ENH: Extend the DisplayHook.compute_result_repr() and write_result_repr() methods to produce and consume the lists of extra formats. Use this capability in the ZMQ shell. Document the extension to the messaging format.

IPython/core/displayhook.py

@@ -1,291 +1,301 b''
 # -*- coding: utf-8 -*-
 """Displayhook for IPython.
 
 Authors:
 
 * Fernando Perez
 * Brian Granger
 """
 
 #-----------------------------------------------------------------------------
 #       Copyright (C) 2008-2010 The IPython Development Team
 #       Copyright (C) 2001-2007 Fernando Perez <[e-mail]>
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import __builtin__
 
 from IPython.config.configurable import Configurable
 from IPython.core import prompts
 import IPython.utils.generics
 import IPython.utils.io
 from IPython.utils.traitlets import Instance, List
 from IPython.utils.warn import warn
 from IPython.core.formatters import DefaultFormatter
 
 #-----------------------------------------------------------------------------
 # Main displayhook class
 #-----------------------------------------------------------------------------
 
 # TODO: The DisplayHook class should be split into two classes, one that
 # manages the prompts and their synchronization and another that just does the
 # displayhook logic and calls into the prompt manager.
 
 # TODO: Move the various attributes (cache_size, colors, input_sep,
 # output_sep, output_sep2, ps1, ps2, ps_out, pad_left). Some of these are also
 # attributes of InteractiveShell. They should be on ONE object only and the
 # other objects should ask that one object for their values.
 
 class DisplayHook(Configurable):
     """The custom IPython displayhook to replace sys.displayhook.
 
     This class does many things, but the basic idea is that it is a callable
     that gets called anytime user code returns a value.
 
     Currently this class does more than just the displayhook logic and that
     extra logic should eventually be moved out of here.
     """
 
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC')
 
     # The default formatter.
     default_formatter = Instance('IPython.core.formatters.FormatterABC')
     def _default_formatter_default(self):
         # FIXME: backwards compatibility for the InteractiveShell.pprint option?
         return DefaultFormatter(config=self.config)
 
     # Any additional FormatterABC instances we use.
     # FIXME: currently unused.
     extra_formatters = List(config=True)
 
     # Each call to the In[] prompt raises it by 1, even the first.
     #prompt_count = Int(0)
 
     def __init__(self, shell=None, cache_size=1000,
                  colors='NoColor', input_sep='\n',
                  output_sep='\n', output_sep2='',
                  ps1 = None, ps2 = None, ps_out = None, pad_left=True,
                  config=None):
         super(DisplayHook, self).__init__(shell=shell, config=config)
 
         cache_size_min = 3
         if cache_size <= 0:
             self.do_full_cache = 0
             cache_size = 0
         elif cache_size < cache_size_min:
             self.do_full_cache = 0
             cache_size = 0
             warn('caching was disabled (min value for cache size is %s).' %
                  cache_size_min,level=3)
         else:
             self.do_full_cache = 1
 
         self.cache_size = cache_size
         self.input_sep = input_sep
 
         # we need a reference to the user-level namespace
         self.shell = shell
 
         # Set input prompt strings and colors
         if cache_size == 0:
             if ps1.find('%n') > -1 or ps1.find(r'\#') > -1 \
                    or ps1.find(r'\N') > -1:
                 ps1 = '>>> '
             if ps2.find('%n') > -1 or ps2.find(r'\#') > -1 \
                    or ps2.find(r'\N') > -1:
                 ps2 = '... '
         self.ps1_str = self._set_prompt_str(ps1,'In [\\#]: ','>>> ')
         self.ps2_str = self._set_prompt_str(ps2,'   .\\D.: ','... ')
         self.ps_out_str = self._set_prompt_str(ps_out,'Out[\\#]: ','')
 
         self.color_table = prompts.PromptColors
         self.prompt1 = prompts.Prompt1(self,sep=input_sep,prompt=self.ps1_str,
                                pad_left=pad_left)
         self.prompt2 = prompts.Prompt2(self,prompt=self.ps2_str,pad_left=pad_left)
         self.prompt_out = prompts.PromptOut(self,sep='',prompt=self.ps_out_str,
                                     pad_left=pad_left)
         self.set_colors(colors)
 
