|
121 | * The :class:`TaskClient` class. | |
120 |
|
122 | |||
121 | Security |
|
123 | Security | |
122 | -------- |
|
124 | -------- | |
123 |
|
125 | |||
124 | By default (as long as `pyOpenSSL` is installed) all network connections between the controller and engines and the controller and clients are secure. What does this mean? First of all, all of the connections will be encrypted using SSL. Second, the connections are authenticated. We handle authentication in a capability based security model [Capability]_. In this model, a "capability (known in some systems as a key) is a communicable, unforgeable token of authority". Put simply, a capability is like a key to your house. If you have the key to your house, you can get in. If not, you can't. |
|
126 | By default (as long as `pyOpenSSL` is installed) all network connections | |
125 |
|
127 | between the controller and engines and the controller and clients are secure. | ||
126 | In our architecture, the controller is the only process that listens on network ports, and is thus responsible to creating these keys. In IPython, these keys are known as Foolscap URLs, or FURLs, because of the underlying network protocol we are using. As a user, you don't need to know anything about the details of these FURLs, other than that when the controller starts, it saves a set of FURLs to files named :file:`something.furl`. The default location of these files is the :file:`~./ipython/security` directory. |
|
128 | What does this mean? First of all, all of the connections will be encrypted | |
127 |
|
129 | using SSL. Second, the connections are authenticated. We handle authentication | ||
128 | To connect and authenticate to the controller an engine or client simply needs to present an appropriate FURL (that was originally created by the controller) to the controller. Thus, the FURL files need to be copied to a location where the clients and engines can find them. Typically, this is the :file:`~./ipython/security` directory on the host where the client/engine is running (which could be a different host than the controller). Once the FURL files are copied over, everything should work fine. |
|
130 | in a capability based security model [Capability]_. In this model, a | |
|
131 | "capability (known in some systems as a key) is a communicable, unforgeable | |||
|
132 | token of authority". Put simply, a capability is like a key to your house. If | |||
|
133 | you have the key to your house, you can get in. If not, you can't. | |||
|
134 | ||||
|
135 | In our architecture, the controller is the only process that listens on | |||
|
136 | network ports, and is thus responsible to creating these keys. In IPython, | |||
|
137 | these keys are known as Foolscap URLs, or FURLs, because of the underlying | |||
|
138 | network protocol we are using. As a user, you don't need to know anything | |||
|
139 | about the details of these FURLs, other than that when the controller starts, | |||
|
140 | it saves a set of FURLs to files named :file:`something.furl`. The default | |||
|
141 | location of these files is the :file:`~./ipython/security` directory. | |||
|
142 | ||||
|
143 | To connect and authenticate to the controller an engine or client simply needs | |||
|
144 | to present an appropriate FURL (that was originally created by the controller) | |||
|
145 | to the controller. Thus, the FURL files need to be copied to a location where | |||
|
146 | the clients and engines can find them. Typically, this is the | |||
|
147 | :file:`~./ipython/security` directory on the host where the client/engine is | |||
|
148 | running (which could be a different host than the controller). Once the FURL | |||
|
149 | files are copied over, everything should work fine. | |||
129 |
|
150 | |||
130 | Currently, there are three FURL files that the controller creates: |
|
151 | Currently, there are three FURL files that the controller creates: | |
131 |
|
152 | |||
132 | ipcontroller-engine.furl |
|
153 | ipcontroller-engine.furl | |
133 | This FURL file is the key that gives an engine the ability to connect |
|
154 | This FURL file is the key that gives an engine the ability to connect | |
134 | to a controller. |
|
155 | to a controller. | |
135 |
|
156 | |||
136 | ipcontroller-tc.furl |
|
157 | ipcontroller-tc.furl | |
137 | This FURL file is the key that a :class:`TaskClient` must use to |
|
158 | This FURL file is the key that a :class:`TaskClient` must use to | |
138 | connect to the task interface of a controller. |
|
159 | connect to the task interface of a controller. | |
139 |
|
160 | |||
140 | ipcontroller-mec.furl |
|
161 | ipcontroller-mec.furl | |
141 | This FURL file is the key that a :class:`MultiEngineClient` must use |
|
162 | This FURL file is the key that a :class:`MultiEngineClient` must use | |
142 | to connect to the multiengine interface of a controller. |
|
163 | to connect to the multiengine interface of a controller. | |
143 |
|
164 | |||
144 | More details of how these FURL files are used are given below. |
|
165 | More details of how these FURL files are used are given below. | |
145 |
|
166 | |||
146 | A detailed description of the security model and its implementation in IPython |
|
167 | A detailed description of the security model and its implementation in IPython | |
147 | can be found :ref:`here <parallelsecurity>`. |
|
168 | can be found :ref:`here <parallelsecurity>`. | |
148 |
|
169 | |||
149 | Getting Started |
|
170 | Getting Started | |
150 | =============== |
|
171 | =============== | |
151 |
|
172 | |||
152 | To use IPython for parallel computing, you need to start one instance of |
|
173 | To use IPython for parallel computing, you need to start one instance of the | |
153 | the controller and one or more instances of the engine. Initially, it is best to simply start a controller and engines on a single host using the :command:`ipcluster` command. To start a controller and 4 engines on you localhost, just do:: |
|
174 | controller and one or more instances of the engine. Initially, it is best to | |
|
175 | simply start a controller and engines on a single host using the | |||
|
176 | :command:`ipcluster` command. To start a controller and 4 engines on your | |||
|
177 | localhost, just do:: | |||
154 |
|
178 | |||
155 | $ ipcluster local -n 4 |
|
179 | $ ipcluster local -n 4 | |
156 |
|
180 | |||
157 |
More details about starting the IPython controller and engines can be found |
|
181 | More details about starting the IPython controller and engines can be found | |
|
182 | :ref:`here <parallel_process>` | |||
158 |
|
183 | |||
159 | Once you have started the IPython controller and one or more engines, you |
|
184 | Once you have started the IPython controller and one or more engines, you | |
160 | are ready to use the engines to do something useful. To make sure |
|
185 | are ready to use the engines to do something useful. To make sure | |
161 | everything is working correctly, try the following commands: |
|
186 | everything is working correctly, try the following commands: | |
162 |
|
187 | |||
163 | .. sourcecode:: ipython |
|
188 | .. sourcecode:: ipython | |
164 |
|
189 | |||
165 | In [1]: from IPython.kernel import client |
|
190 | In [1]: from IPython.kernel import client | |
166 |
|
191 | |||
167 | In [2]: mec = client.MultiEngineClient() |
|
192 | In [2]: mec = client.MultiEngineClient() | |
168 |
|
193 | |||
169 | In [4]: mec.get_ids() |
|
194 | In [4]: mec.get_ids() | |
170 | Out[4]: [0, 1, 2, 3] |
|
195 | Out[4]: [0, 1, 2, 3] | |
171 |
|
196 | |||
172 | In [5]: mec.execute('print "Hello World"') |
|
197 | In [5]: mec.execute('print "Hello World"') | |
173 | Out[5]: |
|
198 | Out[5]: | |
174 | <Results List> |
|
199 | <Results List> | |
175 | [0] In [1]: print "Hello World" |
|
200 | [0] In [1]: print "Hello World" | |
176 | [0] Out[1]: Hello World |
|
201 | [0] Out[1]: Hello World | |
177 |
|
202 | |||
178 | [1] In [1]: print "Hello World" |
|
203 | [1] In [1]: print "Hello World" | |
179 | [1] Out[1]: Hello World |
|
204 | [1] Out[1]: Hello World | |
180 |
|
205 | |||
181 | [2] In [1]: print "Hello World" |
|
206 | [2] In [1]: print "Hello World" | |
182 | [2] Out[1]: Hello World |
|
207 | [2] Out[1]: Hello World | |
183 |
|
208 | |||
184 | [3] In [1]: print "Hello World" |
|
209 | [3] In [1]: print "Hello World" | |
185 | [3] Out[1]: Hello World |
|
210 | [3] Out[1]: Hello World | |
186 |
|
211 | |||
187 | Remember, a client also needs to present a FURL file to the controller. How does this happen? When a multiengine client is created with no arguments, the client tries to find the corresponding FURL file in the local :file:`~./ipython/security` directory. If it finds it, you are set. If you have put the FURL file in a different location or it has a different name, create the client like this:: |
|
212 | Remember, a client also needs to present a FURL file to the controller. How | |
|
213 | does this happen? When a multiengine client is created with no arguments, the | |||
|
214 | client tries to find the corresponding FURL file in the local | |||
|
215 | :file:`~./ipython/security` directory. If it finds it, you are set. If you | |||
|
216 | have put the FURL file in a different location or it has a different name, | |||
|
217 | create the client like this:: | |||
188 |
|
218 | |||
189 | mec = client.MultiEngineClient('/path/to/my/ipcontroller-mec.furl') |
|
219 | mec = client.MultiEngineClient('/path/to/my/ipcontroller-mec.furl') | |
190 |
|
220 | |||
191 | Same thing hold true of creating a task client:: |
|
221 | Same thing hold true of creating a task client:: | |
192 |
|
222 | |||
193 | tc = client.TaskClient('/path/to/my/ipcontroller-tc.furl') |
|
223 | tc = client.TaskClient('/path/to/my/ipcontroller-tc.furl') | |
194 |
|
224 | |||
195 | You are now ready to learn more about the :ref:`MultiEngine <parallelmultiengine>` and :ref:`Task <paralleltask>` interfaces to the controller. |
|
225 | You are now ready to learn more about the :ref:`MultiEngine | |
|
226 | <parallelmultiengine>` and :ref:`Task <paralleltask>` interfaces to the | |||
|
227 | controller. | |||
196 |
|
228 | |||
197 | .. note:: |
|
229 | .. note:: | |
198 |
|
230 | |||
199 | Don't forget that the engine, multiengine client and task client all have |
|
231 | Don't forget that the engine, multiengine client and task client all have | |
200 | *different* furl files. You must move *each* of these around to an |
|
232 | *different* furl files. You must move *each* of these around to an | |
201 | appropriate location so that the engines and clients can use them to |
|
233 | appropriate location so that the engines and clients can use them to | |
202 | connect to the controller. |
|
234 | connect to the controller. | |
203 |
|
235 | |||
204 | .. [Capability] Capability-based security, http://en.wikipedia.org/wiki/Capability-based_security |
|
236 | .. [Capability] Capability-based security, http://en.wikipedia.org/wiki/Capability-based_security | |
205 |
|
237 |
@@ -1,157 +1,176 b'' | |||||
1 | .. _parallelmpi: |
|
1 | .. _parallelmpi: | |
2 |
|
2 | |||
3 | ======================= |
|
3 | ======================= | |
4 | Using MPI with IPython |
|
4 | Using MPI with IPython | |
5 | ======================= |
|
5 | ======================= | |
6 |
|
6 | |||
7 | Often, a parallel algorithm will require moving data between the engines. One way of accomplishing this is by doing a pull and then a push using the multiengine client. However, this will be slow as all the data has to go through the controller to the client and then back through the controller, to its final destination. |
|
7 | Often, a parallel algorithm will require moving data between the engines. One | |
|
8 | way of accomplishing this is by doing a pull and then a push using the | |||
|
9 | multiengine client. However, this will be slow as all the data has to go | |||
|
10 | through the controller to the client and then back through the controller, to | |||
|
11 | its final destination. | |||
8 |
|
12 | |||
9 | A much better way of moving data between engines is to use a message passing library, such as the Message Passing Interface (MPI) [MPI]_. IPython's parallel computing architecture has been designed from the ground up to integrate with MPI. This document describes how to use MPI with IPython. |
|
13 | A much better way of moving data between engines is to use a message passing | |
|
14 | library, such as the Message Passing Interface (MPI) [MPI]_. IPython's | |||
|
15 | parallel computing architecture has been designed from the ground up to | |||
|
16 | integrate with MPI. This document describes how to use MPI with IPython. | |||
10 |
|
17 | |||
11 | Additional installation requirements |
|
18 | Additional installation requirements | |
12 | ==================================== |
|
19 | ==================================== | |
13 |
|
20 | |||
14 | If you want to use MPI with IPython, you will need to install: |
|
21 | If you want to use MPI with IPython, you will need to install: | |
15 |
|
22 | |||
16 | * A standard MPI implementation such as OpenMPI [OpenMPI]_ or MPICH. |
|
23 | * A standard MPI implementation such as OpenMPI [OpenMPI]_ or MPICH. | |
17 | * The mpi4py [mpi4py]_ package. |
|
24 | * The mpi4py [mpi4py]_ package. | |
18 |
|
25 | |||
19 | .. note:: |
|
26 | .. note:: | |
20 |
|
27 | |||
21 | The mpi4py package is not a strict requirement. However, you need to |
|
28 | The mpi4py package is not a strict requirement. However, you need to | |
22 | have *some* way of calling MPI from Python. You also need some way of |
|
29 | have *some* way of calling MPI from Python. You also need some way of | |
23 | making sure that :func:`MPI_Init` is called when the IPython engines start |
|
30 | making sure that :func:`MPI_Init` is called when the IPython engines start | |
24 | up. There are a number of ways of doing this and a good number of |
|
31 | up. There are a number of ways of doing this and a good number of | |
25 | associated subtleties. We highly recommend just using mpi4py as it |
|
32 | associated subtleties. We highly recommend just using mpi4py as it | |
26 | takes care of most of these problems. If you want to do something |
|
33 | takes care of most of these problems. If you want to do something | |
27 | different, let us know and we can help you get started. |
|
34 | different, let us know and we can help you get started. | |
28 |
|
35 | |||
29 | Starting the engines with MPI enabled |
|
36 | Starting the engines with MPI enabled | |
30 | ===================================== |
|
37 | ===================================== | |
31 |
|
38 | |||
32 | To use code that calls MPI, there are typically two things that MPI requires. |
|
39 | To use code that calls MPI, there are typically two things that MPI requires. | |
33 |
|
40 | |||
34 | 1. The process that wants to call MPI must be started using |
|
41 | 1. The process that wants to call MPI must be started using | |
35 | :command:`mpiexec` or a batch system (like PBS) that has MPI support. |
|
42 | :command:`mpiexec` or a batch system (like PBS) that has MPI support. | |
36 | 2. Once the process starts, it must call :func:`MPI_Init`. |
|
43 | 2. Once the process starts, it must call :func:`MPI_Init`. | |
37 |
|
44 | |||
38 |
There are a couple of ways that you can start the IPython engines and get |
|
45 | There are a couple of ways that you can start the IPython engines and get | |
|
46 | these things to happen. | |||
39 |
|
47 | |||
40 | Automatic starting using :command:`mpiexec` and :command:`ipcluster` |
|
48 | Automatic starting using :command:`mpiexec` and :command:`ipcluster` | |
41 | -------------------------------------------------------------------- |
|
49 | -------------------------------------------------------------------- | |
42 |
|
50 | |||
43 |
The easiest approach is to use the `mpiexec` mode of :command:`ipcluster`, |
|
51 | The easiest approach is to use the `mpiexec` mode of :command:`ipcluster`, | |
|
52 | which will first start a controller and then a set of engines using | |||
|
53 | :command:`mpiexec`:: | |||
44 |
|
54 | |||
45 | $ ipcluster mpiexec -n 4 |
|
55 | $ ipcluster mpiexec -n 4 | |
46 |
|
56 | |||
47 | This approach is best as interrupting :command:`ipcluster` will automatically |
|
57 | This approach is best as interrupting :command:`ipcluster` will automatically | |
48 | stop and clean up the controller and engines. |
|
58 | stop and clean up the controller and engines. | |
49 |
|
59 | |||
50 | Manual starting using :command:`mpiexec` |
|
60 | Manual starting using :command:`mpiexec` | |
51 | ---------------------------------------- |
|
61 | ---------------------------------------- | |
52 |
|
62 | |||
53 |
If you want to start the IPython engines using the :command:`mpiexec`, just |
|
63 | If you want to start the IPython engines using the :command:`mpiexec`, just | |
|
64 | do:: | |||
54 |
|
65 | |||
55 | $ mpiexec -n 4 ipengine --mpi=mpi4py |
|
66 | $ mpiexec -n 4 ipengine --mpi=mpi4py | |
56 |
|
67 | |||
57 | This requires that you already have a controller running and that the FURL |
|
68 | This requires that you already have a controller running and that the FURL | |
58 | files for the engines are in place. We also have built in support for |
|
69 | files for the engines are in place. We also have built in support for | |
59 | PyTrilinos [PyTrilinos]_, which can be used (assuming is installed) by |
|
70 | PyTrilinos [PyTrilinos]_, which can be used (assuming is installed) by | |
60 | starting the engines with:: |
|
71 | starting the engines with:: | |
61 |
|
72 | |||
62 | mpiexec -n 4 ipengine --mpi=pytrilinos |
|
73 | mpiexec -n 4 ipengine --mpi=pytrilinos | |
63 |
|
74 | |||
64 | Automatic starting using PBS and :command:`ipcluster` |
|
75 | Automatic starting using PBS and :command:`ipcluster` | |
65 | ----------------------------------------------------- |
|
76 | ----------------------------------------------------- | |
66 |
|
77 | |||
67 |
The :command:`ipcluster` command also has built-in integration with PBS. For |
|
78 | The :command:`ipcluster` command also has built-in integration with PBS. For | |
|
79 | more information on this approach, see our documentation on :ref:`ipcluster | |||
|
80 | <parallel_process>`. | |||
68 |
|
81 | |||
69 | Actually using MPI |
|
82 | Actually using MPI | |
70 | ================== |
|
83 | ================== | |
71 |
|
84 | |||
72 | Once the engines are running with MPI enabled, you are ready to go. You can now call any code that uses MPI in the IPython engines. And, all of this can be done interactively. Here we show a simple example that uses mpi4py [mpi4py]_. |
|
85 | Once the engines are running with MPI enabled, you are ready to go. You can | |
|
86 | now call any code that uses MPI in the IPython engines. And, all of this can | |||
|
87 | be done interactively. Here we show a simple example that uses mpi4py | |||
|
88 | [mpi4py]_. | |||
73 |
|
89 | |||
74 |
First, lets define a simply function that uses MPI to calculate the sum of a |
|
90 | First, lets define a simply function that uses MPI to calculate the sum of a | |
|
91 | distributed array. Save the following text in a file called :file:`psum.py`: | |||
75 |
|
92 | |||
76 | .. sourcecode:: python |
|
93 | .. sourcecode:: python | |
77 |
|
94 | |||
78 | from mpi4py import MPI |
|
95 | from mpi4py import MPI | |
79 | import numpy as np |
|
96 | import numpy as np | |
80 |
|
97 | |||
81 | def psum(a): |
|
98 | def psum(a): | |
82 | s = np.sum(a) |
|
99 | s = np.sum(a) | |
83 | return MPI.COMM_WORLD.Allreduce(s,MPI.SUM) |
|
100 | return MPI.COMM_WORLD.Allreduce(s,MPI.SUM) | |
84 |
|
101 | |||
85 | Now, start an IPython cluster in the same directory as :file:`psum.py`:: |
|
102 | Now, start an IPython cluster in the same directory as :file:`psum.py`:: | |
86 |
|
103 | |||
87 | $ ipcluster mpiexec -n 4 |
|
104 | $ ipcluster mpiexec -n 4 | |
88 |
|
105 | |||
89 | Finally, connect to the cluster and use this function interactively. In this case, we create a random array on each engine and sum up all the random arrays using our :func:`psum` function: |
|
106 | Finally, connect to the cluster and use this function interactively. In this | |
|
107 | case, we create a random array on each engine and sum up all the random arrays | |||
|
108 | using our :func:`psum` function: | |||
90 |
|
109 | |||
91 | .. sourcecode:: ipython |
|
110 | .. sourcecode:: ipython | |
92 |
|
111 | |||
93 | In [1]: from IPython.kernel import client |
|
112 | In [1]: from IPython.kernel import client | |
94 |
|
113 | |||
95 | In [2]: mec = client.MultiEngineClient() |
|
114 | In [2]: mec = client.MultiEngineClient() | |
96 |
|
115 | |||
97 | In [3]: mec.activate() |
|
116 | In [3]: mec.activate() | |
98 |
|
117 | |||
99 | In [4]: px import numpy as np |
|
118 | In [4]: px import numpy as np | |
100 | Parallel execution on engines: all |
|
119 | Parallel execution on engines: all | |
101 | Out[4]: |
|
120 | Out[4]: | |
102 | <Results List> |
|
121 | <Results List> | |
103 | [0] In [13]: import numpy as np |
|
122 | [0] In [13]: import numpy as np | |
104 | [1] In [13]: import numpy as np |
|
123 | [1] In [13]: import numpy as np | |
105 | [2] In [13]: import numpy as np |
|
124 | [2] In [13]: import numpy as np | |
106 | [3] In [13]: import numpy as np |
|
125 | [3] In [13]: import numpy as np | |
107 |
|
126 | |||
108 | In [6]: px a = np.random.rand(100) |
|
127 | In [6]: px a = np.random.rand(100) | |
109 | Parallel execution on engines: all |
|
128 | Parallel execution on engines: all | |
110 | Out[6]: |
|
129 | Out[6]: | |
111 | <Results List> |
|
130 | <Results List> | |
112 | [0] In [15]: a = np.random.rand(100) |
|
131 | [0] In [15]: a = np.random.rand(100) | |
113 | [1] In [15]: a = np.random.rand(100) |
|
132 | [1] In [15]: a = np.random.rand(100) | |
114 | [2] In [15]: a = np.random.rand(100) |
|
133 | [2] In [15]: a = np.random.rand(100) | |
115 | [3] In [15]: a = np.random.rand(100) |
|
134 | [3] In [15]: a = np.random.rand(100) | |
116 |
|
135 | |||
117 | In [7]: px from psum import psum |
|
136 | In [7]: px from psum import psum | |
118 | Parallel execution on engines: all |
|
137 | Parallel execution on engines: all | |
119 | Out[7]: |
|
138 | Out[7]: | |
120 | <Results List> |
|
139 | <Results List> | |
121 | [0] In [16]: from psum import psum |
|
140 | [0] In [16]: from psum import psum | |
122 | [1] In [16]: from psum import psum |
|
141 | [1] In [16]: from psum import psum | |
123 | [2] In [16]: from psum import psum |
|
142 | [2] In [16]: from psum import psum | |
124 | [3] In [16]: from psum import psum |
|
143 | [3] In [16]: from psum import psum | |
125 |
|
144 | |||
126 | In [8]: px s = psum(a) |
|
145 | In [8]: px s = psum(a) | |
127 | Parallel execution on engines: all |
|
146 | Parallel execution on engines: all | |
128 | Out[8]: |
|
147 | Out[8]: | |
129 | <Results List> |
|
148 | <Results List> | |
130 | [0] In [17]: s = psum(a) |
|
149 | [0] In [17]: s = psum(a) | |
131 | [1] In [17]: s = psum(a) |
|
150 | [1] In [17]: s = psum(a) | |
132 | [2] In [17]: s = psum(a) |
|
151 | [2] In [17]: s = psum(a) | |
133 | [3] In [17]: s = psum(a) |
|
152 | [3] In [17]: s = psum(a) | |
134 |
|
153 | |||
135 | In [9]: px print s |
|
154 | In [9]: px print s | |
136 | Parallel execution on engines: all |
|
155 | Parallel execution on engines: all | |
137 | Out[9]: |
|
156 | Out[9]: | |
138 | <Results List> |
|
157 | <Results List> | |
139 | [0] In [18]: print s |
|
158 | [0] In [18]: print s | |
140 | [0] Out[18]: 187.451545803 |
|
159 | [0] Out[18]: 187.451545803 | |
141 |
|
160 | |||
142 | [1] In [18]: print s |
|
161 | [1] In [18]: print s | |
143 | [1] Out[18]: 187.451545803 |
|
162 | [1] Out[18]: 187.451545803 | |
144 |
|
163 | |||
145 | [2] In [18]: print s |
|
164 | [2] In [18]: print s | |
146 | [2] Out[18]: 187.451545803 |
|
165 | [2] Out[18]: 187.451545803 | |
147 |
|
166 | |||
148 | [3] In [18]: print s |
|
167 | [3] In [18]: print s | |
149 | [3] Out[18]: 187.451545803 |
|
168 | [3] Out[18]: 187.451545803 | |
150 |
|
169 | |||
151 | Any Python code that makes calls to MPI can be used in this manner, including |
|
170 | Any Python code that makes calls to MPI can be used in this manner, including | |
152 | compiled C, C++ and Fortran libraries that have been exposed to Python. |
|
171 | compiled C, C++ and Fortran libraries that have been exposed to Python. | |
153 |
|
172 | |||
154 | .. [MPI] Message Passing Interface. http://www-unix.mcs.anl.gov/mpi/ |
|
173 | .. [MPI] Message Passing Interface. http://www-unix.mcs.anl.gov/mpi/ | |
155 | .. [mpi4py] MPI for Python. mpi4py: http://mpi4py.scipy.org/ |
|
174 | .. [mpi4py] MPI for Python. mpi4py: http://mpi4py.scipy.org/ | |
156 | .. [OpenMPI] Open MPI. http://www.open-mpi.org/ |
|
175 | .. [OpenMPI] Open MPI. http://www.open-mpi.org/ | |
157 | .. [PyTrilinos] PyTrilinos. http://trilinos.sandia.gov/packages/pytrilinos/ |
|
176 | .. [PyTrilinos] PyTrilinos. http://trilinos.sandia.gov/packages/pytrilinos/ |
@@ -1,828 +1,835 b'' | |||||
1 | .. _parallelmultiengine: |
|
1 | .. _parallelmultiengine: | |
2 |
|
2 | |||
3 | =============================== |
|
3 | =============================== | |
4 | IPython's multiengine interface |
|
4 | IPython's multiengine interface | |
5 | =============================== |
|
5 | =============================== | |
6 |
|
6 | |||
7 | The multiengine interface represents one possible way of working with a set of |
|
7 | The multiengine interface represents one possible way of working with a set of | |
8 | IPython engines. The basic idea behind the multiengine interface is that the |
|
8 | IPython engines. The basic idea behind the multiengine interface is that the | |
9 | capabilities of each engine are directly and explicitly exposed to the user. |
|
9 | capabilities of each engine are directly and explicitly exposed to the user. | |
10 | Thus, in the multiengine interface, each engine is given an id that is used to |
|
10 | Thus, in the multiengine interface, each engine is given an id that is used to | |
11 | identify the engine and give it work to do. This interface is very intuitive |
|
11 | identify the engine and give it work to do. This interface is very intuitive | |
12 | and is designed with interactive usage in mind, and is thus the best place for |
|
12 | and is designed with interactive usage in mind, and is thus the best place for | |
13 | new users of IPython to begin. |
|
13 | new users of IPython to begin. | |
14 |
|
14 | |||
15 | Starting the IPython controller and engines |
|
15 | Starting the IPython controller and engines | |
16 | =========================================== |
|
16 | =========================================== | |
17 |
|
17 | |||
18 | To follow along with this tutorial, you will need to start the IPython |
|
18 | To follow along with this tutorial, you will need to start the IPython | |
19 | controller and four IPython engines. The simplest way of doing this is to use |
|
19 | controller and four IPython engines. The simplest way of doing this is to use | |
20 | the :command:`ipcluster` command:: |
|
20 | the :command:`ipcluster` command:: | |
21 |
|
21 | |||
22 | $ ipcluster local -n 4 |
|
22 | $ ipcluster local -n 4 | |
23 |
|
23 | |||
24 | For more detailed information about starting the controller and engines, see |
|
24 | For more detailed information about starting the controller and engines, see | |
25 | our :ref:`introduction <ip1par>` to using IPython for parallel computing. |
|
25 | our :ref:`introduction <ip1par>` to using IPython for parallel computing. | |
26 |
|
26 | |||
27 | Creating a ``MultiEngineClient`` instance |
|
27 | Creating a ``MultiEngineClient`` instance | |
28 | ========================================= |
|
28 | ========================================= | |
29 |
|
29 | |||
30 | The first step is to import the IPython :mod:`IPython.kernel.client` module |
|
30 | The first step is to import the IPython :mod:`IPython.kernel.client` module | |
31 | and then create a :class:`MultiEngineClient` instance: |
|
31 | and then create a :class:`MultiEngineClient` instance: | |
32 |
|
32 | |||
33 | .. sourcecode:: ipython |
|
33 | .. sourcecode:: ipython | |
34 |
|
34 | |||
35 | In [1]: from IPython.kernel import client |
|
35 | In [1]: from IPython.kernel import client | |
36 |
|
36 | |||
37 | In [2]: mec = client.MultiEngineClient() |
|
37 | In [2]: mec = client.MultiEngineClient() | |
38 |
|
38 | |||
39 | This form assumes that the :file:`ipcontroller-mec.furl` is in the |
|
39 | This form assumes that the :file:`ipcontroller-mec.furl` is in the | |
40 | :file:`~./ipython/security` directory on the client's host. If not, the |
|
40 | :file:`~./ipython/security` directory on the client's host. If not, the | |
41 | location of the FURL file must be given as an argument to the |
|
41 | location of the FURL file must be given as an argument to the | |
42 | constructor: |
|
42 | constructor: | |
43 |
|
43 | |||
44 | .. sourcecode:: ipython |
|
44 | .. sourcecode:: ipython | |
45 |
|
45 | |||
46 | In [2]: mec = client.MultiEngineClient('/path/to/my/ipcontroller-mec.furl') |
|
46 | In [2]: mec = client.MultiEngineClient('/path/to/my/ipcontroller-mec.furl') | |
47 |
|
47 | |||
48 | To make sure there are engines connected to the controller, use can get a list |
|
48 | To make sure there are engines connected to the controller, use can get a list | |
49 | of engine ids: |
|
49 | of engine ids: | |
50 |
|
50 | |||
51 | .. sourcecode:: ipython |
|
51 | .. sourcecode:: ipython | |
52 |
|
52 | |||
53 | In [3]: mec.get_ids() |
|
53 | In [3]: mec.get_ids() | |
54 | Out[3]: [0, 1, 2, 3] |
|
54 | Out[3]: [0, 1, 2, 3] | |
55 |
|
55 | |||
56 | Here we see that there are four engines ready to do work for us. |
|
56 | Here we see that there are four engines ready to do work for us. | |
57 |
|
57 | |||
58 | Quick and easy parallelism |
|
58 | Quick and easy parallelism | |
59 | ========================== |
|
59 | ========================== | |
60 |
|
60 | |||
61 | In many cases, you simply want to apply a Python function to a sequence of objects, but *in parallel*. The multiengine interface provides two simple ways of accomplishing this: a parallel version of :func:`map` and ``@parallel`` function decorator. |
|
61 | In many cases, you simply want to apply a Python function to a sequence of | |
|
62 | objects, but *in parallel*. The multiengine interface provides two simple ways | |||
|
63 | of accomplishing this: a parallel version of :func:`map` and ``@parallel`` | |||
|
64 | function decorator. | |||
62 |
|
65 | |||
63 | Parallel map |
|
66 | Parallel map | |
64 | ------------ |
|
67 | ------------ | |
65 |
|
68 | |||
66 | Python's builtin :func:`map` functions allows a function to be applied to a |
|
69 | Python's builtin :func:`map` functions allows a function to be applied to a | |
67 | sequence element-by-element. This type of code is typically trivial to |
|
70 | sequence element-by-element. This type of code is typically trivial to | |
68 | parallelize. In fact, the multiengine interface in IPython already has a |
|
71 | parallelize. In fact, the multiengine interface in IPython already has a | |
69 | parallel version of :meth:`map` that works just like its serial counterpart: |
|
72 | parallel version of :meth:`map` that works just like its serial counterpart: | |
70 |
|
73 | |||
71 | .. sourcecode:: ipython |
|
74 | .. sourcecode:: ipython | |
72 |
|
75 | |||
73 | In [63]: serial_result = map(lambda x:x**10, range(32)) |
|
76 | In [63]: serial_result = map(lambda x:x**10, range(32)) | |
74 |
|
77 | |||
75 | In [64]: parallel_result = mec.map(lambda x:x**10, range(32)) |
|
78 | In [64]: parallel_result = mec.map(lambda x:x**10, range(32)) | |
76 |
|
79 | |||
77 | In [65]: serial_result==parallel_result |
|
80 | In [65]: serial_result==parallel_result | |
78 | Out[65]: True |
|
81 | Out[65]: True | |
79 |
|
82 | |||
80 | .. note:: |
|
83 | .. note:: | |
81 |
|
84 | |||
82 | The multiengine interface version of :meth:`map` does not do any load |
|
85 | The multiengine interface version of :meth:`map` does not do any load | |
83 | balancing. For a load balanced version, see the task interface. |
|
86 | balancing. For a load balanced version, see the task interface. | |
84 |
|
87 | |||
85 | .. seealso:: |
|
88 | .. seealso:: | |
86 |
|
89 | |||
87 | The :meth:`map` method has a number of options that can be controlled by |
|
90 | The :meth:`map` method has a number of options that can be controlled by | |
88 | the :meth:`mapper` method. See its docstring for more information. |
|
91 | the :meth:`mapper` method. See its docstring for more information. | |
89 |
|
92 | |||
90 | Parallel function decorator |
|
93 | Parallel function decorator | |
91 | --------------------------- |
|
94 | --------------------------- | |
92 |
|
95 | |||
93 | Parallel functions are just like normal function, but they can be called on sequences and *in parallel*. The multiengine interface provides a decorator that turns any Python function into a parallel function: |
|
96 | Parallel functions are just like normal function, but they can be called on | |
|
97 | sequences and *in parallel*. The multiengine interface provides a decorator | |||
|
98 | that turns any Python function into a parallel function: | |||
94 |
|
99 | |||
95 | .. sourcecode:: ipython |
|
100 | .. sourcecode:: ipython | |
96 |
|
101 | |||
97 | In [10]: @mec.parallel() |
|
102 | In [10]: @mec.parallel() | |
98 | ....: def f(x): |
|
103 | ....: def f(x): | |
99 | ....: return 10.0*x**4 |
|
104 | ....: return 10.0*x**4 | |
100 | ....: |
|
105 | ....: | |
101 |
|
106 | |||
102 | In [11]: f(range(32)) # this is done in parallel |
|
107 | In [11]: f(range(32)) # this is done in parallel | |
103 | Out[11]: |
|
108 | Out[11]: | |
104 | [0.0,10.0,160.0,...] |
|
109 | [0.0,10.0,160.0,...] | |
105 |
|
110 | |||
106 | See the docstring for the :meth:`parallel` decorator for options. |
|
111 | See the docstring for the :meth:`parallel` decorator for options. | |
107 |
|
112 | |||
108 | Running Python commands |
|
113 | Running Python commands | |
109 | ======================= |
|
114 | ======================= | |
110 |
|
115 | |||
111 | The most basic type of operation that can be performed on the engines is to |
|
116 | The most basic type of operation that can be performed on the engines is to | |
112 | execute Python code. Executing Python code can be done in blocking or |
|
117 | execute Python code. Executing Python code can be done in blocking or | |
113 | non-blocking mode (blocking is default) using the :meth:`execute` method. |
|
118 | non-blocking mode (blocking is default) using the :meth:`execute` method. | |
114 |
|
119 | |||
115 | Blocking execution |
|
120 | Blocking execution | |
116 | ------------------ |
|
121 | ------------------ | |
117 |
|
122 | |||
118 | In blocking mode, the :class:`MultiEngineClient` object (called ``mec`` in |
|
123 | In blocking mode, the :class:`MultiEngineClient` object (called ``mec`` in | |
119 | these examples) submits the command to the controller, which places the |
|
124 | these examples) submits the command to the controller, which places the | |
120 | command in the engines' queues for execution. The :meth:`execute` call then |
|
125 | command in the engines' queues for execution. The :meth:`execute` call then | |
121 | blocks until the engines are done executing the command: |
|
126 | blocks until the engines are done executing the command: | |
122 |
|
127 | |||
123 | .. sourcecode:: ipython |
|
128 | .. sourcecode:: ipython | |
124 |
|
129 | |||
125 | # The default is to run on all engines |
|
130 | # The default is to run on all engines | |
126 | In [4]: mec.execute('a=5') |
|
131 | In [4]: mec.execute('a=5') | |
127 | Out[4]: |
|
132 | Out[4]: | |
128 | <Results List> |
|
133 | <Results List> | |
129 | [0] In [1]: a=5 |
|
134 | [0] In [1]: a=5 | |
130 | [1] In [1]: a=5 |
|
135 | [1] In [1]: a=5 | |
131 | [2] In [1]: a=5 |
|
136 | [2] In [1]: a=5 | |
132 | [3] In [1]: a=5 |
|
137 | [3] In [1]: a=5 | |
133 |
|
138 | |||
134 | In [5]: mec.execute('b=10') |
|
139 | In [5]: mec.execute('b=10') | |
135 | Out[5]: |
|
140 | Out[5]: | |
136 | <Results List> |
|
141 | <Results List> | |
137 | [0] In [2]: b=10 |
|
142 | [0] In [2]: b=10 | |
138 | [1] In [2]: b=10 |
|
143 | [1] In [2]: b=10 | |
139 | [2] In [2]: b=10 |
|
144 | [2] In [2]: b=10 | |
140 | [3] In [2]: b=10 |
|
145 | [3] In [2]: b=10 | |
141 |
|
146 | |||
142 | Python commands can be executed on specific engines by calling execute using |
|
147 | Python commands can be executed on specific engines by calling execute using | |
143 | the ``targets`` keyword argument: |
|
148 | the ``targets`` keyword argument: | |
144 |
|
149 | |||
145 | .. sourcecode:: ipython |
|
150 | .. sourcecode:: ipython | |
146 |
|
151 | |||
147 | In [6]: mec.execute('c=a+b',targets=[0,2]) |
|
152 | In [6]: mec.execute('c=a+b',targets=[0,2]) | |
148 | Out[6]: |
|
153 | Out[6]: | |
149 | <Results List> |
|
154 | <Results List> | |
150 | [0] In [3]: c=a+b |
|
155 | [0] In [3]: c=a+b | |
151 | [2] In [3]: c=a+b |
|
156 | [2] In [3]: c=a+b | |
152 |
|
157 | |||
153 |
|
158 | |||
154 | In [7]: mec.execute('c=a-b',targets=[1,3]) |
|
159 | In [7]: mec.execute('c=a-b',targets=[1,3]) | |
155 | Out[7]: |
|
160 | Out[7]: | |
156 | <Results List> |
|
161 | <Results List> | |
157 | [1] In [3]: c=a-b |
|
162 | [1] In [3]: c=a-b | |
158 | [3] In [3]: c=a-b |
|
163 | [3] In [3]: c=a-b | |
159 |
|
164 | |||
160 |
|
165 | |||
161 | In [8]: mec.execute('print c') |
|
166 | In [8]: mec.execute('print c') | |
162 | Out[8]: |
|
167 | Out[8]: | |
163 | <Results List> |
|
168 | <Results List> | |
164 | [0] In [4]: print c |
|
169 | [0] In [4]: print c | |
165 | [0] Out[4]: 15 |
|
170 | [0] Out[4]: 15 | |
166 |
|
171 | |||
167 | [1] In [4]: print c |
|
172 | [1] In [4]: print c | |
168 | [1] Out[4]: -5 |
|
173 | [1] Out[4]: -5 | |
169 |
|
174 | |||
170 | [2] In [4]: print c |
|
175 | [2] In [4]: print c | |
171 | [2] Out[4]: 15 |
|
176 | [2] Out[4]: 15 | |
172 |
|
177 | |||
173 | [3] In [4]: print c |
|
178 | [3] In [4]: print c | |
174 | [3] Out[4]: -5 |
|
179 | [3] Out[4]: -5 | |
175 |
|
180 | |||
176 | This example also shows one of the most important things about the IPython |
|
181 | This example also shows one of the most important things about the IPython | |
177 | engines: they have a persistent user namespaces. The :meth:`execute` method |
|
182 | engines: they have a persistent user namespaces. The :meth:`execute` method | |
178 | returns a Python ``dict`` that contains useful information: |
|
183 | returns a Python ``dict`` that contains useful information: | |
179 |
|
184 | |||
180 | .. sourcecode:: ipython |
|
185 | .. sourcecode:: ipython | |
181 |
|
186 | |||
182 | In [9]: result_dict = mec.execute('d=10; print d') |
|
187 | In [9]: result_dict = mec.execute('d=10; print d') | |
183 |
|
188 | |||
184 | In [10]: for r in result_dict: |
|
189 | In [10]: for r in result_dict: | |
185 | ....: print r |
|
190 | ....: print r | |
186 | ....: |
|
191 | ....: | |
187 | ....: |
|
192 | ....: | |
188 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 0, 'stdout': '10\n'} |
|
193 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 0, 'stdout': '10\n'} | |
189 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 1, 'stdout': '10\n'} |
|
194 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 1, 'stdout': '10\n'} | |
190 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 2, 'stdout': '10\n'} |
|
195 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 2, 'stdout': '10\n'} | |
191 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 3, 'stdout': '10\n'} |
|
196 | {'input': {'translated': 'd=10; print d', 'raw': 'd=10; print d'}, 'number': 5, 'id': 3, 'stdout': '10\n'} | |
192 |
|
197 | |||
193 | Non-blocking execution |
|
198 | Non-blocking execution | |
194 | ---------------------- |
|
199 | ---------------------- | |
195 |
|
200 | |||
196 | In non-blocking mode, :meth:`execute` submits the command to be executed and |
|
201 | In non-blocking mode, :meth:`execute` submits the command to be executed and | |
197 | then returns a :class:`PendingResult` object immediately. The |
|
202 | then returns a :class:`PendingResult` object immediately. The | |
198 | :class:`PendingResult` object gives you a way of getting a result at a later |
|
203 | :class:`PendingResult` object gives you a way of getting a result at a later | |
199 | time through its :meth:`get_result` method or :attr:`r` attribute. This allows |
|
204 | time through its :meth:`get_result` method or :attr:`r` attribute. This allows | |
200 | you to quickly submit long running commands without blocking your local |
|
205 | you to quickly submit long running commands without blocking your local | |
201 | Python/IPython session: |
|
206 | Python/IPython session: | |
202 |
|
207 | |||
203 | .. sourcecode:: ipython |
|
208 | .. sourcecode:: ipython | |
204 |
|
209 | |||
205 | # In blocking mode |
|
210 | # In blocking mode | |
206 | In [6]: mec.execute('import time') |
|
211 | In [6]: mec.execute('import time') | |
207 | Out[6]: |
|
212 | Out[6]: | |
208 | <Results List> |
|
213 | <Results List> | |
209 | [0] In [1]: import time |
|
214 | [0] In [1]: import time | |
210 | [1] In [1]: import time |
|
215 | [1] In [1]: import time | |
211 | [2] In [1]: import time |
|
216 | [2] In [1]: import time | |
212 | [3] In [1]: import time |
|
217 | [3] In [1]: import time | |
213 |
|
218 | |||
214 | # In non-blocking mode |
|
219 | # In non-blocking mode | |
215 | In [7]: pr = mec.execute('time.sleep(10)',block=False) |
|
220 | In [7]: pr = mec.execute('time.sleep(10)',block=False) | |
216 |
|
221 | |||
217 | # Now block for the result |
|
222 | # Now block for the result | |
218 | In [8]: pr.get_result() |
|
223 | In [8]: pr.get_result() | |
219 | Out[8]: |
|
224 | Out[8]: | |
220 | <Results List> |
|
225 | <Results List> | |
221 | [0] In [2]: time.sleep(10) |
|
226 | [0] In [2]: time.sleep(10) | |
222 | [1] In [2]: time.sleep(10) |
|
227 | [1] In [2]: time.sleep(10) | |
223 | [2] In [2]: time.sleep(10) |
|
228 | [2] In [2]: time.sleep(10) | |
224 | [3] In [2]: time.sleep(10) |
|
229 | [3] In [2]: time.sleep(10) | |
225 |
|
230 | |||
226 | # Again in non-blocking mode |
|
231 | # Again in non-blocking mode | |
227 | In [9]: pr = mec.execute('time.sleep(10)',block=False) |
|
232 | In [9]: pr = mec.execute('time.sleep(10)',block=False) | |
228 |
|
233 | |||
229 | # Poll to see if the result is ready |
|
234 | # Poll to see if the result is ready | |
230 | In [10]: pr.get_result(block=False) |
|
235 | In [10]: pr.get_result(block=False) | |
231 |
|
236 | |||
232 | # A shorthand for get_result(block=True) |
|
237 | # A shorthand for get_result(block=True) | |
233 | In [11]: pr.r |
|
238 | In [11]: pr.r | |
234 | Out[11]: |
|
239 | Out[11]: | |
235 | <Results List> |
|
240 | <Results List> | |
236 | [0] In [3]: time.sleep(10) |
|
241 | [0] In [3]: time.sleep(10) | |
237 | [1] In [3]: time.sleep(10) |
|
242 | [1] In [3]: time.sleep(10) | |
238 | [2] In [3]: time.sleep(10) |
|
243 | [2] In [3]: time.sleep(10) | |
239 | [3] In [3]: time.sleep(10) |
|
244 | [3] In [3]: time.sleep(10) | |