         # Store the last prompt string each time, we need it for aligning
         # continuation and auto-rewrite prompts
         self.last_prompt = ''
         self.output_sep = output_sep
         self.output_sep2 = output_sep2
         self._,self.__,self.___ = '','',''
 
         # these are deliberately global:
         to_user_ns = {'_':self._,'__':self.__,'___':self.___}
         self.shell.user_ns.update(to_user_ns)
 
     @property
     def prompt_count(self):
         return self.shell.execution_count
 
     def _set_prompt_str(self,p_str,cache_def,no_cache_def):
         if p_str is None:
             if self.do_full_cache:
                 return cache_def
             else:
                 return no_cache_def
         else:
             return p_str
                 
     def set_colors(self, colors):
         """Set the active color scheme and configure colors for the three
         prompt subsystems."""
 
         # FIXME: This modifying of the global prompts.prompt_specials needs
         # to be fixed. We need to refactor all of the prompts stuff to use
         # proper configuration and traits notifications.
         if colors.lower()=='nocolor':
             prompts.prompt_specials = prompts.prompt_specials_nocolor
         else:
             prompts.prompt_specials = prompts.prompt_specials_color
         
         self.color_table.set_active_scheme(colors)
         self.prompt1.set_colors()
         self.prompt2.set_colors()
         self.prompt_out.set_colors()
 
     #-------------------------------------------------------------------------
     # Methods used in __call__. Override these methods to modify the behavior
     # of the displayhook.
     #-------------------------------------------------------------------------
 
     def check_for_underscore(self):
         """Check if the user has set the '_' variable by hand."""
         # If something injected a '_' variable in __builtin__, delete
         # ipython's automatic one so we don't clobber that.  gettext() in
         # particular uses _, so we need to stay away from it.
         if '_' in __builtin__.__dict__:
             try:
                 del self.shell.user_ns['_']
             except KeyError:
                 pass
 
     def quiet(self):
         """Should we silence the display hook because of ';'?"""
         # do not print output if input ends in ';'
         try:
             if self.shell.input_hist[self.prompt_count].endswith(';\n'):
                 return True
         except IndexError:
             # some uses of ipshellembed may fail here
             pass
         return False
 
     def start_displayhook(self):
         """Start the displayhook, initializing resources."""
         pass
 
     def write_output_prompt(self):
         """Write the output prompt."""
         # Use write, not print which adds an extra space.
         IPython.utils.io.Term.cout.write(self.output_sep)
         outprompt = str(self.prompt_out)
         if self.do_full_cache:
             IPython.utils.io.Term.cout.write(outprompt)
 
     def compute_result_repr(self, result):
         """Compute and return the repr of the object to be displayed.
 
         This method only compute the string form of the repr and should NOT
         actual print or write that to a stream.
         """
         result_repr = self.default_formatter(result)
         if '\n' in result_repr:
             # So that multi-line strings line up with the left column of
             # the screen, instead of having the output prompt mess up
             # their first line.
             outprompt = str(self.prompt_out)
             if outprompt and not outprompt.endswith('\n'):
                 # But avoid extraneous empty lines.
                 result_repr = '\n' + result_repr
 
-        return result_repr
+        extra_formats = []
+        for f in self.extra_formatters:
+            try:
+                data = f(result)
+            except Exception:
+                # FIXME: log the exception.
+                continue
+            if data is not None:
+                extra_formats.append((f.id, f.format, data))
+
+        return result_repr, extra_formats
 
-    def write_result_repr(self, result_repr):
+    def write_result_repr(self, result_repr, extra_formats):
         # We want to print because we want to always make sure we have a 
         # newline, even if all the prompt separators are ''. This is the
         # standard IPython behavior.
         print >>IPython.utils.io.Term.cout, result_repr
 
     def update_user_ns(self, result):
         """Update user_ns with various things like _, __, _1, etc."""
 