240 |
|
245 | |||
241 | Often, it is desirable to wait until a set of :class:`PendingResult` objects |
|
246 | Often, it is desirable to wait until a set of :class:`PendingResult` objects | |
242 | are done. For this, there is a the method :meth:`barrier`. This method takes a |
|
247 | are done. For this, there is a the method :meth:`barrier`. This method takes a | |
243 | tuple of :class:`PendingResult` objects and blocks until all of the associated |
|
248 | tuple of :class:`PendingResult` objects and blocks until all of the associated | |
244 | results are ready: |
|
249 | results are ready: | |
245 |
|
250 | |||
246 | .. sourcecode:: ipython |
|
251 | .. sourcecode:: ipython | |
247 |
|
252 | |||
248 | In [72]: mec.block=False |
|
253 | In [72]: mec.block=False | |
249 |
|
254 | |||
250 | # A trivial list of PendingResults objects |
|
255 | # A trivial list of PendingResults objects | |
251 | In [73]: pr_list = [mec.execute('time.sleep(3)') for i in range(10)] |
|
256 | In [73]: pr_list = [mec.execute('time.sleep(3)') for i in range(10)] | |
252 |
|
257 | |||
253 | # Wait until all of them are done |
|
258 | # Wait until all of them are done | |
254 | In [74]: mec.barrier(pr_list) |
|
259 | In [74]: mec.barrier(pr_list) | |
255 |
|
260 | |||
256 | # Then, their results are ready using get_result or the r attribute |
|
261 | # Then, their results are ready using get_result or the r attribute | |
257 | In [75]: pr_list[0].r |
|
262 | In [75]: pr_list[0].r | |
258 | Out[75]: |
|
263 | Out[75]: | |
259 | <Results List> |
|
264 | <Results List> | |
260 | [0] In [20]: time.sleep(3) |
|
265 | [0] In [20]: time.sleep(3) | |
261 | [1] In [19]: time.sleep(3) |
|
266 | [1] In [19]: time.sleep(3) | |
262 | [2] In [20]: time.sleep(3) |
|
267 | [2] In [20]: time.sleep(3) | |
263 | [3] In [19]: time.sleep(3) |
|
268 | [3] In [19]: time.sleep(3) | |
264 |
|
269 | |||
265 |
|
270 | |||
266 | The ``block`` and ``targets`` keyword arguments and attributes |
|
271 | The ``block`` and ``targets`` keyword arguments and attributes | |
267 | -------------------------------------------------------------- |
|
272 | -------------------------------------------------------------- | |
268 |
|
273 | |||
269 | Most methods in the multiengine interface (like :meth:`execute`) accept |
|
274 | Most methods in the multiengine interface (like :meth:`execute`) accept | |
270 | ``block`` and ``targets`` as keyword arguments. As we have seen above, these |
|
275 | ``block`` and ``targets`` as keyword arguments. As we have seen above, these | |
271 | keyword arguments control the blocking mode and which engines the command is |
|
276 | keyword arguments control the blocking mode and which engines the command is | |
272 | applied to. The :class:`MultiEngineClient` class also has :attr:`block` and |
|
277 | applied to. The :class:`MultiEngineClient` class also has :attr:`block` and | |
273 | :attr:`targets` attributes that control the default behavior when the keyword |
|
278 | :attr:`targets` attributes that control the default behavior when the keyword | |
274 | arguments are not provided. Thus the following logic is used for :attr:`block` |
|
279 | arguments are not provided. Thus the following logic is used for :attr:`block` | |
275 | and :attr:`targets`: |
|
280 | and :attr:`targets`: | |
276 |
|
281 | |||
277 | * If no keyword argument is provided, the instance attributes are used. |
|
282 | * If no keyword argument is provided, the instance attributes are used. | |
278 | * Keyword argument, if provided override the instance attributes. |
|
283 | * Keyword argument, if provided override the instance attributes. | |
279 |
|
284 | |||
280 | The following examples demonstrate how to use the instance attributes: |
|
285 | The following examples demonstrate how to use the instance attributes: | |
281 |
|
286 | |||
282 | .. sourcecode:: ipython |
|
287 | .. sourcecode:: ipython | |
283 |
|
288 | |||
284 | In [16]: mec.targets = [0,2] |
|
289 | In [16]: mec.targets = [0,2] | |
285 |
|
290 | |||
286 | In [17]: mec.block = False |
|
291 | In [17]: mec.block = False | |
287 |
|
292 | |||
288 | In [18]: pr = mec.execute('a=5') |
|
293 | In [18]: pr = mec.execute('a=5') | |
289 |
|
294 | |||
290 | In [19]: pr.r |
|
295 | In [19]: pr.r | |
291 | Out[19]: |
|
296 | Out[19]: | |
292 | <Results List> |
|
297 | <Results List> | |
293 | [0] In [6]: a=5 |
|
298 | [0] In [6]: a=5 | |
294 | [2] In [6]: a=5 |
|
299 | [2] In [6]: a=5 | |
295 |
|
300 | |||
296 | # Note targets='all' means all engines |
|
301 | # Note targets='all' means all engines | |
297 | In [20]: mec.targets = 'all' |
|
302 | In [20]: mec.targets = 'all' | |
298 |
|
303 | |||
299 | In [21]: mec.block = True |
|
304 | In [21]: mec.block = True | |
300 |
|
305 | |||
301 | In [22]: mec.execute('b=10; print b') |
|
306 | In [22]: mec.execute('b=10; print b') | |
302 | Out[22]: |
|
307 | Out[22]: | |
303 | <Results List> |
|
308 | <Results List> | |
304 | [0] In [7]: b=10; print b |
|
309 | [0] In [7]: b=10; print b | |
305 | [0] Out[7]: 10 |
|
310 | [0] Out[7]: 10 | |
306 |
|
311 | |||
307 | [1] In [6]: b=10; print b |
|
312 | [1] In [6]: b=10; print b | |
308 | [1] Out[6]: 10 |
|
313 | [1] Out[6]: 10 | |
309 |
|
314 | |||
310 | [2] In [7]: b=10; print b |
|
315 | [2] In [7]: b=10; print b | |
311 | [2] Out[7]: 10 |
|
316 | [2] Out[7]: 10 | |
312 |
|
317 | |||
313 | [3] In [6]: b=10; print b |
|
318 | [3] In [6]: b=10; print b | |
314 | [3] Out[6]: 10 |
|
319 | [3] Out[6]: 10 | |
315 |
|
320 | |||
316 | The :attr:`block` and :attr:`targets` instance attributes also determine the |
|
321 | The :attr:`block` and :attr:`targets` instance attributes also determine the | |
317 | behavior of the parallel magic commands. |
|
322 | behavior of the parallel magic commands. | |
318 |
|
323 | |||
319 |
|
324 | |||
320 | Parallel magic commands |
|
325 | Parallel magic commands | |
321 | ----------------------- |
|
326 | ----------------------- | |
322 |
|
327 | |||
323 | We provide a few IPython magic commands (``%px``, ``%autopx`` and ``%result``) |
|
328 | We provide a few IPython magic commands (``%px``, ``%autopx`` and ``%result``) | |
324 | that make it more pleasant to execute Python commands on the engines |
|
329 | that make it more pleasant to execute Python commands on the engines | |
325 | interactively. These are simply shortcuts to :meth:`execute` and |
|
330 | interactively. These are simply shortcuts to :meth:`execute` and | |
326 | :meth:`get_result`. The ``%px`` magic executes a single Python command on the |
|
331 | :meth:`get_result`. The ``%px`` magic executes a single Python command on the | |
327 | engines specified by the :attr:`targets` attribute of the |
|
332 | engines specified by the :attr:`targets` attribute of the | |
328 | :class:`MultiEngineClient` instance (by default this is ``'all'``): |
|
333 | :class:`MultiEngineClient` instance (by default this is ``'all'``): | |
329 |
|
334 | |||
330 | .. sourcecode:: ipython |
|
335 | .. sourcecode:: ipython | |
331 |
|
336 | |||
332 | # Make this MultiEngineClient active for parallel magic commands |
|
337 | # Make this MultiEngineClient active for parallel magic commands | |
333 | In [23]: mec.activate() |
|
338 | In [23]: mec.activate() | |
334 |
|
339 | |||
335 | In [24]: mec.block=True |
|
340 | In [24]: mec.block=True | |
336 |
|
341 | |||
337 | In [25]: import numpy |
|
342 | In [25]: import numpy | |
338 |
|
343 | |||
339 | In [26]: %px import numpy |
|
344 | In [26]: %px import numpy | |
340 | Executing command on Controller |
|
345 | Executing command on Controller | |
341 | Out[26]: |
|
346 | Out[26]: | |
342 | <Results List> |
|
347 | <Results List> | |
343 | [0] In [8]: import numpy |
|
348 | [0] In [8]: import numpy | |
344 | [1] In [7]: import numpy |
|
349 | [1] In [7]: import numpy | |
345 | [2] In [8]: import numpy |
|
350 | [2] In [8]: import numpy | |
346 | [3] In [7]: import numpy |
|
351 | [3] In [7]: import numpy | |
347 |
|
352 | |||
348 |
|
353 | |||
349 | In [27]: %px a = numpy.random.rand(2,2) |
|
354 | In [27]: %px a = numpy.random.rand(2,2) | |
350 | Executing command on Controller |
|
355 | Executing command on Controller | |
351 | Out[27]: |
|
356 | Out[27]: | |
352 | <Results List> |
|
357 | <Results List> | |
353 | [0] In [9]: a = numpy.random.rand(2,2) |
|
358 | [0] In [9]: a = numpy.random.rand(2,2) | |
354 | [1] In [8]: a = numpy.random.rand(2,2) |
|
359 | [1] In [8]: a = numpy.random.rand(2,2) | |
355 | [2] In [9]: a = numpy.random.rand(2,2) |
|
360 | [2] In [9]: a = numpy.random.rand(2,2) | |
356 | [3] In [8]: a = numpy.random.rand(2,2) |
|
361 | [3] In [8]: a = numpy.random.rand(2,2) | |
357 |
|
362 | |||
358 |
|
363 | |||
359 | In [28]: %px print numpy.linalg.eigvals(a) |
|
364 | In [28]: %px print numpy.linalg.eigvals(a) | |
360 | Executing command on Controller |
|
365 | Executing command on Controller | |
361 | Out[28]: |
|
366 | Out[28]: | |
362 | <Results List> |
|
367 | <Results List> | |
363 | [0] In [10]: print numpy.linalg.eigvals(a) |
|
368 | [0] In [10]: print numpy.linalg.eigvals(a) | |
364 | [0] Out[10]: [ 1.28167017 0.14197338] |
|
369 | [0] Out[10]: [ 1.28167017 0.14197338] | |
365 |
|
370 | |||
366 | [1] In [9]: print numpy.linalg.eigvals(a) |
|
371 | [1] In [9]: print numpy.linalg.eigvals(a) | |
367 | [1] Out[9]: [-0.14093616 1.27877273] |
|
372 | [1] Out[9]: [-0.14093616 1.27877273] | |
368 |
|
373 | |||
369 | [2] In [10]: print numpy.linalg.eigvals(a) |
|
374 | [2] In [10]: print numpy.linalg.eigvals(a) | |
370 | [2] Out[10]: [-0.37023573 1.06779409] |
|
375 | [2] Out[10]: [-0.37023573 1.06779409] | |
371 |
|
376 | |||
372 | [3] In [9]: print numpy.linalg.eigvals(a) |
|
377 | [3] In [9]: print numpy.linalg.eigvals(a) | |
373 | [3] Out[9]: [ 0.83664764 -0.25602658] |
|
378 | [3] Out[9]: [ 0.83664764 -0.25602658] | |
374 |
|
379 | |||
375 | The ``%result`` magic gets and prints the stdin/stdout/stderr of the last |
|
380 | The ``%result`` magic gets and prints the stdin/stdout/stderr of the last | |
376 | command executed on each engine. It is simply a shortcut to the |
|
381 | command executed on each engine. It is simply a shortcut to the | |
377 | :meth:`get_result` method: |
|
382 | :meth:`get_result` method: | |
378 |
|
383 | |||
379 | .. sourcecode:: ipython |
|
384 | .. sourcecode:: ipython | |
380 |
|
385 | |||
381 | In [29]: %result |
|
386 | In [29]: %result | |
382 | Out[29]: |
|
387 | Out[29]: | |
383 | <Results List> |
|
388 | <Results List> | |
384 | [0] In [10]: print numpy.linalg.eigvals(a) |
|
389 | [0] In [10]: print numpy.linalg.eigvals(a) | |
385 | [0] Out[10]: [ 1.28167017 0.14197338] |
|
390 | [0] Out[10]: [ 1.28167017 0.14197338] | |
386 |
|
391 | |||
387 | [1] In [9]: print numpy.linalg.eigvals(a) |
|
392 | [1] In [9]: print numpy.linalg.eigvals(a) | |
388 | [1] Out[9]: [-0.14093616 1.27877273] |
|
393 | [1] Out[9]: [-0.14093616 1.27877273] | |
389 |
|
394 | |||
390 | [2] In [10]: print numpy.linalg.eigvals(a) |
|
395 | [2] In [10]: print numpy.linalg.eigvals(a) | |
391 | [2] Out[10]: [-0.37023573 1.06779409] |
|
396 | [2] Out[10]: [-0.37023573 1.06779409] | |
392 |
|
397 | |||
393 | [3] In [9]: print numpy.linalg.eigvals(a) |
|
398 | [3] In [9]: print numpy.linalg.eigvals(a) | |
394 | [3] Out[9]: [ 0.83664764 -0.25602658] |
|
399 | [3] Out[9]: [ 0.83664764 -0.25602658] | |
395 |
|
400 | |||
396 | The ``%autopx`` magic switches to a mode where everything you type is executed |
|
401 | The ``%autopx`` magic switches to a mode where everything you type is executed | |
397 | on the engines given by the :attr:`targets` attribute: |
|
402 | on the engines given by the :attr:`targets` attribute: | |
398 |
|
403 | |||
399 | .. sourcecode:: ipython |
|
404 | .. sourcecode:: ipython | |
400 |
|
405 | |||
401 | In [30]: mec.block=False |
|
406 | In [30]: mec.block=False | |
402 |
|
407 | |||
403 | In [31]: %autopx |
|
408 | In [31]: %autopx | |
404 | Auto Parallel Enabled |
|
409 | Auto Parallel Enabled | |
405 | Type %autopx to disable |
|
410 | Type %autopx to disable | |
406 |
|
411 | |||
407 | In [32]: max_evals = [] |
|
412 | In [32]: max_evals = [] | |
408 | <IPython.kernel.multiengineclient.PendingResult object at 0x17b8a70> |
|
413 | <IPython.kernel.multiengineclient.PendingResult object at 0x17b8a70> | |
409 |
|
414 | |||
410 | In [33]: for i in range(100): |
|
415 | In [33]: for i in range(100): | |
411 | ....: a = numpy.random.rand(10,10) |
|
416 | ....: a = numpy.random.rand(10,10) | |
412 | ....: a = a+a.transpose() |
|
417 | ....: a = a+a.transpose() | |
413 | ....: evals = numpy.linalg.eigvals(a) |
|
418 | ....: evals = numpy.linalg.eigvals(a) | |
414 | ....: max_evals.append(evals[0].real) |
|
419 | ....: max_evals.append(evals[0].real) | |
415 | ....: |
|
420 | ....: | |
416 | ....: |
|
421 | ....: | |
417 | <IPython.kernel.multiengineclient.PendingResult object at 0x17af8f0> |
|
422 | <IPython.kernel.multiengineclient.PendingResult object at 0x17af8f0> | |
418 |
|
423 | |||
419 | In [34]: %autopx |
|
424 | In [34]: %autopx | |
420 | Auto Parallel Disabled |
|
425 | Auto Parallel Disabled | |
421 |
|
426 | |||
422 | In [35]: mec.block=True |
|
427 | In [35]: mec.block=True | |
423 |
|
428 | |||
424 | In [36]: px print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) |
|
429 | In [36]: px print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) | |
425 | Executing command on Controller |
|
430 | Executing command on Controller | |
426 | Out[36]: |
|
431 | Out[36]: | |
427 | <Results List> |
|
432 | <Results List> | |
428 | [0] In [13]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) |
|
433 | [0] In [13]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) | |
429 | [0] Out[13]: Average max eigenvalue is: 10.1387247332 |
|
434 | [0] Out[13]: Average max eigenvalue is: 10.1387247332 | |
430 |
|
435 | |||
431 | [1] In [12]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) |
|
436 | [1] In [12]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) | |
432 | [1] Out[12]: Average max eigenvalue is: 10.2076902286 |
|
437 | [1] Out[12]: Average max eigenvalue is: 10.2076902286 | |
433 |
|
438 | |||
434 | [2] In [13]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) |
|
439 | [2] In [13]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) | |
435 | [2] Out[13]: Average max eigenvalue is: 10.1891484655 |
|
440 | [2] Out[13]: Average max eigenvalue is: 10.1891484655 | |
436 |
|
441 | |||
437 | [3] In [12]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) |
|
442 | [3] In [12]: print "Average max eigenvalue is: ", sum(max_evals)/len(max_evals) | |
438 | [3] Out[12]: Average max eigenvalue is: 10.1158837784 |
|
443 | [3] Out[12]: Average max eigenvalue is: 10.1158837784 | |
439 |
|
444 | |||
440 |
|
445 | |||
441 | Moving Python objects around |
|
446 | Moving Python objects around | |
442 | ============================ |
|
447 | ============================ | |
443 |
|
448 | |||
444 | In addition to executing code on engines, you can transfer Python objects to |
|
449 | In addition to executing code on engines, you can transfer Python objects to | |
445 | and from your IPython session and the engines. In IPython, these operations |
|
450 | and from your IPython session and the engines. In IPython, these operations | |
446 | are called :meth:`push` (sending an object to the engines) and :meth:`pull` |
|
451 | are called :meth:`push` (sending an object to the engines) and :meth:`pull` | |
447 | (getting an object from the engines). |
|
452 | (getting an object from the engines). | |
448 |
|
453 | |||
449 | Basic push and pull |
|
454 | Basic push and pull | |
450 | ------------------- |
|
455 | ------------------- | |
451 |
|
456 | |||
452 | Here are some examples of how you use :meth:`push` and :meth:`pull`: |
|
457 | Here are some examples of how you use :meth:`push` and :meth:`pull`: | |
453 |
|
458 | |||
454 | .. sourcecode:: ipython |
|
459 | .. sourcecode:: ipython | |
455 |
|
460 | |||
456 | In [38]: mec.push(dict(a=1.03234,b=3453)) |
|
461 | In [38]: mec.push(dict(a=1.03234,b=3453)) | |
457 | Out[38]: [None, None, None, None] |
|
462 | Out[38]: [None, None, None, None] | |
458 |
|
463 | |||
459 | In [39]: mec.pull('a') |
|
464 | In [39]: mec.pull('a') | |
460 | Out[39]: [1.03234, 1.03234, 1.03234, 1.03234] |
|
465 | Out[39]: [1.03234, 1.03234, 1.03234, 1.03234] | |
461 |
|
466 | |||
462 | In [40]: mec.pull('b',targets=0) |
|
467 | In [40]: mec.pull('b',targets=0) | |
463 | Out[40]: [3453] |
|
468 | Out[40]: [3453] | |
464 |
|
469 | |||
465 | In [41]: mec.pull(('a','b')) |
|
470 | In [41]: mec.pull(('a','b')) | |
466 | Out[41]: [[1.03234, 3453], [1.03234, 3453], [1.03234, 3453], [1.03234, 3453]] |
|
471 | Out[41]: [[1.03234, 3453], [1.03234, 3453], [1.03234, 3453], [1.03234, 3453]] | |
467 |
|
472 | |||
468 | In [42]: mec.zip_pull(('a','b')) |
|
473 | In [42]: mec.zip_pull(('a','b')) | |
469 | Out[42]: [(1.03234, 1.03234, 1.03234, 1.03234), (3453, 3453, 3453, 3453)] |
|
474 | Out[42]: [(1.03234, 1.03234, 1.03234, 1.03234), (3453, 3453, 3453, 3453)] | |
470 |
|
475 | |||
471 | In [43]: mec.push(dict(c='speed')) |
|
476 | In [43]: mec.push(dict(c='speed')) | |
472 | Out[43]: [None, None, None, None] |
|
477 | Out[43]: [None, None, None, None] | |
473 |
|
478 | |||
474 | In [44]: %px print c |
|
479 | In [44]: %px print c | |
475 | Executing command on Controller |
|
480 | Executing command on Controller | |
476 | Out[44]: |
|
481 | Out[44]: | |
477 | <Results List> |
|
482 | <Results List> | |
478 | [0] In [14]: print c |
|
483 | [0] In [14]: print c | |
479 | [0] Out[14]: speed |
|
484 | [0] Out[14]: speed | |
480 |
|
485 | |||
481 | [1] In [13]: print c |
|
486 | [1] In [13]: print c | |
482 | [1] Out[13]: speed |
|
487 | [1] Out[13]: speed | |
483 |
|
488 | |||
484 | [2] In [14]: print c |
|
489 | [2] In [14]: print c | |
485 | [2] Out[14]: speed |
|
490 | [2] Out[14]: speed | |
486 |
|
491 | |||
487 | [3] In [13]: print c |
|
492 | [3] In [13]: print c | |
488 | [3] Out[13]: speed |
|
493 | [3] Out[13]: speed | |
489 |
|
494 | |||
490 | In non-blocking mode :meth:`push` and :meth:`pull` also return |
|
495 | In non-blocking mode :meth:`push` and :meth:`pull` also return | |
491 | :class:`PendingResult` objects: |
|
496 | :class:`PendingResult` objects: | |
492 |
|
497 | |||
493 | .. sourcecode:: ipython |
|
498 | .. sourcecode:: ipython | |
494 |
|
499 | |||
495 | In [47]: mec.block=False |
|
500 | In [47]: mec.block=False | |
496 |
|
501 | |||
497 | In [48]: pr = mec.pull('a') |
|
502 | In [48]: pr = mec.pull('a') | |
498 |
|
503 | |||
499 | In [49]: pr.r |
|
504 | In [49]: pr.r | |
500 | Out[49]: [1.03234, 1.03234, 1.03234, 1.03234] |
|
505 | Out[49]: [1.03234, 1.03234, 1.03234, 1.03234] | |
501 |
|
506 | |||
502 |
|
507 | |||
503 | Push and pull for functions |
|
508 | Push and pull for functions | |
504 | --------------------------- |
|
509 | --------------------------- | |
505 |
|
510 | |||
506 | Functions can also be pushed and pulled using :meth:`push_function` and |
|
511 | Functions can also be pushed and pulled using :meth:`push_function` and | |
507 | :meth:`pull_function`: |
|
512 | :meth:`pull_function`: | |
508 |
|
513 | |||
509 | .. sourcecode:: ipython |
|
514 | .. sourcecode:: ipython | |
510 |
|
515 | |||
511 | In [52]: mec.block=True |
|
516 | In [52]: mec.block=True | |
512 |
|
517 | |||
513 | In [53]: def f(x): |
|
518 | In [53]: def f(x): | |
514 | ....: return 2.0*x**4 |
|
519 | ....: return 2.0*x**4 | |
515 | ....: |
|
520 | ....: | |
516 |
|
521 | |||
517 | In [54]: mec.push_function(dict(f=f)) |
|
522 | In [54]: mec.push_function(dict(f=f)) | |
518 | Out[54]: [None, None, None, None] |
|
523 | Out[54]: [None, None, None, None] | |
519 |
|
524 | |||
520 | In [55]: mec.execute('y = f(4.0)') |
|
525 | In [55]: mec.execute('y = f(4.0)') | |
521 | Out[55]: |
|
526 | Out[55]: | |
522 | <Results List> |
|
527 | <Results List> | |
523 | [0] In [15]: y = f(4.0) |
|
528 | [0] In [15]: y = f(4.0) | |
524 | [1] In [14]: y = f(4.0) |
|
529 | [1] In [14]: y = f(4.0) | |
525 | [2] In [15]: y = f(4.0) |
|
530 | [2] In [15]: y = f(4.0) | |
526 | [3] In [14]: y = f(4.0) |
|
531 | [3] In [14]: y = f(4.0) | |
527 |
|
532 | |||
528 |
|
533 | |||
529 | In [56]: px print y |
|
534 | In [56]: px print y | |
530 | Executing command on Controller |
|
535 | Executing command on Controller | |
531 | Out[56]: |
|
536 | Out[56]: | |
532 | <Results List> |
|
537 | <Results List> | |
533 | [0] In [16]: print y |
|
538 | [0] In [16]: print y | |
534 | [0] Out[16]: 512.0 |
|
539 | [0] Out[16]: 512.0 | |
535 |
|
540 | |||
536 | [1] In [15]: print y |
|
541 | [1] In [15]: print y | |
537 | [1] Out[15]: 512.0 |
|
542 | [1] Out[15]: 512.0 | |
538 |
|
543 | |||
539 | [2] In [16]: print y |
|
544 | [2] In [16]: print y | |
540 | [2] Out[16]: 512.0 |
|
545 | [2] Out[16]: 512.0 | |
541 |
|
546 | |||
542 | [3] In [15]: print y |
|
547 | [3] In [15]: print y | |
543 | [3] Out[15]: 512.0 |
|
548 | [3] Out[15]: 512.0 | |
544 |
|
549 | |||
545 |
|
550 | |||
546 | Dictionary interface |
|
551 | Dictionary interface | |
547 | -------------------- |
|
552 | -------------------- | |
548 |
|
553 | |||
549 | As a shorthand to :meth:`push` and :meth:`pull`, the |
|
554 | As a shorthand to :meth:`push` and :meth:`pull`, the | |
550 | :class:`MultiEngineClient` class implements some of the Python dictionary |
|
555 | :class:`MultiEngineClient` class implements some of the Python dictionary | |
551 | interface. This make the remote namespaces of the engines appear as a local |
|
556 | interface. This make the remote namespaces of the engines appear as a local | |
552 | dictionary. Underneath, this uses :meth:`push` and :meth:`pull`: |
|
557 | dictionary. Underneath, this uses :meth:`push` and :meth:`pull`: | |
553 |
|
558 | |||
554 | .. sourcecode:: ipython |
|
559 | .. sourcecode:: ipython | |
555 |
|
560 | |||
556 | In [50]: mec.block=True |
|
561 | In [50]: mec.block=True | |
557 |
|
562 | |||
558 | In [51]: mec['a']=['foo','bar'] |
|
563 | In [51]: mec['a']=['foo','bar'] | |
559 |
|
564 | |||
560 | In [52]: mec['a'] |
|
565 | In [52]: mec['a'] | |
561 | Out[52]: [['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar']] |
|
566 | Out[52]: [['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar']] | |
562 |
|
567 | |||
563 | Scatter and gather |
|
568 | Scatter and gather | |
564 | ------------------ |
|
569 | ------------------ | |
565 |
|
570 | |||
566 | Sometimes it is useful to partition a sequence and push the partitions to |
|
571 | Sometimes it is useful to partition a sequence and push the partitions to | |
567 | different engines. In MPI language, this is know as scatter/gather and we |
|
572 | different engines. In MPI language, this is know as scatter/gather and we | |
568 | follow that terminology. However, it is important to remember that in |
|
573 | follow that terminology. However, it is important to remember that in | |
569 | IPython's :class:`MultiEngineClient` class, :meth:`scatter` is from the |
|
574 | IPython's :class:`MultiEngineClient` class, :meth:`scatter` is from the | |
570 | interactive IPython session to the engines and :meth:`gather` is from the |
|
575 | interactive IPython session to the engines and :meth:`gather` is from the | |
571 | engines back to the interactive IPython session. For scatter/gather operations |
|
576 | engines back to the interactive IPython session. For scatter/gather operations | |
572 | between engines, MPI should be used: |
|
577 | between engines, MPI should be used: | |
573 |
|
578 | |||
574 | .. sourcecode:: ipython |
|
579 | .. sourcecode:: ipython | |
575 |
|
580 | |||
576 | In [58]: mec.scatter('a',range(16)) |
|
581 | In [58]: mec.scatter('a',range(16)) | |
577 | Out[58]: [None, None, None, None] |
|
582 | Out[58]: [None, None, None, None] | |
578 |
|
583 | |||
579 | In [59]: px print a |
|
584 | In [59]: px print a | |
580 | Executing command on Controller |
|
585 | Executing command on Controller | |
581 | Out[59]: |
|
586 | Out[59]: | |
582 | <Results List> |
|
587 | <Results List> | |
583 | [0] In [17]: print a |
|
588 | [0] In [17]: print a | |
584 | [0] Out[17]: [0, 1, 2, 3] |
|
589 | [0] Out[17]: [0, 1, 2, 3] | |
585 |
|
590 | |||
586 | [1] In [16]: print a |
|
591 | [1] In [16]: print a | |
587 | [1] Out[16]: [4, 5, 6, 7] |
|
592 | [1] Out[16]: [4, 5, 6, 7] | |
588 |
|
593 | |||
589 | [2] In [17]: print a |
|
594 | [2] In [17]: print a | |
590 | [2] Out[17]: [8, 9, 10, 11] |
|
595 | [2] Out[17]: [8, 9, 10, 11] | |
591 |
|
596 | |||
592 | [3] In [16]: print a |
|
597 | [3] In [16]: print a | |
593 | [3] Out[16]: [12, 13, 14, 15] |
|
598 | [3] Out[16]: [12, 13, 14, 15] | |
594 |
|
599 | |||
595 |
|
600 | |||
596 | In [60]: mec.gather('a') |
|
601 | In [60]: mec.gather('a') | |
597 | Out[60]: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15] |
|
602 | Out[60]: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15] | |
598 |
|
603 | |||
599 | Other things to look at |
|
604 | Other things to look at | |
600 | ======================= |
|
605 | ======================= | |
601 |
|
606 | |||
602 | How to do parallel list comprehensions |
|
607 | How to do parallel list comprehensions | |
603 | -------------------------------------- |
|
608 | -------------------------------------- | |
604 |
|
609 | |||
605 | In many cases list comprehensions are nicer than using the map function. While |
|
610 | In many cases list comprehensions are nicer than using the map function. While | |
606 | we don't have fully parallel list comprehensions, it is simple to get the |
|
611 | we don't have fully parallel list comprehensions, it is simple to get the | |
607 | basic effect using :meth:`scatter` and :meth:`gather`: |
|
612 | basic effect using :meth:`scatter` and :meth:`gather`: | |
608 |
|
613 | |||
609 | .. sourcecode:: ipython |
|
614 | .. sourcecode:: ipython | |
610 |
|
615 | |||
611 | In [66]: mec.scatter('x',range(64)) |
|
616 | In [66]: mec.scatter('x',range(64)) | |
612 | Out[66]: [None, None, None, None] |
|
617 | Out[66]: [None, None, None, None] | |
613 |
|
618 | |||
614 | In [67]: px y = [i**10 for i in x] |
|
619 | In [67]: px y = [i**10 for i in x] | |
615 | Executing command on Controller |
|
620 | Executing command on Controller | |
616 | Out[67]: |
|
621 | Out[67]: | |
617 | <Results List> |
|
622 | <Results List> | |
618 | [0] In [19]: y = [i**10 for i in x] |
|
623 | [0] In [19]: y = [i**10 for i in x] | |
619 | [1] In [18]: y = [i**10 for i in x] |
|
624 | [1] In [18]: y = [i**10 for i in x] | |
620 | [2] In [19]: y = [i**10 for i in x] |
|
625 | [2] In [19]: y = [i**10 for i in x] | |
621 | [3] In [18]: y = [i**10 for i in x] |
|
626 | [3] In [18]: y = [i**10 for i in x] | |
622 |
|
627 | |||
623 |
|
628 | |||
624 | In [68]: y = mec.gather('y') |
|
629 | In [68]: y = mec.gather('y') | |
625 |
|
630 | |||
626 | In [69]: print y |
|
631 | In [69]: print y | |
627 | [0, 1, 1024, 59049, 1048576, 9765625, 60466176, 282475249, 1073741824,...] |
|
632 | [0, 1, 1024, 59049, 1048576, 9765625, 60466176, 282475249, 1073741824,...] | |
628 |
|
633 | |||
629 | Parallel exceptions |
|
634 | Parallel exceptions | |
630 | ------------------- |
|
635 | ------------------- | |
631 |
|
636 | |||
632 | In the multiengine interface, parallel commands can raise Python exceptions, |
|
637 | In the multiengine interface, parallel commands can raise Python exceptions, | |
633 | just like serial commands. But, it is a little subtle, because a single |
|
638 | just like serial commands. But, it is a little subtle, because a single | |
634 | parallel command can actually raise multiple exceptions (one for each engine |
|
639 | parallel command can actually raise multiple exceptions (one for each engine | |
635 | the command was run on). To express this idea, the MultiEngine interface has a |
|
640 | the command was run on). To express this idea, the MultiEngine interface has a | |
636 | :exc:`CompositeError` exception class that will be raised in most cases. The |
|
641 | :exc:`CompositeError` exception class that will be raised in most cases. The | |
637 | :exc:`CompositeError` class is a special type of exception that wraps one or |
|
642 | :exc:`CompositeError` class is a special type of exception that wraps one or | |
638 | more other types of exceptions. Here is how it works: |
|
643 | more other types of exceptions. Here is how it works: | |
639 |
|
644 | |||
640 | .. sourcecode:: ipython |
|
645 | .. sourcecode:: ipython | |
641 |
|
646 | |||
642 | In [76]: mec.block=True |
|
647 | In [76]: mec.block=True | |
643 |
|
648 | |||
644 | In [77]: mec.execute('1/0') |
|
649 | In [77]: mec.execute('1/0') | |
645 | --------------------------------------------------------------------------- |
|
650 | --------------------------------------------------------------------------- | |
646 | CompositeError Traceback (most recent call last) |
|
651 | CompositeError Traceback (most recent call last) | |
647 |
|
652 | |||
648 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() |
|
653 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() | |
649 |
|
654 | |||
650 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in execute(self, lines, targets, block) |
|
655 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in execute(self, lines, targets, block) | |
651 | 432 targets, block = self._findTargetsAndBlock(targets, block) |
|
656 | 432 targets, block = self._findTargetsAndBlock(targets, block) | |
652 | 433 result = blockingCallFromThread(self.smultiengine.execute, lines, |
|
657 | 433 result = blockingCallFromThread(self.smultiengine.execute, lines, | |
653 | --> 434 targets=targets, block=block) |
|
658 | --> 434 targets=targets, block=block) | |
654 | 435 if block: |
|
659 | 435 if block: | |
655 | 436 result = ResultList(result) |
|
660 | 436 result = ResultList(result) | |
656 |
|
661 | |||
657 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) |
|
662 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) | |
658 | 72 result.raiseException() |
|
663 | 72 result.raiseException() | |
659 | 73 except Exception, e: |
|
664 | 73 except Exception, e: | |
660 | ---> 74 raise e |
|
665 | ---> 74 raise e | |
661 | 75 return result |
|
666 | 75 return result | |
662 | 76 |
|
667 | 76 | |
663 |
|
668 | |||
664 | CompositeError: one or more exceptions from call to method: execute |
|
669 | CompositeError: one or more exceptions from call to method: execute | |
665 | [0:execute]: ZeroDivisionError: integer division or modulo by zero |
|
670 | [0:execute]: ZeroDivisionError: integer division or modulo by zero | |
666 | [1:execute]: ZeroDivisionError: integer division or modulo by zero |
|
671 | [1:execute]: ZeroDivisionError: integer division or modulo by zero | |
667 | [2:execute]: ZeroDivisionError: integer division or modulo by zero |
|
672 | [2:execute]: ZeroDivisionError: integer division or modulo by zero | |
668 | [3:execute]: ZeroDivisionError: integer division or modulo by zero |
|
673 | [3:execute]: ZeroDivisionError: integer division or modulo by zero | |
669 |
|
674 | |||
670 | Notice how the error message printed when :exc:`CompositeError` is raised has information about the individual exceptions that were raised on each engine. If you want, you can even raise one of these original exceptions: |
|
675 | Notice how the error message printed when :exc:`CompositeError` is raised has | |
|
676 | information about the individual exceptions that were raised on each engine. | |||
|
677 | If you want, you can even raise one of these original exceptions: | |||
671 |
|
678 | |||
672 | .. sourcecode:: ipython |
|
679 | .. sourcecode:: ipython | |
673 |
|
680 | |||
674 | In [80]: try: |
|
681 | In [80]: try: | |
675 | ....: mec.execute('1/0') |
|
682 | ....: mec.execute('1/0') | |
676 | ....: except client.CompositeError, e: |
|
683 | ....: except client.CompositeError, e: | |
677 | ....: e.raise_exception() |
|
684 | ....: e.raise_exception() | |
678 | ....: |
|
685 | ....: | |
679 | ....: |
|
686 | ....: | |
680 | --------------------------------------------------------------------------- |
|
687 | --------------------------------------------------------------------------- | |
681 | ZeroDivisionError Traceback (most recent call last) |
|
688 | ZeroDivisionError Traceback (most recent call last) | |
682 |
|
689 | |||
683 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() |
|
690 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() | |
684 |
|
691 | |||
685 | /ipython1-client-r3021/ipython1/kernel/error.pyc in raise_exception(self, excid) |
|
692 | /ipython1-client-r3021/ipython1/kernel/error.pyc in raise_exception(self, excid) | |
686 | 156 raise IndexError("an exception with index %i does not exist"%excid) |
|
693 | 156 raise IndexError("an exception with index %i does not exist"%excid) | |
687 | 157 else: |
|
694 | 157 else: | |
688 | --> 158 raise et, ev, etb |
|
695 | --> 158 raise et, ev, etb | |
689 | 159 |
|
696 | 159 | |
690 | 160 def collect_exceptions(rlist, method): |
|
697 | 160 def collect_exceptions(rlist, method): | |
691 |
|
698 | |||
692 | ZeroDivisionError: integer division or modulo by zero |
|
699 | ZeroDivisionError: integer division or modulo by zero | |
693 |
|
700 | |||
694 | If you are working in IPython, you can simple type ``%debug`` after one of |
|
701 | If you are working in IPython, you can simple type ``%debug`` after one of | |
695 | these :exc:`CompositeError` exceptions is raised, and inspect the exception |
|
702 | these :exc:`CompositeError` exceptions is raised, and inspect the exception | |
696 | instance: |
|
703 | instance: | |
697 |
|
704 | |||
698 | .. sourcecode:: ipython |
|
705 | .. sourcecode:: ipython | |
699 |
|
706 | |||
700 | In [81]: mec.execute('1/0') |
|
707 | In [81]: mec.execute('1/0') | |
701 | --------------------------------------------------------------------------- |
|
708 | --------------------------------------------------------------------------- | |
702 | CompositeError Traceback (most recent call last) |
|
709 | CompositeError Traceback (most recent call last) | |
703 |
|
710 | |||
704 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() |
|
711 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() | |
705 |
|
712 | |||
706 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in execute(self, lines, targets, block) |
|
713 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in execute(self, lines, targets, block) | |
707 | 432 targets, block = self._findTargetsAndBlock(targets, block) |
|
714 | 432 targets, block = self._findTargetsAndBlock(targets, block) | |
708 | 433 result = blockingCallFromThread(self.smultiengine.execute, lines, |
|
715 | 433 result = blockingCallFromThread(self.smultiengine.execute, lines, | |
709 | --> 434 targets=targets, block=block) |
|
716 | --> 434 targets=targets, block=block) | |
710 | 435 if block: |
|
717 | 435 if block: | |
711 | 436 result = ResultList(result) |
|
718 | 436 result = ResultList(result) | |
712 |
|
719 | |||
713 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) |
|
720 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) | |
714 | 72 result.raiseException() |
|
721 | 72 result.raiseException() | |
715 | 73 except Exception, e: |
|
722 | 73 except Exception, e: | |
716 | ---> 74 raise e |
|
723 | ---> 74 raise e | |
717 | 75 return result |
|
724 | 75 return result | |
718 | 76 |
|
725 | 76 | |
719 |
|
726 | |||
720 | CompositeError: one or more exceptions from call to method: execute |
|
727 | CompositeError: one or more exceptions from call to method: execute | |
721 | [0:execute]: ZeroDivisionError: integer division or modulo by zero |
|
728 | [0:execute]: ZeroDivisionError: integer division or modulo by zero | |
722 | [1:execute]: ZeroDivisionError: integer division or modulo by zero |
|
729 | [1:execute]: ZeroDivisionError: integer division or modulo by zero | |
723 | [2:execute]: ZeroDivisionError: integer division or modulo by zero |
|
730 | [2:execute]: ZeroDivisionError: integer division or modulo by zero | |
724 | [3:execute]: ZeroDivisionError: integer division or modulo by zero |
|
731 | [3:execute]: ZeroDivisionError: integer division or modulo by zero | |
725 |
|
732 | |||
726 | In [82]: %debug |
|
733 | In [82]: %debug | |
727 | > |
|
734 | > | |
728 |
|
735 | |||
729 | /ipython1-client-r3021/ipython1/kernel/twistedutil.py(74)blockingCallFromThread() |
|
736 | /ipython1-client-r3021/ipython1/kernel/twistedutil.py(74)blockingCallFromThread() | |
730 | 73 except Exception, e: |
|
737 | 73 except Exception, e: | |
731 | ---> 74 raise e |
|
738 | ---> 74 raise e | |
732 | 75 return result |
|
739 | 75 return result | |
733 |
|
740 | |||
734 | # With the debugger running, e is the exceptions instance. We can tab complete |
|
741 | # With the debugger running, e is the exceptions instance. We can tab complete | |
735 | # on it and see the extra methods that are available. |
|
742 | # on it and see the extra methods that are available. | |
736 | ipdb> e. |
|
743 | ipdb> e. | |
737 | e.__class__ e.__getitem__ e.__new__ e.__setstate__ e.args |
|
744 | e.__class__ e.__getitem__ e.__new__ e.__setstate__ e.args | |
738 | e.__delattr__ e.__getslice__ e.__reduce__ e.__str__ e.elist |
|
745 | e.__delattr__ e.__getslice__ e.__reduce__ e.__str__ e.elist | |
739 | e.__dict__ e.__hash__ e.__reduce_ex__ e.__weakref__ e.message |
|
746 | e.__dict__ e.__hash__ e.__reduce_ex__ e.__weakref__ e.message | |
740 | e.__doc__ e.__init__ e.__repr__ e._get_engine_str e.print_tracebacks |
|
747 | e.__doc__ e.__init__ e.__repr__ e._get_engine_str e.print_tracebacks | |
741 | e.__getattribute__ e.__module__ e.__setattr__ e._get_traceback e.raise_exception |
|
748 | e.__getattribute__ e.__module__ e.__setattr__ e._get_traceback e.raise_exception | |
742 | ipdb> e.print_tracebacks() |
|
749 | ipdb> e.print_tracebacks() | |
743 | [0:execute]: |
|
750 | [0:execute]: | |
744 | --------------------------------------------------------------------------- |
|
751 | --------------------------------------------------------------------------- | |
745 | ZeroDivisionError Traceback (most recent call last) |
|
752 | ZeroDivisionError Traceback (most recent call last) | |
746 |
|
753 | |||
747 | /ipython1-client-r3021/docs/examples/<string> in <module>() |
|
754 | /ipython1-client-r3021/docs/examples/<string> in <module>() | |
748 |
|
755 | |||
749 | ZeroDivisionError: integer division or modulo by zero |
|
756 | ZeroDivisionError: integer division or modulo by zero | |
750 |
|
757 | |||
751 | [1:execute]: |
|
758 | [1:execute]: | |
752 | --------------------------------------------------------------------------- |
|
759 | --------------------------------------------------------------------------- | |
753 | ZeroDivisionError Traceback (most recent call last) |
|
760 | ZeroDivisionError Traceback (most recent call last) | |
754 |
|
761 | |||
755 | /ipython1-client-r3021/docs/examples/<string> in <module>() |
|
762 | /ipython1-client-r3021/docs/examples/<string> in <module>() | |
756 |
|
763 | |||
757 | ZeroDivisionError: integer division or modulo by zero |
|
764 | ZeroDivisionError: integer division or modulo by zero | |
758 |
|
765 | |||
759 | [2:execute]: |
|
766 | [2:execute]: | |
760 | --------------------------------------------------------------------------- |
|
767 | --------------------------------------------------------------------------- | |
761 | ZeroDivisionError Traceback (most recent call last) |
|
768 | ZeroDivisionError Traceback (most recent call last) | |
762 |
|
769 | |||
763 | /ipython1-client-r3021/docs/examples/<string> in <module>() |
|
770 | /ipython1-client-r3021/docs/examples/<string> in <module>() | |
764 |
|
771 | |||
765 | ZeroDivisionError: integer division or modulo by zero |
|
772 | ZeroDivisionError: integer division or modulo by zero | |
766 |
|
773 | |||
767 | [3:execute]: |
|
774 | [3:execute]: | |
768 | --------------------------------------------------------------------------- |
|
775 | --------------------------------------------------------------------------- | |
769 | ZeroDivisionError Traceback (most recent call last) |
|
776 | ZeroDivisionError Traceback (most recent call last) | |
770 |
|
777 | |||
771 | /ipython1-client-r3021/docs/examples/<string> in <module>() |
|
778 | /ipython1-client-r3021/docs/examples/<string> in <module>() | |
772 |
|
779 | |||
773 | ZeroDivisionError: integer division or modulo by zero |
|
780 | ZeroDivisionError: integer division or modulo by zero | |
774 |
|
781 | |||
775 | .. note:: |
|
782 | .. note:: | |
776 |
|
783 | |||
777 | The above example appears to be broken right now because of a change in |
|
784 | The above example appears to be broken right now because of a change in | |
778 | how we are using Twisted. |
|
785 | how we are using Twisted. | |
779 |
|
786 | |||
780 | All of this same error handling magic even works in non-blocking mode: |
|
787 | All of this same error handling magic even works in non-blocking mode: | |
781 |
|
788 | |||
782 | .. sourcecode:: ipython |
|
789 | .. sourcecode:: ipython | |
783 |
|
790 | |||
784 | In [83]: mec.block=False |
|
791 | In [83]: mec.block=False | |
785 |
|
792 | |||
786 | In [84]: pr = mec.execute('1/0') |
|
793 | In [84]: pr = mec.execute('1/0') | |
787 |
|
794 | |||
788 | In [85]: pr.r |
|
795 | In [85]: pr.r | |
789 | --------------------------------------------------------------------------- |
|
796 | --------------------------------------------------------------------------- | |
790 | CompositeError Traceback (most recent call last) |
|
797 | CompositeError Traceback (most recent call last) | |
791 |
|
798 | |||
792 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() |
|
799 | /ipython1-client-r3021/docs/examples/<ipython console> in <module>() | |
793 |
|
800 | |||
794 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in _get_r(self) |
|
801 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in _get_r(self) | |
795 | 170 |
|
802 | 170 | |
796 | 171 def _get_r(self): |
|
803 | 171 def _get_r(self): | |
797 | --> 172 return self.get_result(block=True) |
|
804 | --> 172 return self.get_result(block=True) | |
798 | 173 |
|
805 | 173 | |
799 | 174 r = property(_get_r) |
|
806 | 174 r = property(_get_r) | |
800 |
|
807 | |||
801 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in get_result(self, default, block) |
|
808 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in get_result(self, default, block) | |
802 | 131 return self.result |
|
809 | 131 return self.result | |
803 | 132 try: |
|
810 | 132 try: | |
804 | --> 133 result = self.client.get_pending_deferred(self.result_id, block) |
|
811 | --> 133 result = self.client.get_pending_deferred(self.result_id, block) | |
805 | 134 except error.ResultNotCompleted: |
|
812 | 134 except error.ResultNotCompleted: | |
806 | 135 return default |
|
813 | 135 return default | |
807 |
|
814 | |||
808 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in get_pending_deferred(self, deferredID, block) |
|
815 | /ipython1-client-r3021/ipython1/kernel/multiengineclient.pyc in get_pending_deferred(self, deferredID, block) | |
809 | 385 |
|
816 | 385 | |
810 | 386 def get_pending_deferred(self, deferredID, block): |
|
817 | 386 def get_pending_deferred(self, deferredID, block): | |