         # Avoid recursive reference when displaying _oh/Out
         if result is not self.shell.user_ns['_oh']:
             if len(self.shell.user_ns['_oh']) >= self.cache_size and self.do_full_cache:
                 warn('Output cache limit (currently '+
                       `self.cache_size`+' entries) hit.\n'
                      'Flushing cache and resetting history counter...\n'
                      'The only history variables available will be _,__,___ and _1\n'
                      'with the current result.')
 
                 self.flush()
             # Don't overwrite '_' and friends if '_' is in __builtin__ (otherwise
             # we cause buggy behavior for things like gettext).
             if '_' not in __builtin__.__dict__:
                 self.___ = self.__
                 self.__ = self._
                 self._ = result
                 self.shell.user_ns.update({'_':self._,'__':self.__,'___':self.___})
 
             # hackish access to top-level  namespace to create _1,_2... dynamically
             to_main = {}
             if self.do_full_cache:
                 new_result = '_'+`self.prompt_count`
                 to_main[new_result] = result
                 self.shell.user_ns.update(to_main)
                 self.shell.user_ns['_oh'][self.prompt_count] = result
 
     def log_output(self, result):
         """Log the output."""
         if self.shell.logger.log_output:
             self.shell.logger.log_write(repr(result), 'output')
 
     def finish_displayhook(self):
         """Finish up all displayhook activities."""
         IPython.utils.io.Term.cout.write(self.output_sep2)
         IPython.utils.io.Term.cout.flush()
 
     def __call__(self, result=None):
         """Printing with history cache management.
         
         This is invoked everytime the interpreter needs to print, and is
         activated by setting the variable sys.displayhook to it.
         """
         self.check_for_underscore()
         if result is not None and not self.quiet():
             self.start_displayhook()
             self.write_output_prompt()
-            result_repr = self.compute_result_repr(result)
-            self.write_result_repr(result_repr)
+            result_repr, extra_formats = self.compute_result_repr(result)
+            self.write_result_repr(result_repr, extra_formats)
             self.update_user_ns(result)
             self.log_output(result)
             self.finish_displayhook()
 
     def flush(self):
         if not self.do_full_cache:
             raise ValueError,"You shouldn't have reached the cache flush "\
                   "if full caching is not enabled!"
         # delete auto-generated vars from global namespace
         
         for n in range(1,self.prompt_count + 1):
             key = '_'+`n`
             try:
                 del self.shell.user_ns[key]
             except: pass
         self.shell.user_ns['_oh'].clear()
         
         if '_' not in __builtin__.__dict__:
             self.shell.user_ns.update({'_':None,'__':None, '___':None})
         import gc
         gc.collect() # xxx needed?
 

IPython/zmq/zmqshell.py

@@ -1,580 +1,581 b''
 """A ZMQ-based subclass of InteractiveShell.
 
 This code is meant to ease the refactoring of the base InteractiveShell into
 something with a cleaner architecture for 2-process use, without actually
 breaking InteractiveShell itself.  So we're doing something a bit ugly, where
 we subclass and override what we want to fix.  Once this is working well, we
 can go back to the base class and refactor the code for a cleaner inheritance
 implementation that doesn't rely on so much monkeypatching.
 
 But this lets us maintain a fully working IPython as we develop the new
 machinery.  This should thus be thought of as scaffolding.
 """
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 # Stdlib
 import inspect
 import os
 import re
 
 # Our own
 from IPython.core.interactiveshell import (
     InteractiveShell, InteractiveShellABC
 )
 from IPython.core import page
 from IPython.core.displayhook import DisplayHook
 from IPython.core.macro import Macro
 from IPython.core.payloadpage import install_payload_page
 from IPython.utils import io
 from IPython.utils.path import get_py_filename
 from IPython.utils.text import StringTypes
 from IPython.utils.traitlets import Instance, Type, Dict
 from IPython.utils.warn import warn
 from IPython.zmq.session import extract_header
 from session import Session
 
 #-----------------------------------------------------------------------------
 # Globals and side-effects
 #-----------------------------------------------------------------------------
 