811 | --> 387 return blockingCallFromThread(self.smultiengine.get_pending_deferred, deferredID, block) |
|
818 | --> 387 return blockingCallFromThread(self.smultiengine.get_pending_deferred, deferredID, block) | |
812 | 388 |
|
819 | 388 | |
813 | 389 def barrier(self, pendingResults): |
|
820 | 389 def barrier(self, pendingResults): | |
814 |
|
821 | |||
815 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) |
|
822 | /ipython1-client-r3021/ipython1/kernel/twistedutil.pyc in blockingCallFromThread(f, *a, **kw) | |
816 | 72 result.raiseException() |
|
823 | 72 result.raiseException() | |
817 | 73 except Exception, e: |
|
824 | 73 except Exception, e: | |
818 | ---> 74 raise e |
|
825 | ---> 74 raise e | |
819 | 75 return result |
|
826 | 75 return result | |
820 | 76 |
|
827 | 76 | |
821 |
|
828 | |||
822 | CompositeError: one or more exceptions from call to method: execute |
|
829 | CompositeError: one or more exceptions from call to method: execute | |
823 | [0:execute]: ZeroDivisionError: integer division or modulo by zero |
|
830 | [0:execute]: ZeroDivisionError: integer division or modulo by zero | |
824 | [1:execute]: ZeroDivisionError: integer division or modulo by zero |
|
831 | [1:execute]: ZeroDivisionError: integer division or modulo by zero | |
825 | [2:execute]: ZeroDivisionError: integer division or modulo by zero |
|
832 | [2:execute]: ZeroDivisionError: integer division or modulo by zero | |
826 | [3:execute]: ZeroDivisionError: integer division or modulo by zero |
|
833 | [3:execute]: ZeroDivisionError: integer division or modulo by zero | |
827 |
|
834 | |||
828 |
|
835 |
@@ -1,351 +1,387 b'' | |||||
1 | .. _parallel_process: |
|
1 | .. _parallel_process: | |
2 |
|
2 | |||
3 | =========================================== |
|
3 | =========================================== | |
4 | Starting the IPython controller and engines |
|
4 | Starting the IPython controller and engines | |
5 | =========================================== |
|
5 | =========================================== | |
6 |
|
6 | |||
7 | To use IPython for parallel computing, you need to start one instance of |
|
7 | To use IPython for parallel computing, you need to start one instance of | |
8 | the controller and one or more instances of the engine. The controller |
|
8 | the controller and one or more instances of the engine. The controller | |
9 | and each engine can run on different machines or on the same machine. |
|
9 | and each engine can run on different machines or on the same machine. | |
10 | Because of this, there are many different possibilities. |
|
10 | Because of this, there are many different possibilities. | |
11 |
|
11 | |||
12 | Broadly speaking, there are two ways of going about starting a controller and engines: |
|
12 | Broadly speaking, there are two ways of going about starting a controller and engines: | |
13 |
|
13 | |||
14 | * In an automated manner using the :command:`ipcluster` command. |
|
14 | * In an automated manner using the :command:`ipcluster` command. | |
15 | * In a more manual way using the :command:`ipcontroller` and |
|
15 | * In a more manual way using the :command:`ipcontroller` and | |
16 | :command:`ipengine` commands. |
|
16 | :command:`ipengine` commands. | |
17 |
|
17 | |||
18 |
This document describes both of these methods. |
|
18 | This document describes both of these methods. We recommend that new users | |
|
19 | start with the :command:`ipcluster` command as it simplifies many common usage | |||
|
20 | cases. | |||
19 |
|
21 | |||
20 | General considerations |
|
22 | General considerations | |
21 | ====================== |
|
23 | ====================== | |
22 |
|
24 | |||
23 | Before delving into the details about how you can start a controller and engines using the various methods, we outline some of the general issues that come up when starting the controller and engines. These things come up no matter which method you use to start your IPython cluster. |
|
25 | Before delving into the details about how you can start a controller and | |
|
26 | engines using the various methods, we outline some of the general issues that | |||
|
27 | come up when starting the controller and engines. These things come up no | |||
|
28 | matter which method you use to start your IPython cluster. | |||
24 |
|
29 | |||
25 |
Let's say that you want to start the controller on ``host0`` and engines on |
|
30 | Let's say that you want to start the controller on ``host0`` and engines on | |
|
31 | hosts ``host1``-``hostn``. The following steps are then required: | |||
26 |
|
32 | |||
27 | 1. Start the controller on ``host0`` by running :command:`ipcontroller` on |
|
33 | 1. Start the controller on ``host0`` by running :command:`ipcontroller` on | |
28 | ``host0``. |
|
34 | ``host0``. | |
29 | 2. Move the FURL file (:file:`ipcontroller-engine.furl`) created by the |
|
35 | 2. Move the FURL file (:file:`ipcontroller-engine.furl`) created by the | |
30 | controller from ``host0`` to hosts ``host1``-``hostn``. |
|
36 | controller from ``host0`` to hosts ``host1``-``hostn``. | |
31 | 3. Start the engines on hosts ``host1``-``hostn`` by running |
|
37 | 3. Start the engines on hosts ``host1``-``hostn`` by running | |
32 | :command:`ipengine`. This command has to be told where the FURL file |
|
38 | :command:`ipengine`. This command has to be told where the FURL file | |
33 | (:file:`ipcontroller-engine.furl`) is located. |
|
39 | (:file:`ipcontroller-engine.furl`) is located. | |
34 |
|
40 | |||
35 | At this point, the controller and engines will be connected. By default, the |
|
41 | At this point, the controller and engines will be connected. By default, the | |
36 | FURL files created by the controller are put into the |
|
42 | FURL files created by the controller are put into the | |
37 | :file:`~/.ipython/security` directory. If the engines share a filesystem with |
|
43 | :file:`~/.ipython/security` directory. If the engines share a filesystem with | |
38 | the controller, step 2 can be skipped as the engines will automatically look |
|
44 | the controller, step 2 can be skipped as the engines will automatically look | |
39 | at that location. |
|
45 | at that location. | |
40 |
|
46 | |||
41 | The final step required required to actually use the running controller from a |
|
47 | The final step required required to actually use the running controller from a | |
42 | client is to move the FURL files :file:`ipcontroller-mec.furl` and |
|
48 | client is to move the FURL files :file:`ipcontroller-mec.furl` and | |
43 | :file:`ipcontroller-tc.furl` from ``host0`` to the host where the clients will |
|
49 | :file:`ipcontroller-tc.furl` from ``host0`` to the host where the clients will | |
44 | be run. If these file are put into the :file:`~/.ipython/security` directory of the client's host, they will be found automatically. Otherwise, the full path to them has to be passed to the client's constructor. |
|
50 | be run. If these file are put into the :file:`~/.ipython/security` directory | |
|
51 | of the client's host, they will be found automatically. Otherwise, the full | |||
|
52 | path to them has to be passed to the client's constructor. | |||
45 |
|
53 | |||
46 | Using :command:`ipcluster` |
|
54 | Using :command:`ipcluster` | |
47 | ========================== |
|
55 | ========================== | |
48 |
|
56 | |||
49 |
The :command:`ipcluster` command provides a simple way of starting a |
|
57 | The :command:`ipcluster` command provides a simple way of starting a | |
|
58 | controller and engines in the following situations: | |||
50 |
|
59 | |||
51 | 1. When the controller and engines are all run on localhost. This is useful |
|
60 | 1. When the controller and engines are all run on localhost. This is useful | |
52 | for testing or running on a multicore computer. |
|
61 | for testing or running on a multicore computer. | |
53 | 2. When engines are started using the :command:`mpirun` command that comes |
|
62 | 2. When engines are started using the :command:`mpirun` command that comes | |
54 | with most MPI [MPI]_ implementations |
|
63 | with most MPI [MPI]_ implementations | |
55 | 3. When engines are started using the PBS [PBS]_ batch system. |
|
64 | 3. When engines are started using the PBS [PBS]_ batch system. | |
56 | 4. When the controller is started on localhost and the engines are started on |
|
65 | 4. When the controller is started on localhost and the engines are started on | |
57 | remote nodes using :command:`ssh`. |
|
66 | remote nodes using :command:`ssh`. | |
58 |
|
67 | |||
59 | .. note:: |
|
68 | .. note:: | |
60 |
|
69 | |||
61 | It is also possible for advanced users to add support to |
|
70 | It is also possible for advanced users to add support to | |
62 | :command:`ipcluster` for starting controllers and engines using other |
|
71 | :command:`ipcluster` for starting controllers and engines using other | |
63 | methods (like Sun's Grid Engine for example). |
|
72 | methods (like Sun's Grid Engine for example). | |
64 |
|
73 | |||
65 | .. note:: |
|
74 | .. note:: | |
66 |
|
75 | |||
67 | Currently :command:`ipcluster` requires that the |
|
76 | Currently :command:`ipcluster` requires that the | |
68 | :file:`~/.ipython/security` directory live on a shared filesystem that is |
|
77 | :file:`~/.ipython/security` directory live on a shared filesystem that is | |
69 | seen by both the controller and engines. If you don't have a shared file |
|
78 | seen by both the controller and engines. If you don't have a shared file | |
70 | system you will need to use :command:`ipcontroller` and |
|
79 | system you will need to use :command:`ipcontroller` and | |
71 | :command:`ipengine` directly. This constraint can be relaxed if you are |
|
80 | :command:`ipengine` directly. This constraint can be relaxed if you are | |
72 | using the :command:`ssh` method to start the cluster. |
|
81 | using the :command:`ssh` method to start the cluster. | |
73 |
|
82 | |||
74 | Underneath the hood, :command:`ipcluster` just uses :command:`ipcontroller` |
|
83 | Underneath the hood, :command:`ipcluster` just uses :command:`ipcontroller` | |
75 | and :command:`ipengine` to perform the steps described above. |
|
84 | and :command:`ipengine` to perform the steps described above. | |
76 |
|
85 | |||
77 | Using :command:`ipcluster` in local mode |
|
86 | Using :command:`ipcluster` in local mode | |
78 | ---------------------------------------- |
|
87 | ---------------------------------------- | |
79 |
|
88 | |||
80 | To start one controller and 4 engines on localhost, just do:: |
|
89 | To start one controller and 4 engines on localhost, just do:: | |
81 |
|
90 | |||
82 | $ ipcluster local -n 4 |
|
91 | $ ipcluster local -n 4 | |
83 |
|
92 | |||
84 | To see other command line options for the local mode, do:: |
|
93 | To see other command line options for the local mode, do:: | |
85 |
|
94 | |||
86 | $ ipcluster local -h |
|
95 | $ ipcluster local -h | |
87 |
|
96 | |||
88 | Using :command:`ipcluster` in mpiexec/mpirun mode |
|
97 | Using :command:`ipcluster` in mpiexec/mpirun mode | |
89 | ------------------------------------------------- |
|
98 | ------------------------------------------------- | |
90 |
|
99 | |||
91 | The mpiexec/mpirun mode is useful if you: |
|
100 | The mpiexec/mpirun mode is useful if you: | |
92 |
|
101 | |||
93 | 1. Have MPI installed. |
|
102 | 1. Have MPI installed. | |
94 | 2. Your systems are configured to use the :command:`mpiexec` or |
|
103 | 2. Your systems are configured to use the :command:`mpiexec` or | |
95 | :command:`mpirun` commands to start MPI processes. |
|
104 | :command:`mpirun` commands to start MPI processes. | |
96 |
|
105 | |||
97 | .. note:: |
|
106 | .. note:: | |
98 |
|
107 | |||
99 | The preferred command to use is :command:`mpiexec`. However, we also |
|
108 | The preferred command to use is :command:`mpiexec`. However, we also | |
100 | support :command:`mpirun` for backwards compatibility. The underlying |
|
109 | support :command:`mpirun` for backwards compatibility. The underlying | |
101 | logic used is exactly the same, the only difference being the name of the |
|
110 | logic used is exactly the same, the only difference being the name of the | |
102 | command line program that is called. |
|
111 | command line program that is called. | |
103 |
|
112 | |||
104 | If these are satisfied, you can start an IPython cluster using:: |
|
113 | If these are satisfied, you can start an IPython cluster using:: | |
105 |
|
114 | |||
106 | $ ipcluster mpiexec -n 4 |
|
115 | $ ipcluster mpiexec -n 4 | |
107 |
|
116 | |||
108 | This does the following: |
|
117 | This does the following: | |
109 |
|
118 | |||
110 | 1. Starts the IPython controller on current host. |
|
119 | 1. Starts the IPython controller on current host. | |
111 | 2. Uses :command:`mpiexec` to start 4 engines. |
|
120 | 2. Uses :command:`mpiexec` to start 4 engines. | |
112 |
|
121 | |||
113 | On newer MPI implementations (such as OpenMPI), this will work even if you don't make any calls to MPI or call :func:`MPI_Init`. However, older MPI implementations actually require each process to call :func:`MPI_Init` upon starting. The easiest way of having this done is to install the mpi4py [mpi4py]_ package and then call ipcluster with the ``--mpi`` option:: |
|
122 | On newer MPI implementations (such as OpenMPI), this will work even if you | |
|
123 | don't make any calls to MPI or call :func:`MPI_Init`. However, older MPI | |||
|
124 | implementations actually require each process to call :func:`MPI_Init` upon | |||
|
125 | starting. The easiest way of having this done is to install the mpi4py | |||
|
126 | [mpi4py]_ package and then call ipcluster with the ``--mpi`` option:: | |||
114 |
|
127 | |||
115 | $ ipcluster mpiexec -n 4 --mpi=mpi4py |
|
128 | $ ipcluster mpiexec -n 4 --mpi=mpi4py | |
116 |
|
129 | |||
117 | Unfortunately, even this won't work for some MPI implementations. If you are having problems with this, you will likely have to use a custom Python executable that itself calls :func:`MPI_Init` at the appropriate time. Fortunately, mpi4py comes with such a custom Python executable that is easy to install and use. However, this custom Python executable approach will not work with :command:`ipcluster` currently. |
|
130 | Unfortunately, even this won't work for some MPI implementations. If you are | |
|
131 | having problems with this, you will likely have to use a custom Python | |||
|
132 | executable that itself calls :func:`MPI_Init` at the appropriate time. | |||
|
133 | Fortunately, mpi4py comes with such a custom Python executable that is easy to | |||
|
134 | install and use. However, this custom Python executable approach will not work | |||
|
135 | with :command:`ipcluster` currently. | |||
118 |
|
136 | |||
119 | Additional command line options for this mode can be found by doing:: |
|
137 | Additional command line options for this mode can be found by doing:: | |
120 |
|
138 | |||
121 | $ ipcluster mpiexec -h |
|
139 | $ ipcluster mpiexec -h | |
122 |
|
140 | |||
123 | More details on using MPI with IPython can be found :ref:`here <parallelmpi>`. |
|
141 | More details on using MPI with IPython can be found :ref:`here <parallelmpi>`. | |
124 |
|
142 | |||
125 |
|
143 | |||
126 | Using :command:`ipcluster` in PBS mode |
|
144 | Using :command:`ipcluster` in PBS mode | |
127 | -------------------------------------- |
|
145 | -------------------------------------- | |
128 |
|
146 | |||
129 | The PBS mode uses the Portable Batch System [PBS]_ to start the engines. To use this mode, you first need to create a PBS script template that will be used to start the engines. Here is a sample PBS script template: |
|
147 | The PBS mode uses the Portable Batch System [PBS]_ to start the engines. To | |
|
148 | use this mode, you first need to create a PBS script template that will be | |||
|
149 | used to start the engines. Here is a sample PBS script template: | |||
130 |
|
150 | |||
131 | .. sourcecode:: bash |
|
151 | .. sourcecode:: bash | |
132 |
|
152 | |||
133 | #PBS -N ipython |
|
153 | #PBS -N ipython | |
134 | #PBS -j oe |
|
154 | #PBS -j oe | |
135 | #PBS -l walltime=00:10:00 |
|
155 | #PBS -l walltime=00:10:00 | |
136 | #PBS -l nodes=${n/4}:ppn=4 |
|
156 | #PBS -l nodes=${n/4}:ppn=4 | |
137 | #PBS -q parallel |
|
157 | #PBS -q parallel | |
138 |
|
158 | |||
139 | cd $$PBS_O_WORKDIR |
|
159 | cd $$PBS_O_WORKDIR | |
140 | export PATH=$$HOME/usr/local/bin |
|
160 | export PATH=$$HOME/usr/local/bin | |
141 | export PYTHONPATH=$$HOME/usr/local/lib/python2.4/site-packages |
|
161 | export PYTHONPATH=$$HOME/usr/local/lib/python2.4/site-packages | |
142 | /usr/local/bin/mpiexec -n ${n} ipengine --logfile=$$PBS_O_WORKDIR/ipengine |
|
162 | /usr/local/bin/mpiexec -n ${n} ipengine --logfile=$$PBS_O_WORKDIR/ipengine | |
143 |
|
163 | |||
144 | There are a few important points about this template: |
|
164 | There are a few important points about this template: | |
145 |
|
165 | |||
146 | 1. This template will be rendered at runtime using IPython's :mod:`Itpl` |
|
166 | 1. This template will be rendered at runtime using IPython's :mod:`Itpl` | |
147 | template engine. |
|
167 | template engine. | |
148 |
|
168 | |||
149 | 2. Instead of putting in the actual number of engines, use the notation |
|
169 | 2. Instead of putting in the actual number of engines, use the notation | |
150 | ``${n}`` to indicate the number of engines to be started. You can also uses |
|
170 | ``${n}`` to indicate the number of engines to be started. You can also uses | |
151 | expressions like ``${n/4}`` in the template to indicate the number of |
|
171 | expressions like ``${n/4}`` in the template to indicate the number of | |
152 | nodes. |
|
172 | nodes. | |
153 |
|
173 | |||
154 | 3. Because ``$`` is a special character used by the template engine, you must |
|
174 | 3. Because ``$`` is a special character used by the template engine, you must | |
155 | escape any ``$`` by using ``$$``. This is important when referring to |
|
175 | escape any ``$`` by using ``$$``. This is important when referring to | |
156 | environment variables in the template. |
|
176 | environment variables in the template. | |
157 |
|
177 | |||
158 | 4. Any options to :command:`ipengine` should be given in the batch script |
|
178 | 4. Any options to :command:`ipengine` should be given in the batch script | |
159 | template. |
|
179 | template. | |
160 |
|
180 | |||
161 | 5. Depending on the configuration of you system, you may have to set |
|
181 | 5. Depending on the configuration of you system, you may have to set | |
162 | environment variables in the script template. |
|
182 | environment variables in the script template. | |
163 |
|
183 | |||
164 |
Once you have created such a script, save it with a name like |
|
184 | Once you have created such a script, save it with a name like | |
|
185 | :file:`pbs.template`. Now you are ready to start your job:: | |||
165 |
|
186 | |||
166 | $ ipcluster pbs -n 128 --pbs-script=pbs.template |
|
187 | $ ipcluster pbs -n 128 --pbs-script=pbs.template | |
167 |
|
188 | |||
168 | Additional command line options for this mode can be found by doing:: |
|
189 | Additional command line options for this mode can be found by doing:: | |
169 |
|
190 | |||
170 | $ ipcluster pbs -h |
|
191 | $ ipcluster pbs -h | |
171 |
|
192 | |||
172 | Using :command:`ipcluster` in SSH mode |
|
193 | Using :command:`ipcluster` in SSH mode | |
173 | -------------------------------------- |
|
194 | -------------------------------------- | |
174 |
|
195 | |||
175 | The SSH mode uses :command:`ssh` to execute :command:`ipengine` on remote |
|
196 | The SSH mode uses :command:`ssh` to execute :command:`ipengine` on remote | |
176 | nodes and the :command:`ipcontroller` on localhost. |
|
197 | nodes and the :command:`ipcontroller` on localhost. | |
177 |
|
198 | |||
178 |
When using using this mode it highly recommended that you have set up SSH keys |
|
199 | When using using this mode it highly recommended that you have set up SSH keys | |
|
200 | and are using ssh-agent [SSH]_ for password-less logins. | |||
179 |
|
201 | |||
180 |
To use this mode you need a python file describing the cluster, here is an |
|
202 | To use this mode you need a python file describing the cluster, here is an | |
|
203 | example of such a "clusterfile": | |||
181 |
|
204 | |||
182 | .. sourcecode:: python |
|
205 | .. sourcecode:: python | |
183 |
|
206 | |||
184 | send_furl = True |
|
207 | send_furl = True | |
185 | engines = { 'host1.example.com' : 2, |
|
208 | engines = { 'host1.example.com' : 2, | |
186 | 'host2.example.com' : 5, |
|
209 | 'host2.example.com' : 5, | |
187 | 'host3.example.com' : 1, |
|
210 | 'host3.example.com' : 1, | |
188 | 'host4.example.com' : 8 } |
|
211 | 'host4.example.com' : 8 } | |
189 |
|
212 | |||
190 |
Since this is a regular python file usual python syntax applies. Things to |
|
213 | Since this is a regular python file usual python syntax applies. Things to | |
|
214 | note: | |||
191 |
|
215 | |||
192 | * The `engines` dict, where the keys is the host we want to run engines on and |
|
216 | * The `engines` dict, where the keys is the host we want to run engines on and | |
193 | the value is the number of engines to run on that host. |
|
217 | the value is the number of engines to run on that host. | |
194 | * send_furl can either be `True` or `False`, if `True` it will copy over the |
|
218 | * send_furl can either be `True` or `False`, if `True` it will copy over the | |
195 | furl needed for :command:`ipengine` to each host. |
|
219 | furl needed for :command:`ipengine` to each host. | |
196 |
|
220 | |||
197 | The ``--clusterfile`` command line option lets you specify the file to use for |
|
221 | The ``--clusterfile`` command line option lets you specify the file to use for | |
198 | the cluster definition. Once you have your cluster file and you can |
|
222 | the cluster definition. Once you have your cluster file and you can | |
199 | :command:`ssh` into the remote hosts with out an password you are ready to |
|
223 | :command:`ssh` into the remote hosts with out an password you are ready to | |
200 | start your cluster like so: |
|
224 | start your cluster like so: | |
201 |
|
225 | |||
202 | .. sourcecode:: bash |
|
226 | .. sourcecode:: bash | |
203 |
|
227 | |||
204 | $ ipcluster ssh --clusterfile /path/to/my/clusterfile.py |
|
228 | $ ipcluster ssh --clusterfile /path/to/my/clusterfile.py | |
205 |
|
229 | |||
206 |
|
230 | |||
207 |
Two helper shell scripts are used to start and stop :command:`ipengine` on |
|
231 | Two helper shell scripts are used to start and stop :command:`ipengine` on | |
|
232 | remote hosts: | |||
208 |
|
233 | |||
209 | * sshx.sh |
|
234 | * sshx.sh | |
210 | * engine_killer.sh |
|
235 | * engine_killer.sh | |
211 |
|
236 | |||
212 | Defaults for both of these are contained in the source code for :command:`ipcluster`. The default scripts are written to a local file in a tmep directory and then copied to a temp directory on the remote host and executed from there. On most Unix, Linux and OS X systems this is /tmp. |
|
237 | Defaults for both of these are contained in the source code for | |
|
238 | :command:`ipcluster`. The default scripts are written to a local file in a | |||
|
239 | tmep directory and then copied to a temp directory on the remote host and | |||
|
240 | executed from there. On most Unix, Linux and OS X systems this is /tmp. | |||
213 |
|
241 | |||
214 | The default sshx.sh is the following: |
|
242 | The default sshx.sh is the following: | |
215 |
|
243 | |||
216 | .. sourcecode:: bash |
|
244 | .. sourcecode:: bash | |
217 |
|
245 | |||
218 | #!/bin/sh |
|
246 | #!/bin/sh | |
219 | "$@" &> /dev/null & |
|
247 | "$@" &> /dev/null & | |
220 | echo $! |
|
248 | echo $! | |
221 |
|
249 | |||
222 | If you want to use a custom sshx.sh script you need to use the ``--sshx`` |
|
250 | If you want to use a custom sshx.sh script you need to use the ``--sshx`` | |
223 | option and specify the file to use. Using a custom sshx.sh file could be |
|
251 | option and specify the file to use. Using a custom sshx.sh file could be | |
224 | helpful when you need to setup the environment on the remote host before |
|
252 | helpful when you need to setup the environment on the remote host before | |
225 | executing :command:`ipengine`. |
|
253 | executing :command:`ipengine`. | |
226 |
|
254 | |||
227 | For a detailed options list: |
|
255 | For a detailed options list: | |
228 |
|
256 | |||
229 | .. sourcecode:: bash |
|
257 | .. sourcecode:: bash | |
230 |
|
258 | |||
231 | $ ipcluster ssh -h |
|
259 | $ ipcluster ssh -h | |
232 |
|
260 | |||
233 | Current limitations of the SSH mode of :command:`ipcluster` are: |
|
261 | Current limitations of the SSH mode of :command:`ipcluster` are: | |
234 |
|
262 | |||
235 | * Untested on Windows. Would require a working :command:`ssh` on Windows. |
|
263 | * Untested on Windows. Would require a working :command:`ssh` on Windows. | |
236 | Also, we are using shell scripts to setup and execute commands on remote |
|
264 | Also, we are using shell scripts to setup and execute commands on remote | |
237 | hosts. |
|
265 | hosts. | |
238 | * :command:`ipcontroller` is started on localhost, with no option to start it |
|
266 | * :command:`ipcontroller` is started on localhost, with no option to start it | |
239 | on a remote node. |
|
267 | on a remote node. | |
240 |
|
268 | |||
241 | Using the :command:`ipcontroller` and :command:`ipengine` commands |
|
269 | Using the :command:`ipcontroller` and :command:`ipengine` commands | |
242 | ================================================================== |
|
270 | ================================================================== | |
243 |
|
271 | |||
244 | It is also possible to use the :command:`ipcontroller` and :command:`ipengine` commands to start your controller and engines. This approach gives you full control over all aspects of the startup process. |
|
272 | It is also possible to use the :command:`ipcontroller` and :command:`ipengine` | |
|
273 | commands to start your controller and engines. This approach gives you full | |||
|
274 | control over all aspects of the startup process. | |||
245 |
|
275 | |||
246 | Starting the controller and engine on your local machine |
|
276 | Starting the controller and engine on your local machine | |
247 | -------------------------------------------------------- |
|
277 | -------------------------------------------------------- | |
248 |
|
278 | |||
249 | To use :command:`ipcontroller` and :command:`ipengine` to start things on your |
|
279 | To use :command:`ipcontroller` and :command:`ipengine` to start things on your | |
250 | local machine, do the following. |
|
280 | local machine, do the following. | |
251 |
|
281 | |||
252 | First start the controller:: |
|
282 | First start the controller:: | |
253 |
|
283 | |||
254 | $ ipcontroller |
|
284 | $ ipcontroller | |
255 |
|
285 | |||
256 |
Next, start however many instances of the engine you want using (repeatedly) |
|
286 | Next, start however many instances of the engine you want using (repeatedly) | |
|
287 | the command:: | |||
257 |
|
288 | |||
258 | $ ipengine |
|
289 | $ ipengine | |
259 |
|
290 | |||
260 | The engines should start and automatically connect to the controller using the FURL files in :file:`~./ipython/security`. You are now ready to use the controller and engines from IPython. |
|
291 | The engines should start and automatically connect to the controller using the | |
|
292 | FURL files in :file:`~./ipython/security`. You are now ready to use the | |||
|
293 | controller and engines from IPython. | |||
261 |
|
294 | |||
262 | .. warning:: |
|
295 | .. warning:: | |
263 |
|
296 | |||
264 | The order of the above operations is very important. You *must* |
|
297 | The order of the above operations is very important. You *must* | |
265 | start the controller before the engines, since the engines connect |
|
298 | start the controller before the engines, since the engines connect | |
266 | to the controller as they get started. |
|
299 | to the controller as they get started. | |
267 |
|
300 | |||
268 | .. note:: |
|
301 | .. note:: | |
269 |
|
302 | |||
270 | On some platforms (OS X), to put the controller and engine into the |
|
303 | On some platforms (OS X), to put the controller and engine into the | |
271 | background you may need to give these commands in the form ``(ipcontroller |
|
304 | background you may need to give these commands in the form ``(ipcontroller | |
272 | &)`` and ``(ipengine &)`` (with the parentheses) for them to work |
|
305 | &)`` and ``(ipengine &)`` (with the parentheses) for them to work | |
273 | properly. |
|
306 | properly. | |
274 |
|
307 | |||
275 | Starting the controller and engines on different hosts |
|
308 | Starting the controller and engines on different hosts | |
276 | ------------------------------------------------------ |
|
309 | ------------------------------------------------------ | |
277 |
|
310 | |||
278 | When the controller and engines are running on different hosts, things are |
|
311 | When the controller and engines are running on different hosts, things are | |
279 | slightly more complicated, but the underlying ideas are the same: |
|
312 | slightly more complicated, but the underlying ideas are the same: | |
280 |
|
313 | |||
281 | 1. Start the controller on a host using :command:`ipcontroller`. |
|
314 | 1. Start the controller on a host using :command:`ipcontroller`. | |
282 |
2. Copy :file:`ipcontroller-engine.furl` from :file:`~./ipython/security` on |
|
315 | 2. Copy :file:`ipcontroller-engine.furl` from :file:`~./ipython/security` on | |
|
316 | the controller's host to the host where the engines will run. | |||
283 | 3. Use :command:`ipengine` on the engine's hosts to start the engines. |
|
317 | 3. Use :command:`ipengine` on the engine's hosts to start the engines. | |
284 |
|
318 | |||
285 |
The only thing you have to be careful of is to tell :command:`ipengine` where |
|
319 | The only thing you have to be careful of is to tell :command:`ipengine` where | |
|
320 | the :file:`ipcontroller-engine.furl` file is located. There are two ways you | |||
|
321 | can do this: | |||
286 |
|
322 | |||
287 | * Put :file:`ipcontroller-engine.furl` in the :file:`~./ipython/security` |
|
323 | * Put :file:`ipcontroller-engine.furl` in the :file:`~./ipython/security` | |
288 | directory on the engine's host, where it will be found automatically. |
|
324 | directory on the engine's host, where it will be found automatically. | |
289 | * Call :command:`ipengine` with the ``--furl-file=full_path_to_the_file`` |
|
325 | * Call :command:`ipengine` with the ``--furl-file=full_path_to_the_file`` | |
290 | flag. |
|
326 | flag. | |
291 |
|
327 | |||
292 | The ``--furl-file`` flag works like this:: |
|
328 | The ``--furl-file`` flag works like this:: | |
293 |
|
329 | |||
294 | $ ipengine --furl-file=/path/to/my/ipcontroller-engine.furl |
|
330 | $ ipengine --furl-file=/path/to/my/ipcontroller-engine.furl | |
295 |
|
331 | |||
296 | .. note:: |
|
332 | .. note:: | |
297 |
|
333 | |||
298 | If the controller's and engine's hosts all have a shared file system |
|
334 | If the controller's and engine's hosts all have a shared file system | |
299 | (:file:`~./ipython/security` is the same on all of them), then things |
|
335 | (:file:`~./ipython/security` is the same on all of them), then things | |
300 | will just work! |
|
336 | will just work! | |
301 |
|
337 | |||
302 | Make FURL files persistent |
|
338 | Make FURL files persistent | |
303 | --------------------------- |
|
339 | --------------------------- | |
304 |
|
340 | |||
305 | At fist glance it may seem that that managing the FURL files is a bit |
|
341 | At fist glance it may seem that that managing the FURL files is a bit | |
306 | annoying. Going back to the house and key analogy, copying the FURL around |
|
342 | annoying. Going back to the house and key analogy, copying the FURL around | |
307 | each time you start the controller is like having to make a new key every time |
|
343 | each time you start the controller is like having to make a new key every time | |
308 | you want to unlock the door and enter your house. As with your house, you want |
|
344 | you want to unlock the door and enter your house. As with your house, you want | |
309 | to be able to create the key (or FURL file) once, and then simply use it at |
|
345 | to be able to create the key (or FURL file) once, and then simply use it at | |
310 | any point in the future. |
|
346 | any point in the future. | |
311 |
|
347 | |||
312 | This is possible, but before you do this, you **must** remove any old FURL |
|
348 | This is possible, but before you do this, you **must** remove any old FURL | |
313 | files in the :file:`~/.ipython/security` directory. |
|
349 | files in the :file:`~/.ipython/security` directory. | |
314 |
|
350 | |||
315 | .. warning:: |
|
351 | .. warning:: | |
316 |
|
352 | |||
317 | You **must** remove old FURL files before using persistent FURL files. |
|
353 | You **must** remove old FURL files before using persistent FURL files. | |
318 |
|
354 | |||
319 | Then, The only thing you have to do is decide what ports the controller will |
|
355 | Then, The only thing you have to do is decide what ports the controller will | |
320 | listen on for the engines and clients. This is done as follows:: |
|
356 | listen on for the engines and clients. This is done as follows:: | |
321 |
|
357 | |||
322 | $ ipcontroller -r --client-port=10101 --engine-port=10102 |
|
358 | $ ipcontroller -r --client-port=10101 --engine-port=10102 | |
323 |
|
359 | |||
324 | These options also work with all of the various modes of |
|
360 | These options also work with all of the various modes of | |
325 | :command:`ipcluster`:: |
|
361 | :command:`ipcluster`:: | |
326 |
|
362 | |||
327 | $ ipcluster local -n 2 -r --client-port=10101 --engine-port=10102 |
|
363 | $ ipcluster local -n 2 -r --client-port=10101 --engine-port=10102 | |
328 |
|
364 | |||
329 | Then, just copy the furl files over the first time and you are set. You can |
|
365 | Then, just copy the furl files over the first time and you are set. You can | |
330 | start and stop the controller and engines any many times as you want in the |
|
366 | start and stop the controller and engines any many times as you want in the | |
331 | future, just make sure to tell the controller to use the *same* ports. |
|
367 | future, just make sure to tell the controller to use the *same* ports. | |
332 |
|
368 | |||
333 | .. note:: |
|
369 | .. note:: | |
334 |
|
370 | |||
335 | You may ask the question: what ports does the controller listen on if you |
|
371 | You may ask the question: what ports does the controller listen on if you | |
336 | don't tell is to use specific ones? The default is to use high random port |
|
372 | don't tell is to use specific ones? The default is to use high random port | |
337 | numbers. We do this for two reasons: i) to increase security through |
|
373 | numbers. We do this for two reasons: i) to increase security through | |
338 | obscurity and ii) to multiple controllers on a given host to start and |
|
374 | obscurity and ii) to multiple controllers on a given host to start and | |
339 | automatically use different ports. |
|
375 | automatically use different ports. | |
340 |
|
376 | |||
341 | Log files |
|
377 | Log files | |
342 | --------- |
|
378 | --------- | |
343 |
|
379 | |||
344 | All of the components of IPython have log files associated with them. |
|
380 | All of the components of IPython have log files associated with them. | |
345 | These log files can be extremely useful in debugging problems with |
|
381 | These log files can be extremely useful in debugging problems with | |
346 | IPython and can be found in the directory :file:`~/.ipython/log`. Sending |
|
382 | IPython and can be found in the directory :file:`~/.ipython/log`. Sending | |
347 | the log files to us will often help us to debug any problems. |
|
383 | the log files to us will often help us to debug any problems. | |
348 |
|
384 | |||
349 |
|
385 | |||
350 | .. [PBS] Portable Batch System. http://www.openpbs.org/ |
|
386 | .. [PBS] Portable Batch System. http://www.openpbs.org/ | |
351 | .. [SSH] SSH-Agent http://en.wikipedia.org/wiki/Ssh-agent |
|
387 | .. [SSH] SSH-Agent http://en.wikipedia.org/wiki/Ssh-agent |
@@ -1,363 +1,364 b'' | |||||
1 | .. _parallelsecurity: |
|
1 | .. _parallelsecurity: | |
2 |
|
2 | |||
3 | =========================== |
|
3 | =========================== | |
4 | Security details of IPython |
|
4 | Security details of IPython | |
5 | =========================== |
|
5 | =========================== | |
6 |
|
6 | |||
7 | IPython's :mod:`IPython.kernel` package exposes the full power of the Python |
|
7 | IPython's :mod:`IPython.kernel` package exposes the full power of the Python | |
8 | interpreter over a TCP/IP network for the purposes of parallel computing. This |
|
8 | interpreter over a TCP/IP network for the purposes of parallel computing. This | |
9 | feature brings up the important question of IPython's security model. This |
|
9 | feature brings up the important question of IPython's security model. This | |
10 | document gives details about this model and how it is implemented in IPython's |
|
10 | document gives details about this model and how it is implemented in IPython's | |
11 | architecture. |
|
11 | architecture. | |
12 |
|
12 | |||
13 | Processs and network topology |
|
13 | Processs and network topology | |
14 | ============================= |
|
14 | ============================= | |
15 |
|
15 | |||
16 | To enable parallel computing, IPython has a number of different processes that |
|
16 | To enable parallel computing, IPython has a number of different processes that | |
17 | run. These processes are discussed at length in the IPython documentation and |
|
17 | run. These processes are discussed at length in the IPython documentation and | |
18 | are summarized here: |
|
18 | are summarized here: | |
19 |
|
19 | |||
20 | * The IPython *engine*. This process is a full blown Python |
|
20 | * The IPython *engine*. This process is a full blown Python | |
21 | interpreter in which user code is executed. Multiple |
|
21 | interpreter in which user code is executed. Multiple | |
22 | engines are started to make parallel computing possible. |
|
22 | engines are started to make parallel computing possible. | |
23 | * The IPython *controller*. This process manages a set of |
|
23 | * The IPython *controller*. This process manages a set of | |
24 | engines, maintaining a queue for each and presenting |
|
24 | engines, maintaining a queue for each and presenting | |
25 | an asynchronous interface to the set of engines. |
|
25 | an asynchronous interface to the set of engines. | |
26 | * The IPython *client*. This process is typically an |
|
26 | * The IPython *client*. This process is typically an | |
27 | interactive Python process that is used to coordinate the |
|
27 | interactive Python process that is used to coordinate the | |
28 | engines to get a parallel computation done. |
|
28 | engines to get a parallel computation done. | |
29 |
|
29 | |||
30 | Collectively, these three processes are called the IPython *kernel*. |
|
30 | Collectively, these three processes are called the IPython *kernel*. | |
31 |
|
31 | |||
32 | These three processes communicate over TCP/IP connections with a well defined |
|
32 | These three processes communicate over TCP/IP connections with a well defined | |
33 | topology. The IPython controller is the only process that listens on TCP/IP |
|
33 | topology. The IPython controller is the only process that listens on TCP/IP | |
34 | sockets. Upon starting, an engine connects to a controller and registers |
|
34 | sockets. Upon starting, an engine connects to a controller and registers | |
35 | itself with the controller. These engine/controller TCP/IP connections persist |
|
35 | itself with the controller. These engine/controller TCP/IP connections persist | |
36 | for the lifetime of each engine. |
|
36 | for the lifetime of each engine. | |
37 |
|
37 | |||
38 | The IPython client also connects to the controller using one or more TCP/IP |
|
38 | The IPython client also connects to the controller using one or more TCP/IP | |
39 | connections. These connections persist for the lifetime of the client only. |
|
39 | connections. These connections persist for the lifetime of the client only. | |
40 |
|
40 | |||
41 | A given IPython controller and set of engines typically has a relatively short |
|
41 | A given IPython controller and set of engines typically has a relatively short | |
42 | lifetime. Typically this lifetime corresponds to the duration of a single |
|
42 | lifetime. Typically this lifetime corresponds to the duration of a single | |
43 | parallel simulation performed by a single user. Finally, the controller, |
|
43 | parallel simulation performed by a single user. Finally, the controller, | |
44 | engines and client processes typically execute with the permissions of that |
|
44 | engines and client processes typically execute with the permissions of that | |
45 | same user. More specifically, the controller and engines are *not* executed as |
|
45 | same user. More specifically, the controller and engines are *not* executed as | |
46 | root or with any other superuser permissions. |
|
46 | root or with any other superuser permissions. | |
47 |
|
47 | |||
48 | Application logic |
|
48 | Application logic | |
49 | ================= |
|
49 | ================= | |
50 |
|
50 | |||
51 | When running the IPython kernel to perform a parallel computation, a user |
|
51 | When running the IPython kernel to perform a parallel computation, a user | |
52 | utilizes the IPython client to send Python commands and data through the |
|
52 | utilizes the IPython client to send Python commands and data through the | |
53 | IPython controller to the IPython engines, where those commands are executed |
|
53 | IPython controller to the IPython engines, where those commands are executed | |
54 | and the data processed. The design of IPython ensures that the client is the |
|
54 | and the data processed. The design of IPython ensures that the client is the | |
55 |
only access point for the capabilities of the engines. |
|
55 | only access point for the capabilities of the engines. That is, the only way | |
|
56 | of addressing the engines is through a client. | |||
56 |
|
57 | |||
57 | A user can utilize the client to instruct the IPython engines to execute |
|
58 | A user can utilize the client to instruct the IPython engines to execute | |
58 | arbitrary Python commands. These Python commands can include calls to the |
|
59 | arbitrary Python commands. These Python commands can include calls to the | |
59 | system shell, access the filesystem, etc., as required by the user's |
|
60 | system shell, access the filesystem, etc., as required by the user's | |