 # Install the payload version of page.
 install_payload_page()
 
 #-----------------------------------------------------------------------------
 # Functions and classes
 #-----------------------------------------------------------------------------
 
 class ZMQDisplayHook(DisplayHook):
 
     session = Instance(Session)
     pub_socket = Instance('zmq.Socket')
     parent_header = Dict({})
 
     def set_parent(self, parent):
         """Set the parent for outbound messages."""
         self.parent_header = extract_header(parent)
 
     def start_displayhook(self):
         self.msg = self.session.msg(u'pyout', {}, parent=self.parent_header)
 
     def write_output_prompt(self):
         """Write the output prompt."""
         if self.do_full_cache:
             self.msg['content']['execution_count'] = self.prompt_count
 
-    def write_result_repr(self, result_repr):
+    def write_result_repr(self, result_repr, extra_formats):
         self.msg['content']['data'] = result_repr
+        self.msg['content']['extra_formats'] = extra_formats
 
     def finish_displayhook(self):
         """Finish up all displayhook activities."""
         self.pub_socket.send_json(self.msg)
         self.msg = None
 
 
 class ZMQInteractiveShell(InteractiveShell):
     """A subclass of InteractiveShell for ZMQ."""
 
     displayhook_class = Type(ZMQDisplayHook)
     keepkernel_on_exit = None
 
     def init_environment(self):
         """Configure the user's environment.
 
         """
         env = os.environ
         # These two ensure 'ls' produces nice coloring on BSD-derived systems
         env['TERM'] = 'xterm-color'
         env['CLICOLOR'] = '1'
         # Since normal pagers don't work at all (over pexpect we don't have
         # single-key control of the subprocess), try to disable paging in
         # subprocesses as much as possible.
         env['PAGER'] = 'cat'
         env['GIT_PAGER'] = 'cat'
 
     def auto_rewrite_input(self, cmd):
         """Called to show the auto-rewritten input for autocall and friends.
 
         FIXME: this payload is currently not correctly processed by the
         frontend.
         """
         new = self.displayhook.prompt1.auto_rewrite() + cmd
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.auto_rewrite_input',
             transformed_input=new,
             )
         self.payload_manager.write_payload(payload)
         
     def ask_exit(self):
         """Engage the exit actions."""
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.ask_exit',
             exit=True,
             keepkernel=self.keepkernel_on_exit,
             )
         self.payload_manager.write_payload(payload)
 
     def _showtraceback(self, etype, evalue, stb):
 
         exc_content = {
             u'traceback' : stb,
             u'ename' : unicode(etype.__name__),
             u'evalue' : unicode(evalue)
         }
 
         dh = self.displayhook
         exc_msg = dh.session.msg(u'pyerr', exc_content, dh.parent_header)
         # Send exception info over pub socket for other clients than the caller
         # to pick up
         dh.pub_socket.send_json(exc_msg)
 
         # FIXME - Hack: store exception info in shell object.  Right now, the
         # caller is reading this info after the fact, we need to fix this logic
         # to remove this hack.  Even uglier, we need to store the error status
         # here, because in the main loop, the logic that sets it is being
         # skipped because runlines swallows the exceptions.
         exc_content[u'status'] = u'error'
         self._reply_content = exc_content
         # /FIXME
         
         return exc_content
 
     #------------------------------------------------------------------------
     # Magic overrides
     #------------------------------------------------------------------------
     # Once the base class stops inheriting from magic, this code needs to be
     # moved into a separate machinery as well.  For now, at least isolate here
     # the magics which this class needs to implement differently from the base
     # class, or that are unique to it.
 
     def magic_doctest_mode(self,parameter_s=''):
         """Toggle doctest mode on and off.
 
         This mode is intended to make IPython behave as much as possible like a
         plain Python shell, from the perspective of how its prompts, exceptions
         and output look.  This makes it easy to copy and paste parts of a
         session into doctests.  It does so by:
 
         - Changing the prompts to the classic ``>>>`` ones.
         - Changing the exception reporting mode to 'Plain'.
         - Disabling pretty-printing of output.
 