60 | application code. From this perspective, when a user runs an IPython engine on |
|
61 | application code. From this perspective, when a user runs an IPython engine on | |
61 | a host, that engine has the same capabilities and permissions as the user |
|
62 | a host, that engine has the same capabilities and permissions as the user | |
62 | themselves (as if they were logged onto the engine's host with a terminal). |
|
63 | themselves (as if they were logged onto the engine's host with a terminal). | |
63 |
|
64 | |||
64 | Secure network connections |
|
65 | Secure network connections | |
65 | ========================== |
|
66 | ========================== | |
66 |
|
67 | |||
67 | Overview |
|
68 | Overview | |
68 | -------- |
|
69 | -------- | |
69 |
|
70 | |||
70 | All TCP/IP connections between the client and controller as well as the |
|
71 | All TCP/IP connections between the client and controller as well as the | |
71 | engines and controller are fully encrypted and authenticated. This section |
|
72 | engines and controller are fully encrypted and authenticated. This section | |
72 | describes the details of the encryption and authentication approached used |
|
73 | describes the details of the encryption and authentication approached used | |
73 | within IPython. |
|
74 | within IPython. | |
74 |
|
75 | |||
75 | IPython uses the Foolscap network protocol [Foolscap]_ for all communications |
|
76 | IPython uses the Foolscap network protocol [Foolscap]_ for all communications | |
76 | between processes. Thus, the details of IPython's security model are directly |
|
77 | between processes. Thus, the details of IPython's security model are directly | |
77 | related to those of Foolscap. Thus, much of the following discussion is |
|
78 | related to those of Foolscap. Thus, much of the following discussion is | |
78 | actually just a discussion of the security that is built in to Foolscap. |
|
79 | actually just a discussion of the security that is built in to Foolscap. | |
79 |
|
80 | |||
80 | Encryption |
|
81 | Encryption | |
81 | ---------- |
|
82 | ---------- | |
82 |
|
83 | |||
83 | For encryption purposes, IPython and Foolscap use the well known Secure Socket |
|
84 | For encryption purposes, IPython and Foolscap use the well known Secure Socket | |
84 | Layer (SSL) protocol [RFC5246]_. We use the implementation of this protocol |
|
85 | Layer (SSL) protocol [RFC5246]_. We use the implementation of this protocol | |
85 | provided by the OpenSSL project through the pyOpenSSL [pyOpenSSL]_ Python |
|
86 | provided by the OpenSSL project through the pyOpenSSL [pyOpenSSL]_ Python | |
86 | bindings to OpenSSL. |
|
87 | bindings to OpenSSL. | |
87 |
|
88 | |||
88 | Authentication |
|
89 | Authentication | |
89 | -------------- |
|
90 | -------------- | |
90 |
|
91 | |||
91 | IPython clients and engines must also authenticate themselves with the |
|
92 | IPython clients and engines must also authenticate themselves with the | |
92 | controller. This is handled in a capabilities based security model |
|
93 | controller. This is handled in a capabilities based security model | |
93 | [Capability]_. In this model, the controller creates a strong cryptographic |
|
94 | [Capability]_. In this model, the controller creates a strong cryptographic | |
94 | key or token that represents each set of capability that the controller |
|
95 | key or token that represents each set of capability that the controller | |
95 | offers. Any party who has this key and presents it to the controller has full |
|
96 | offers. Any party who has this key and presents it to the controller has full | |
96 | access to the corresponding capabilities of the controller. This model is |
|
97 | access to the corresponding capabilities of the controller. This model is | |
97 | analogous to using a physical key to gain access to physical items |
|
98 | analogous to using a physical key to gain access to physical items | |
98 | (capabilities) behind a locked door. |
|
99 | (capabilities) behind a locked door. | |
99 |
|
100 | |||
100 | For a capabilities based authentication system to prevent unauthorized access, |
|
101 | For a capabilities based authentication system to prevent unauthorized access, | |
101 | two things must be ensured: |
|
102 | two things must be ensured: | |
102 |
|
103 | |||
103 | * The keys must be cryptographically strong. Otherwise attackers could gain |
|
104 | * The keys must be cryptographically strong. Otherwise attackers could gain | |
104 | access by a simple brute force key guessing attack. |
|
105 | access by a simple brute force key guessing attack. | |
105 | * The actual keys must be distributed only to authorized parties. |
|
106 | * The actual keys must be distributed only to authorized parties. | |
106 |
|
107 | |||
107 | The keys in Foolscap are called Foolscap URL's or FURLs. The following section |
|
108 | The keys in Foolscap are called Foolscap URL's or FURLs. The following section | |
108 | gives details about how these FURLs are created in Foolscap. The IPython |
|
109 | gives details about how these FURLs are created in Foolscap. The IPython | |
109 | controller creates a number of FURLs for different purposes: |
|
110 | controller creates a number of FURLs for different purposes: | |
110 |
|
111 | |||
111 | * One FURL that grants IPython engines access to the controller. Also |
|
112 | * One FURL that grants IPython engines access to the controller. Also | |
112 | implicit in this access is permission to execute code sent by an |
|
113 | implicit in this access is permission to execute code sent by an | |
113 | authenticated IPython client. |
|
114 | authenticated IPython client. | |
114 | * Two or more FURLs that grant IPython clients access to the controller. |
|
115 | * Two or more FURLs that grant IPython clients access to the controller. | |
115 | Implicit in this access is permission to give the controller's engine code |
|
116 | Implicit in this access is permission to give the controller's engine code | |
116 | to execute. |
|
117 | to execute. | |
117 |
|
118 | |||
118 | Upon starting, the controller creates these different FURLS and writes them |
|
119 | Upon starting, the controller creates these different FURLS and writes them | |
119 |
files in the user-read-only directory :file:`$HOME/.ipython/security`. Thus, |
|
120 | files in the user-read-only directory :file:`$HOME/.ipython/security`. Thus, | |
120 | user who starts the controller has access to the FURLs. |
|
121 | only the user who starts the controller has access to the FURLs. | |
121 |
|
122 | |||
122 | For an IPython client or engine to authenticate with a controller, it must |
|
123 | For an IPython client or engine to authenticate with a controller, it must | |
123 | present the appropriate FURL to the controller upon connecting. If the |
|
124 | present the appropriate FURL to the controller upon connecting. If the | |
124 | FURL matches what the controller expects for a given capability, access is |
|
125 | FURL matches what the controller expects for a given capability, access is | |
125 | granted. If not, access is denied. The exchange of FURLs is done after |
|
126 | granted. If not, access is denied. The exchange of FURLs is done after | |
126 | encrypted communications channels have been established to prevent attackers |
|
127 | encrypted communications channels have been established to prevent attackers | |
127 | from capturing them. |
|
128 | from capturing them. | |
128 |
|
129 | |||
129 | .. note:: |
|
130 | .. note:: | |
130 |
|
131 | |||
131 | The FURL is similar to an unsigned private key in SSH. |
|
132 | The FURL is similar to an unsigned private key in SSH. | |
132 |
|
133 | |||
133 | Details of the Foolscap handshake |
|
134 | Details of the Foolscap handshake | |
134 | --------------------------------- |
|
135 | --------------------------------- | |
135 |
|
136 | |||
136 | In this section we detail the precise security handshake that takes place at |
|
137 | In this section we detail the precise security handshake that takes place at | |
137 | the beginning of any network connection in IPython. For the purposes of this |
|
138 | the beginning of any network connection in IPython. For the purposes of this | |
138 | discussion, the SERVER is the IPython controller process and the CLIENT is the |
|
139 | discussion, the SERVER is the IPython controller process and the CLIENT is the | |
139 | IPython engine or client process. |
|
140 | IPython engine or client process. | |
140 |
|
141 | |||
141 | Upon starting, all IPython processes do the following: |
|
142 | Upon starting, all IPython processes do the following: | |
142 |
|
143 | |||
143 | 1. Create a public key x509 certificate (ISO/IEC 9594). |
|
144 | 1. Create a public key x509 certificate (ISO/IEC 9594). | |
144 | 2. Create a hash of the contents of the certificate using the SHA-1 algorithm. |
|
145 | 2. Create a hash of the contents of the certificate using the SHA-1 algorithm. | |
145 | The base-32 encoded version of this hash is saved by the process as its |
|
146 | The base-32 encoded version of this hash is saved by the process as its | |
146 | process id (actually in Foolscap, this is the Tub id, but here refer to |
|
147 | process id (actually in Foolscap, this is the Tub id, but here refer to | |
147 | it as the process id). |
|
148 | it as the process id). | |
148 |
|
149 | |||
149 | Upon starting, the IPython controller also does the following: |
|
150 | Upon starting, the IPython controller also does the following: | |
150 |
|
151 | |||
151 | 1. Save the x509 certificate to disk in a secure location. The CLIENT |
|
152 | 1. Save the x509 certificate to disk in a secure location. The CLIENT | |
152 | certificate is never saved to disk. |
|
153 | certificate is never saved to disk. | |
153 | 2. Create a FURL for each capability that the controller has. There are |
|
154 | 2. Create a FURL for each capability that the controller has. There are | |
154 | separate capabilities the controller offers for clients and engines. The |
|
155 | separate capabilities the controller offers for clients and engines. The | |
155 | FURL is created using: a) the process id of the SERVER, b) the IP |
|
156 | FURL is created using: a) the process id of the SERVER, b) the IP | |
156 | address and port the SERVER is listening on and c) a 160 bit, |
|
157 | address and port the SERVER is listening on and c) a 160 bit, | |
157 | cryptographically secure string that represents the capability (the |
|
158 | cryptographically secure string that represents the capability (the | |
158 | "capability id"). |
|
159 | "capability id"). | |
159 | 3. The FURLs are saved to disk in a secure location on the SERVER's host. |
|
160 | 3. The FURLs are saved to disk in a secure location on the SERVER's host. | |
160 |
|
161 | |||
161 | For a CLIENT to be able to connect to the SERVER and access a capability of |
|
162 | For a CLIENT to be able to connect to the SERVER and access a capability of | |
162 | that SERVER, the CLIENT must have knowledge of the FURL for that SERVER's |
|
163 | that SERVER, the CLIENT must have knowledge of the FURL for that SERVER's | |
163 | capability. This typically requires that the file containing the FURL be |
|
164 | capability. This typically requires that the file containing the FURL be | |
164 | moved from the SERVER's host to the CLIENT's host. This is done by the end |
|
165 | moved from the SERVER's host to the CLIENT's host. This is done by the end | |
165 | user who started the SERVER and wishes to have a CLIENT connect to the SERVER. |
|
166 | user who started the SERVER and wishes to have a CLIENT connect to the SERVER. | |
166 |
|
167 | |||
167 | When a CLIENT connects to the SERVER, the following handshake protocol takes |
|
168 | When a CLIENT connects to the SERVER, the following handshake protocol takes | |
168 | place: |
|
169 | place: | |
169 |
|
170 | |||
170 | 1. The CLIENT tells the SERVER what process (or Tub) id it expects the SERVER |
|
171 | 1. The CLIENT tells the SERVER what process (or Tub) id it expects the SERVER | |
171 | to have. |
|
172 | to have. | |
172 | 2. If the SERVER has that process id, it notifies the CLIENT that it will now |
|
173 | 2. If the SERVER has that process id, it notifies the CLIENT that it will now | |
173 | enter encrypted mode. If the SERVER has a different id, the SERVER aborts. |
|
174 | enter encrypted mode. If the SERVER has a different id, the SERVER aborts. | |
174 | 3. Both CLIENT and SERVER initiate the SSL handshake protocol. |
|
175 | 3. Both CLIENT and SERVER initiate the SSL handshake protocol. | |
175 | 4. Both CLIENT and SERVER request the certificate of their peer and verify |
|
176 | 4. Both CLIENT and SERVER request the certificate of their peer and verify | |
176 | that certificate. If this succeeds, all further communications are |
|
177 | that certificate. If this succeeds, all further communications are | |
177 | encrypted. |
|
178 | encrypted. | |
178 | 5. Both CLIENT and SERVER send a hello block containing connection parameters |
|
179 | 5. Both CLIENT and SERVER send a hello block containing connection parameters | |
179 | and their process id. |
|
180 | and their process id. | |
180 | 6. The CLIENT and SERVER check that their peer's stated process id matches the |
|
181 | 6. The CLIENT and SERVER check that their peer's stated process id matches the | |
181 | hash of the x509 certificate the peer presented. If not, the connection is |
|
182 | hash of the x509 certificate the peer presented. If not, the connection is | |
182 | aborted. |
|
183 | aborted. | |
183 | 7. The CLIENT verifies that the SERVER's stated id matches the id of the |
|
184 | 7. The CLIENT verifies that the SERVER's stated id matches the id of the | |
184 | SERVER the CLIENT is intending to connect to. If not, the connection is |
|
185 | SERVER the CLIENT is intending to connect to. If not, the connection is | |
185 | aborted. |
|
186 | aborted. | |
186 | 8. The CLIENT and SERVER elect a master who decides on the final connection |
|
187 | 8. The CLIENT and SERVER elect a master who decides on the final connection | |
187 | parameters. |
|
188 | parameters. | |
188 |
|
189 | |||
189 | The public/private key pair associated with each process's x509 certificate |
|
190 | The public/private key pair associated with each process's x509 certificate | |
190 | are completely hidden from this handshake protocol. There are however, used |
|
191 | are completely hidden from this handshake protocol. There are however, used | |
191 | internally by OpenSSL as part of the SSL handshake protocol. Each process |
|
192 | internally by OpenSSL as part of the SSL handshake protocol. Each process | |
192 | keeps their own private key hidden and sends its peer only the public key |
|
193 | keeps their own private key hidden and sends its peer only the public key | |
193 | (embedded in the certificate). |
|
194 | (embedded in the certificate). | |
194 |
|
195 | |||
195 | Finally, when the CLIENT requests access to a particular SERVER capability, |
|
196 | Finally, when the CLIENT requests access to a particular SERVER capability, | |
196 | the following happens: |
|
197 | the following happens: | |
197 |
|
198 | |||
198 | 1. The CLIENT asks the SERVER for access to a capability by presenting that |
|
199 | 1. The CLIENT asks the SERVER for access to a capability by presenting that | |
199 | capabilities id. |
|
200 | capabilities id. | |
200 | 2. If the SERVER has a capability with that id, access is granted. If not, |
|
201 | 2. If the SERVER has a capability with that id, access is granted. If not, | |
201 | access is not granted. |
|
202 | access is not granted. | |
202 | 3. Once access has been gained, the CLIENT can use the capability. |
|
203 | 3. Once access has been gained, the CLIENT can use the capability. | |
203 |
|
204 | |||
204 | Specific security vulnerabilities |
|
205 | Specific security vulnerabilities | |
205 | ================================= |
|
206 | ================================= | |
206 |
|
207 | |||
207 | There are a number of potential security vulnerabilities present in IPython's |
|
208 | There are a number of potential security vulnerabilities present in IPython's | |
208 | architecture. In this section we discuss those vulnerabilities and detail how |
|
209 | architecture. In this section we discuss those vulnerabilities and detail how | |
209 | the security architecture described above prevents them from being exploited. |
|
210 | the security architecture described above prevents them from being exploited. | |
210 |
|
211 | |||
211 | Unauthorized clients |
|
212 | Unauthorized clients | |
212 | -------------------- |
|
213 | -------------------- | |
213 |
|
214 | |||
214 | The IPython client can instruct the IPython engines to execute arbitrary |
|
215 | The IPython client can instruct the IPython engines to execute arbitrary | |
215 | Python code with the permissions of the user who started the engines. If an |
|
216 | Python code with the permissions of the user who started the engines. If an | |
216 | attacker were able to connect their own hostile IPython client to the IPython |
|
217 | attacker were able to connect their own hostile IPython client to the IPython | |
217 | controller, they could instruct the engines to execute code. |
|
218 | controller, they could instruct the engines to execute code. | |
218 |
|
219 | |||
219 | This attack is prevented by the capabilities based client authentication |
|
220 | This attack is prevented by the capabilities based client authentication | |
220 | performed after the encrypted channel has been established. The relevant |
|
221 | performed after the encrypted channel has been established. The relevant | |
221 | authentication information is encoded into the FURL that clients must |
|
222 | authentication information is encoded into the FURL that clients must | |
222 | present to gain access to the IPython controller. By limiting the distribution |
|
223 | present to gain access to the IPython controller. By limiting the distribution | |
223 | of those FURLs, a user can grant access to only authorized persons. |
|
224 | of those FURLs, a user can grant access to only authorized persons. | |
224 |
|
225 | |||
225 | It is highly unlikely that a client FURL could be guessed by an attacker |
|
226 | It is highly unlikely that a client FURL could be guessed by an attacker | |
226 | in a brute force guessing attack. A given instance of the IPython controller |
|
227 | in a brute force guessing attack. A given instance of the IPython controller | |
227 | only runs for a relatively short amount of time (on the order of hours). Thus |
|
228 | only runs for a relatively short amount of time (on the order of hours). Thus | |
228 | an attacker would have only a limited amount of time to test a search space of |
|
229 | an attacker would have only a limited amount of time to test a search space of | |
229 | size 2**320. Furthermore, even if a controller were to run for a longer amount |
|
230 | size 2**320. Furthermore, even if a controller were to run for a longer amount | |
230 | of time, this search space is quite large (larger for instance than that of |
|
231 | of time, this search space is quite large (larger for instance than that of | |
231 | typical username/password pair). |
|
232 | typical username/password pair). | |
232 |
|
233 | |||
233 | Unauthorized engines |
|
234 | Unauthorized engines | |
234 | -------------------- |
|
235 | -------------------- | |
235 |
|
236 | |||
236 | If an attacker were able to connect a hostile engine to a user's controller, |
|
237 | If an attacker were able to connect a hostile engine to a user's controller, | |
237 | the user might unknowingly send sensitive code or data to the hostile engine. |
|
238 | the user might unknowingly send sensitive code or data to the hostile engine. | |
238 | This attacker's engine would then have full access to that code and data. |
|
239 | This attacker's engine would then have full access to that code and data. | |
239 |
|
240 | |||
240 | This type of attack is prevented in the same way as the unauthorized client |
|
241 | This type of attack is prevented in the same way as the unauthorized client | |
241 | attack, through the usage of the capabilities based authentication scheme. |
|
242 | attack, through the usage of the capabilities based authentication scheme. | |
242 |
|
243 | |||
243 | Unauthorized controllers |
|
244 | Unauthorized controllers | |
244 | ------------------------ |
|
245 | ------------------------ | |
245 |
|
246 | |||
246 | It is also possible that an attacker could try to convince a user's IPython |
|
247 | It is also possible that an attacker could try to convince a user's IPython | |
247 | client or engine to connect to a hostile IPython controller. That controller |
|
248 | client or engine to connect to a hostile IPython controller. That controller | |
248 | would then have full access to the code and data sent between the IPython |
|
249 | would then have full access to the code and data sent between the IPython | |
249 | client and the IPython engines. |
|
250 | client and the IPython engines. | |
250 |
|
251 | |||
251 | Again, this attack is prevented through the FURLs, which ensure that a |
|
252 | Again, this attack is prevented through the FURLs, which ensure that a | |
252 | client or engine connects to the correct controller. It is also important to |
|
253 | client or engine connects to the correct controller. It is also important to | |
253 | note that the FURLs also encode the IP address and port that the |
|
254 | note that the FURLs also encode the IP address and port that the | |
254 | controller is listening on, so there is little chance of mistakenly connecting |
|
255 | controller is listening on, so there is little chance of mistakenly connecting | |
255 | to a controller running on a different IP address and port. |
|
256 | to a controller running on a different IP address and port. | |
256 |
|
257 | |||
257 | When starting an engine or client, a user must specify which FURL to use |
|
258 | When starting an engine or client, a user must specify which FURL to use | |
258 | for that connection. Thus, in order to introduce a hostile controller, the |
|
259 | for that connection. Thus, in order to introduce a hostile controller, the | |
259 | attacker must convince the user to use the FURLs associated with the |
|
260 | attacker must convince the user to use the FURLs associated with the | |
260 | hostile controller. As long as a user is diligent in only using FURLs from |
|
261 | hostile controller. As long as a user is diligent in only using FURLs from | |
261 | trusted sources, this attack is not possible. |
|
262 | trusted sources, this attack is not possible. | |
262 |
|
263 | |||
263 | Other security measures |
|
264 | Other security measures | |
264 | ======================= |
|
265 | ======================= | |
265 |
|
266 | |||
266 | A number of other measures are taken to further limit the security risks |
|
267 | A number of other measures are taken to further limit the security risks | |
267 | involved in running the IPython kernel. |
|
268 | involved in running the IPython kernel. | |
268 |
|
269 | |||
269 | First, by default, the IPython controller listens on random port numbers. |
|
270 | First, by default, the IPython controller listens on random port numbers. | |
270 | While this can be overridden by the user, in the default configuration, an |
|
271 | While this can be overridden by the user, in the default configuration, an | |
271 | attacker would have to do a port scan to even find a controller to attack. |
|
272 | attacker would have to do a port scan to even find a controller to attack. | |
272 | When coupled with the relatively short running time of a typical controller |
|
273 | When coupled with the relatively short running time of a typical controller | |
273 | (on the order of hours), an attacker would have to work extremely hard and |
|
274 | (on the order of hours), an attacker would have to work extremely hard and | |
274 | extremely *fast* to even find a running controller to attack. |
|
275 | extremely *fast* to even find a running controller to attack. | |
275 |
|
276 | |||
276 | Second, much of the time, especially when run on supercomputers or clusters, |
|
277 | Second, much of the time, especially when run on supercomputers or clusters, | |
277 | the controller is running behind a firewall. Thus, for engines or client to |
|
278 | the controller is running behind a firewall. Thus, for engines or client to | |
278 | connect to the controller: |
|
279 | connect to the controller: | |
279 |
|
280 | |||
280 | * The different processes have to all be behind the firewall. |
|
281 | * The different processes have to all be behind the firewall. | |
281 |
|
282 | |||
282 | or: |
|
283 | or: | |
283 |
|
284 | |||
284 | * The user has to use SSH port forwarding to tunnel the |
|
285 | * The user has to use SSH port forwarding to tunnel the | |
285 | connections through the firewall. |
|
286 | connections through the firewall. | |
286 |
|
287 | |||
287 | In either case, an attacker is presented with addition barriers that prevent |
|
288 | In either case, an attacker is presented with addition barriers that prevent | |
288 | attacking or even probing the system. |
|
289 | attacking or even probing the system. | |
289 |
|
290 | |||
290 | Summary |
|
291 | Summary | |
291 | ======= |
|
292 | ======= | |
292 |
|
293 | |||
293 | IPython's architecture has been carefully designed with security in mind. The |
|
294 | IPython's architecture has been carefully designed with security in mind. The | |
294 | capabilities based authentication model, in conjunction with the encrypted |
|
295 | capabilities based authentication model, in conjunction with the encrypted | |
295 | TCP/IP channels, address the core potential vulnerabilities in the system, |
|
296 | TCP/IP channels, address the core potential vulnerabilities in the system, | |
296 | while still enabling user's to use the system in open networks. |
|
297 | while still enabling user's to use the system in open networks. | |
297 |
|
298 | |||
298 | Other questions |
|
299 | Other questions | |
299 | =============== |
|
300 | =============== | |
300 |
|
301 | |||
301 | About keys |
|
302 | About keys | |
302 | ---------- |
|
303 | ---------- | |
303 |
|
304 | |||
304 | Can you clarify the roles of the certificate and its keys versus the FURL, |
|
305 | Can you clarify the roles of the certificate and its keys versus the FURL, | |
305 | which is also called a key? |
|
306 | which is also called a key? | |
306 |
|
307 | |||
307 | The certificate created by IPython processes is a standard public key x509 |
|
308 | The certificate created by IPython processes is a standard public key x509 | |
308 | certificate, that is used by the SSL handshake protocol to setup encrypted |
|
309 | certificate, that is used by the SSL handshake protocol to setup encrypted | |
309 | channel between the controller and the IPython engine or client. This public |
|
310 | channel between the controller and the IPython engine or client. This public | |
310 | and private key associated with this certificate are used only by the SSL |
|
311 | and private key associated with this certificate are used only by the SSL | |
311 | handshake protocol in setting up this encrypted channel. |
|
312 | handshake protocol in setting up this encrypted channel. | |
312 |
|
313 | |||
313 | The FURL serves a completely different and independent purpose from the |
|
314 | The FURL serves a completely different and independent purpose from the | |
314 | key pair associated with the certificate. When we refer to a FURL as a |
|
315 | key pair associated with the certificate. When we refer to a FURL as a | |
315 | key, we are using the word "key" in the capabilities based security model |
|
316 | key, we are using the word "key" in the capabilities based security model | |
316 | sense. This has nothing to do with "key" in the public/private key sense used |
|
317 | sense. This has nothing to do with "key" in the public/private key sense used | |
317 | in the SSL protocol. |
|
318 | in the SSL protocol. | |
318 |
|
319 | |||
319 | With that said the FURL is used as an cryptographic key, to grant |
|
320 | With that said the FURL is used as an cryptographic key, to grant | |
320 | IPython engines and clients access to particular capabilities that the |
|
321 | IPython engines and clients access to particular capabilities that the | |
321 | controller offers. |
|
322 | controller offers. | |
322 |
|
323 | |||
323 | Self signed certificates |
|
324 | Self signed certificates | |
324 | ------------------------ |
|
325 | ------------------------ | |
325 |
|
326 | |||
326 | Is the controller creating a self-signed certificate? Is this created for per |
|
327 | Is the controller creating a self-signed certificate? Is this created for per | |
327 | instance/session, one-time-setup or each-time the controller is started? |
|
328 | instance/session, one-time-setup or each-time the controller is started? | |
328 |
|
329 | |||
329 | The Foolscap network protocol, which handles the SSL protocol details, creates |
|
330 | The Foolscap network protocol, which handles the SSL protocol details, creates | |
330 | a self-signed x509 certificate using OpenSSL for each IPython process. The |
|
331 | a self-signed x509 certificate using OpenSSL for each IPython process. The | |
331 | lifetime of the certificate is handled differently for the IPython controller |
|
332 | lifetime of the certificate is handled differently for the IPython controller | |
332 | and the engines/client. |
|
333 | and the engines/client. | |
333 |
|
334 | |||
334 | For the IPython engines and client, the certificate is only held in memory for |
|
335 | For the IPython engines and client, the certificate is only held in memory for | |
335 | the lifetime of its process. It is never written to disk. |
|
336 | the lifetime of its process. It is never written to disk. | |
336 |
|
337 | |||
337 | For the controller, the certificate can be created anew each time the |
|
338 | For the controller, the certificate can be created anew each time the | |
338 | controller starts or it can be created once and reused each time the |
|
339 | controller starts or it can be created once and reused each time the | |
339 | controller starts. If at any point, the certificate is deleted, a new one is |
|
340 | controller starts. If at any point, the certificate is deleted, a new one is | |
340 | created the next time the controller starts. |
|
341 | created the next time the controller starts. | |
341 |
|
342 | |||
342 | SSL private key |
|
343 | SSL private key | |
343 | --------------- |
|
344 | --------------- | |
344 |
|
345 | |||
345 | How the private key (associated with the certificate) is distributed? |
|
346 | How the private key (associated with the certificate) is distributed? | |
346 |
|
347 | |||
347 | In the usual implementation of the SSL protocol, the private key is never |
|
348 | In the usual implementation of the SSL protocol, the private key is never | |
348 | distributed. We follow this standard always. |
|
349 | distributed. We follow this standard always. | |
349 |
|
350 | |||
350 | SSL versus Foolscap authentication |
|
351 | SSL versus Foolscap authentication | |
351 | ---------------------------------- |
|
352 | ---------------------------------- | |
352 |
|
353 | |||
353 | Many SSL connections only perform one sided authentication (the server to the |
|
354 | Many SSL connections only perform one sided authentication (the server to the | |
354 | client). How is the client authentication in IPython's system related to SSL |
|
355 | client). How is the client authentication in IPython's system related to SSL | |
355 | authentication? |
|
356 | authentication? | |
356 |
|
357 | |||
357 | We perform a two way SSL handshake in which both parties request and verify |
|
358 | We perform a two way SSL handshake in which both parties request and verify | |
358 | the certificate of their peer. This mutual authentication is handled by the |
|
359 | the certificate of their peer. This mutual authentication is handled by the | |
359 | SSL handshake and is separate and independent from the additional |
|
360 | SSL handshake and is separate and independent from the additional | |
360 | authentication steps that the CLIENT and SERVER perform after an encrypted |
|
361 | authentication steps that the CLIENT and SERVER perform after an encrypted | |
361 | channel is established. |
|
362 | channel is established. | |
362 |
|
363 | |||
363 | .. [RFC5246] <http://tools.ietf.org/html/rfc5246> |
|
364 | .. [RFC5246] <http://tools.ietf.org/html/rfc5246> |
@@ -1,99 +1,120 b'' | |||||
1 | .. _paralleltask: |
|
1 | .. _paralleltask: | |
2 |
|
2 | |||
3 | ========================== |
|
3 | ========================== | |
4 | The IPython task interface |
|
4 | The IPython task interface | |
5 | ========================== |
|
5 | ========================== | |
6 |
|
6 | |||
7 | The task interface to the controller presents the engines as a fault tolerant, dynamic load-balanced system or workers. Unlike the multiengine interface, in the task interface, the user have no direct access to individual engines. In some ways, this interface is simpler, but in other ways it is more powerful. |
|
7 | The task interface to the controller presents the engines as a fault tolerant, | |
8 |
|
8 | dynamic load-balanced system or workers. Unlike the multiengine interface, in | ||
9 | Best of all the user can use both of these interfaces running at the same time to take advantage or both of their strengths. When the user can break up the user's work into segments that do not depend on previous execution, the task interface is ideal. But it also has more power and flexibility, allowing the user to guide the distribution of jobs, without having to assign tasks to engines explicitly. |
|
9 | the task interface, the user have no direct access to individual engines. In | |
|
10 | some ways, this interface is simpler, but in other ways it is more powerful. | |||
|
11 | ||||
|
12 | Best of all the user can use both of these interfaces running at the same time | |||
|
13 | to take advantage or both of their strengths. When the user can break up the | |||
|
14 | user's work into segments that do not depend on previous execution, the task | |||
|
15 | interface is ideal. But it also has more power and flexibility, allowing the | |||
|
16 | user to guide the distribution of jobs, without having to assign tasks to | |||
|
17 | engines explicitly. | |||
10 |
|
18 | |||
11 | Starting the IPython controller and engines |
|
19 | Starting the IPython controller and engines | |
12 | =========================================== |
|
20 | =========================================== | |
13 |
|
21 | |||
14 | To follow along with this tutorial, you will need to start the IPython |
|
22 | To follow along with this tutorial, you will need to start the IPython | |
15 | controller and four IPython engines. The simplest way of doing this is to use |
|
23 | controller and four IPython engines. The simplest way of doing this is to use | |
16 | the :command:`ipcluster` command:: |
|
24 | the :command:`ipcluster` command:: | |
17 |
|
25 | |||
18 | $ ipcluster local -n 4 |
|
26 | $ ipcluster local -n 4 | |
19 |
|
27 | |||
20 | For more detailed information about starting the controller and engines, see |
|
28 | For more detailed information about starting the controller and engines, see | |
21 | our :ref:`introduction <ip1par>` to using IPython for parallel computing. |
|
29 | our :ref:`introduction <ip1par>` to using IPython for parallel computing. | |
22 |
|
30 | |||
23 | Creating a ``TaskClient`` instance |
|
31 | Creating a ``TaskClient`` instance | |
24 | ========================================= |
|
32 | ========================================= | |
25 |
|
33 | |||
26 | The first step is to import the IPython :mod:`IPython.kernel.client` module |
|
34 | The first step is to import the IPython :mod:`IPython.kernel.client` module | |
27 | and then create a :class:`TaskClient` instance: |
|
35 | and then create a :class:`TaskClient` instance: | |
28 |
|
36 | |||
29 | .. sourcecode:: ipython |
|
37 | .. sourcecode:: ipython | |
30 |
|
38 | |||
31 | In [1]: from IPython.kernel import client |
|
39 | In [1]: from IPython.kernel import client | |
32 |
|
40 | |||
33 | In [2]: tc = client.TaskClient() |
|
41 | In [2]: tc = client.TaskClient() | |
34 |
|
42 | |||
35 | This form assumes that the :file:`ipcontroller-tc.furl` is in the |
|
43 | This form assumes that the :file:`ipcontroller-tc.furl` is in the | |
36 | :file:`~./ipython/security` directory on the client's host. If not, the |
|
44 | :file:`~./ipython/security` directory on the client's host. If not, the | |
37 | location of the FURL file must be given as an argument to the |
|
45 | location of the FURL file must be given as an argument to the | |
38 | constructor: |
|
46 | constructor: | |
39 |
|
47 | |||
40 | .. sourcecode:: ipython |
|
48 | .. sourcecode:: ipython | |
41 |
|
49 | |||
42 | In [2]: mec = client.TaskClient('/path/to/my/ipcontroller-tc.furl') |
|
50 | In [2]: mec = client.TaskClient('/path/to/my/ipcontroller-tc.furl') | |
43 |
|
51 | |||
44 | Quick and easy parallelism |
|
52 | Quick and easy parallelism | |
45 | ========================== |
|
53 | ========================== | |
46 |
|
54 | |||
47 | In many cases, you simply want to apply a Python function to a sequence of objects, but *in parallel*. Like the multiengine interface, the task interface provides two simple ways of accomplishing this: a parallel version of :func:`map` and ``@parallel`` function decorator. However, the verions in the task interface have one important difference: they are dynamically load balanced. Thus, if the execution time per item varies significantly, you should use the versions in the task interface. |
|
55 | In many cases, you simply want to apply a Python function to a sequence of | |
|
56 | objects, but *in parallel*. Like the multiengine interface, the task interface | |||
|
57 | provides two simple ways of accomplishing this: a parallel version of | |||
|
58 | :func:`map` and ``@parallel`` function decorator. However, the verions in the | |||
|
59 | task interface have one important difference: they are dynamically load | |||
|
60 | balanced. Thus, if the execution time per item varies significantly, you | |||
|
61 | should use the versions in the task interface. | |||
48 |
|
62 | |||
49 | Parallel map |
|
63 | Parallel map | |
50 | ------------ |
|
64 | ------------ | |
51 |
|
65 | |||
52 |
The parallel :meth:`map` in the task interface is similar to that in the |
|
66 | The parallel :meth:`map` in the task interface is similar to that in the | |
|
67 | multiengine interface: | |||
53 |
|
68 | |||
54 | .. sourcecode:: ipython |
|
69 | .. sourcecode:: ipython | |
55 |
|
70 | |||
56 | In [63]: serial_result = map(lambda x:x**10, range(32)) |
|
71 | In [63]: serial_result = map(lambda x:x**10, range(32)) | |
57 |
|
72 | |||
58 | In [64]: parallel_result = tc.map(lambda x:x**10, range(32)) |
|
73 | In [64]: parallel_result = tc.map(lambda x:x**10, range(32)) | |
59 |
|
74 | |||
60 | In [65]: serial_result==parallel_result |
|
75 | In [65]: serial_result==parallel_result | |
61 | Out[65]: True |
|
76 | Out[65]: True | |
62 |
|
77 | |||
63 | Parallel function decorator |
|
78 | Parallel function decorator | |
64 | --------------------------- |
|
79 | --------------------------- | |
65 |
|
80 | |||
66 | Parallel functions are just like normal function, but they can be called on sequences and *in parallel*. The multiengine interface provides a decorator that turns any Python function into a parallel function: |
|
81 | Parallel functions are just like normal function, but they can be called on | |
|
82 | sequences and *in parallel*. The multiengine interface provides a decorator | |||
|
83 | that turns any Python function into a parallel function: | |||
67 |
|
84 | |||
68 | .. sourcecode:: ipython |
|
85 | .. sourcecode:: ipython | |
69 |
|
86 | |||
70 | In [10]: @tc.parallel() |
|
87 | In [10]: @tc.parallel() | |
71 | ....: def f(x): |
|
88 | ....: def f(x): | |
72 | ....: return 10.0*x**4 |
|
89 | ....: return 10.0*x**4 | |
73 | ....: |
|
90 | ....: | |
74 |
|
91 | |||
75 | In [11]: f(range(32)) # this is done in parallel |
|
92 | In [11]: f(range(32)) # this is done in parallel | |
76 | Out[11]: |
|
93 | Out[11]: | |
77 | [0.0,10.0,160.0,...] |
|
94 | [0.0,10.0,160.0,...] | |
78 |
|
95 | |||
79 | More details |
|
96 | More details | |
80 | ============ |
|
97 | ============ | |
81 |
|
98 | |||
82 |
The :class:`TaskClient` has many more powerful features that allow quite a bit |
|
99 | The :class:`TaskClient` has many more powerful features that allow quite a bit | |
|
100 | of flexibility in how tasks are defined and run. The next places to look are | |||
|
101 | in the following classes: | |||
83 |
|
102 | |||
84 | * :class:`IPython.kernel.client.TaskClient` |
|
103 | * :class:`IPython.kernel.client.TaskClient` | |
85 | * :class:`IPython.kernel.client.StringTask` |
|
104 | * :class:`IPython.kernel.client.StringTask` | |
86 | * :class:`IPython.kernel.client.MapTask` |
|
105 | * :class:`IPython.kernel.client.MapTask` | |
87 |
|
106 | |||
88 | The following is an overview of how to use these classes together: |
|
107 | The following is an overview of how to use these classes together: | |
89 |
|
108 | |||
90 | 1. Create a :class:`TaskClient`. |
|
109 | 1. Create a :class:`TaskClient`. | |
91 | 2. Create one or more instances of :class:`StringTask` or :class:`MapTask` |
|
110 | 2. Create one or more instances of :class:`StringTask` or :class:`MapTask` | |
92 | to define your tasks. |
|
111 | to define your tasks. | |
93 | 3. Submit your tasks to using the :meth:`run` method of your |
|
112 | 3. Submit your tasks to using the :meth:`run` method of your | |
94 | :class:`TaskClient` instance. |
|
113 | :class:`TaskClient` instance. | |
95 | 4. Use :meth:`TaskClient.get_task_result` to get the results of the |
|
114 | 4. Use :meth:`TaskClient.get_task_result` to get the results of the | |
96 | tasks. |
|
115 | tasks. | |
97 |
|
116 | |||
98 | We are in the process of developing more detailed information about the task interface. For now, the docstrings of the :class:`TaskClient`, :class:`StringTask` and :class:`MapTask` classes should be consulted. |
|
117 | We are in the process of developing more detailed information about the task | |
|
118 | interface. For now, the docstrings of the :class:`TaskClient`, | |||
|
119 | :class:`StringTask` and :class:`MapTask` classes should be consulted. | |||
99 |
|
120 |
1 | NO CONTENT: file was removed, binary diff hidden |
|
NO CONTENT: file was removed, binary diff hidden |
1 | NO CONTENT: file was removed, binary diff hidden |
|
NO CONTENT: file was removed, binary diff hidden |
1 | NO CONTENT: file was removed |
|
NO CONTENT: file was removed |
General Comments 0
You need to be logged in to leave comments.
Login now