         Note that IPython also supports the pasting of code snippets that have
         leading '>>>' and '...' prompts in them.  This means that you can paste
         doctests from files or docstrings (even if they have leading
         whitespace), and the code will execute correctly.  You can then use
         '%history -t' to see the translated history; this will give you the
         input after removal of all the leading prompts and whitespace, which
         can be pasted back into an editor.
 
         With these features, you can switch into this mode easily whenever you
         need to do testing and changes to doctests, without having to leave
         your existing IPython session.
         """
 
         from IPython.utils.ipstruct import Struct
 
         # Shorthands
         shell = self.shell
         # dstore is a data store kept in the instance metadata bag to track any
         # changes we make, so we can undo them later.
         dstore = shell.meta.setdefault('doctest_mode', Struct())
         save_dstore = dstore.setdefault
 
         # save a few values we'll need to recover later
         mode = save_dstore('mode', False)
         save_dstore('rc_pprint', shell.pprint)
         save_dstore('xmode', shell.InteractiveTB.mode)
         
         if mode == False:
             # turn on
             shell.pprint = False
             shell.magic_xmode('Plain')
         else:
             # turn off
             shell.pprint = dstore.rc_pprint
             shell.magic_xmode(dstore.xmode)
 
         # Store new mode and inform on console
         dstore.mode = bool(1-int(mode))
         mode_label = ['OFF','ON'][dstore.mode]
         print('Doctest mode is:', mode_label)
         
         # Send the payload back so that clients can modify their prompt display
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.magic_doctest_mode',
             mode=dstore.mode)
         self.payload_manager.write_payload(payload)
 
     def magic_edit(self,parameter_s='',last_call=['','']):
         """Bring up an editor and execute the resulting code.
 
         Usage:
           %edit [options] [args]
 
         %edit runs IPython's editor hook.  The default version of this hook is
         set to call the __IPYTHON__.rc.editor command.  This is read from your
         environment variable $EDITOR.  If this isn't found, it will default to
         vi under Linux/Unix and to notepad under Windows.  See the end of this
         docstring for how to change the editor hook.
 
         You can also set the value of this editor via the command line option
         '-editor' or in your ipythonrc file. This is useful if you wish to use
         specifically for IPython an editor different from your typical default
         (and for Windows users who typically don't set environment variables).
 
         This command allows you to conveniently edit multi-line code right in
         your IPython session.
         
         If called without arguments, %edit opens up an empty editor with a
         temporary file and will execute the contents of this file when you
         close it (don't forget to save it!).
 
 
         Options:
 
         -n <number>: open the editor at a specified line number.  By default,
         the IPython editor hook uses the unix syntax 'editor +N filename', but
         you can configure this by providing your own modified hook if your
         favorite editor supports line-number specifications with a different
         syntax.
         
         -p: this will call the editor with the same data as the previous time
         it was used, regardless of how long ago (in your current session) it
         was.
 
         -r: use 'raw' input.  This option only applies to input taken from the
         user's history.  By default, the 'processed' history is used, so that
         magics are loaded in their transformed version to valid Python.  If
         this option is given, the raw input as typed as the command line is
         used instead.  When you exit the editor, it will be executed by
         IPython's own processor.
         
         -x: do not execute the edited code immediately upon exit. This is
         mainly useful if you are editing programs which need to be called with
         command line arguments, which you can then do using %run.
 
 
         Arguments:
 
         If arguments are given, the following possibilites exist:
 
         - The arguments are numbers or pairs of colon-separated numbers (like
         1 4:8 9). These are interpreted as lines of previous input to be
         loaded into the editor. The syntax is the same of the %macro command.
 
         - If the argument doesn't start with a number, it is evaluated as a
         variable and its contents loaded into the editor. You can thus edit
         any string which contains python code (including the result of
         previous edits).
 
         - If the argument is the name of an object (other than a string),
         IPython will try to locate the file where it was defined and open the
         editor at the point where it is defined. You can use `%edit function`
         to load an editor exactly at the point where 'function' is defined,
         edit it and have the file be executed automatically.
 
         If the object is a macro (see %macro for details), this opens up your
         specified editor with a temporary file containing the macro's data.
         Upon exit, the macro is reloaded with the contents of the file.
 
         Note: opening at an exact line is only supported under Unix, and some
         editors (like kedit and gedit up to Gnome 2.8) do not understand the
         '+NUMBER' parameter necessary for this feature. Good editors like
         (X)Emacs, vi, jed, pico and joe all do.
 
         - If the argument is not found as a variable, IPython will look for a
         file with that name (adding .py if necessary) and load it into the
         editor. It will execute its contents with execfile() when you exit,
         loading any code in the file into your interactive namespace.
 
         After executing your code, %edit will return as output the code you
         typed in the editor (except when it was an existing file). This way
         you can reload the code in further invocations of %edit as a variable,
         via _<NUMBER> or Out[<NUMBER>], where <NUMBER> is the prompt number of
         the output.
 
         Note that %edit is also available through the alias %ed.
 
         This is an example of creating a simple function inside the editor and
         then modifying it. First, start up the editor:
 
         In [1]: ed
         Editing... done. Executing edited code...
         Out[1]: 'def foo():n    print "foo() was defined in an editing session"n'
 
         We can then call the function foo():
         
         In [2]: foo()
         foo() was defined in an editing session
 
         Now we edit foo.  IPython automatically loads the editor with the
         (temporary) file where foo() was previously defined:
         
         In [3]: ed foo
         Editing... done. Executing edited code...
 
         And if we call foo() again we get the modified version:
         
         In [4]: foo()
         foo() has now been changed!
 
         Here is an example of how to edit a code snippet successive
         times. First we call the editor:
 
         In [5]: ed
         Editing... done. Executing edited code...
         hello
         Out[5]: "print 'hello'n"
 
         Now we call it again with the previous output (stored in _):
 
         In [6]: ed _
         Editing... done. Executing edited code...
         hello world
         Out[6]: "print 'hello world'n"
 
         Now we call it with the output #8 (stored in _8, also as Out[8]):
 
         In [7]: ed _8
         Editing... done. Executing edited code...
         hello again
         Out[7]: "print 'hello again'n"
 
 
         Changing the default editor hook:
 
         If you wish to write your own editor hook, you can put it in a
         configuration file which you load at startup time.  The default hook
         is defined in the IPython.core.hooks module, and you can use that as a
         starting example for further modifications.  That file also has
         general instructions on how to set a new hook for use once you've
         defined it."""
         
         # FIXME: This function has become a convoluted mess.  It needs a
         # ground-up rewrite with clean, simple logic.
 
         def make_filename(arg):
             "Make a filename from the given args"
             try:
                 filename = get_py_filename(arg)
             except IOError:
                 if args.endswith('.py'):
                     filename = arg
                 else:
                     filename = None
             return filename
 
         # custom exceptions
         class DataIsObject(Exception): pass
 
         opts,args = self.parse_options(parameter_s,'prn:')
         # Set a few locals from the options for convenience:
         opts_p = opts.has_key('p')
         opts_r = opts.has_key('r')
         
         # Default line number value
         lineno = opts.get('n',None)
         if lineno is not None:
             try:
                 lineno = int(lineno)
             except:
                 warn("The -n argument must be an integer.")
                 return
 
         if opts_p:
             args = '_%s' % last_call[0]
             if not self.shell.user_ns.has_key(args):
                 args = last_call[1]
             
         # use last_call to remember the state of the previous call, but don't
         # let it be clobbered by successive '-p' calls.
         try:
             last_call[0] = self.shell.displayhook.prompt_count
             if not opts_p:
                 last_call[1] = parameter_s
         except:
             pass
 
         # by default this is done with temp files, except when the given
         # arg is a filename
         use_temp = 1
 
         if re.match(r'\d',args):
             # Mode where user specifies ranges of lines, like in %macro.
             # This means that you can't edit files whose names begin with
             # numbers this way. Tough.
             ranges = args.split()
             data = ''.join(self.extract_input_slices(ranges,opts_r))
         elif args.endswith('.py'):
             filename = make_filename(args)
             data = ''
             use_temp = 0
         elif args:
             try:
                 # Load the parameter given as a variable. If not a string,
                 # process it as an object instead (below)
 
                 #print '*** args',args,'type',type(args)  # dbg
                 data = eval(args,self.shell.user_ns)
                 if not type(data) in StringTypes:
                     raise DataIsObject
 
             except (NameError,SyntaxError):
                 # given argument is not a variable, try as a filename
                 filename = make_filename(args)
                 if filename is None:
                     warn("Argument given (%s) can't be found as a variable "
                          "or as a filename." % args)
                     return
 
                 data = ''
                 use_temp = 0
             except DataIsObject:
 
                 # macros have a special edit function
                 if isinstance(data,Macro):
                     self._edit_macro(args,data)
                     return
                                 
                 # For objects, try to edit the file where they are defined
                 try:
                     filename = inspect.getabsfile(data)
                     if 'fakemodule' in filename.lower() and inspect.isclass(data):                     
                         # class created by %edit? Try to find source
                         # by looking for method definitions instead, the
                         # __module__ in those classes is FakeModule.
                         attrs = [getattr(data, aname) for aname in dir(data)]
                         for attr in attrs:
                             if not inspect.ismethod(attr):
                                 continue
                             filename = inspect.getabsfile(attr)
                             if filename and 'fakemodule' not in filename.lower():
                                 # change the attribute to be the edit target instead
                                 data = attr 
                                 break
                     
                     datafile = 1
                 except TypeError:
                     filename = make_filename(args)
                     datafile = 1
                     warn('Could not find file where `%s` is defined.\n'
                          'Opening a file named `%s`' % (args,filename))
                 # Now, make sure we can actually read the source (if it was in
                 # a temp file it's gone by now).
                 if datafile:
                     try:
                         if lineno is None:
                             lineno = inspect.getsourcelines(data)[1]
                     except IOError:
                         filename = make_filename(args)
                         if filename is None:
                             warn('The file `%s` where `%s` was defined cannot '
                                  'be read.' % (filename,data))
                             return
                 use_temp = 0
         else:
             data = ''
 
         if use_temp:
             filename = self.shell.mktempfile(data)
             print('IPython will make a temporary file named:', filename)
 
         # Make sure we send to the client an absolute path, in case the working
         # directory of client and kernel don't match
         filename = os.path.abspath(filename)
 
         payload = {
             'source' : 'IPython.zmq.zmqshell.ZMQInteractiveShell.edit_magic',
             'filename' : filename,
             'line_number' : lineno
         }
         self.payload_manager.write_payload(payload)
 
     def magic_gui(self, *args, **kwargs):
         raise NotImplementedError(
             'GUI support must be enabled in command line options.')
 
     def magic_pylab(self, *args, **kwargs):
         raise NotImplementedError(
             'pylab support must be enabled in command line options.')
 
     # A few magics that are adapted to the specifics of using pexpect and a
     # remote terminal
 
     def magic_clear(self, arg_s):
         """Clear the terminal."""
         if os.name == 'posix':
             self.shell.system("clear")
         else:
             self.shell.system("cls")
 
     if os.name == 'nt':
         # This is the usual name in windows
         magic_cls = magic_clear
 
     # Terminal pagers won't work over pexpect, but we do have our own pager
     
     def magic_less(self, arg_s):
         """Show a file through the pager.
 
         Files ending in .py are syntax-highlighted."""
         cont = open(arg_s).read()
         if arg_s.endswith('.py'):
             cont = self.shell.pycolorize(cont)
         page.page(cont)
 
     magic_more = magic_less
 
     # Man calls a pager, so we also need to redefine it
     if os.name == 'posix':
         def magic_man(self, arg_s):
             """Find the man page for the given command and display in pager."""
             page.page(self.shell.getoutput('man %s | col -b' % arg_s,
                                            split=False))
 
     # FIXME: this is specific to the GUI, so we should let the gui app load
     # magics at startup that are only for the gui.  Once the gui app has proper
     # profile and configuration management, we can have it initialize a kernel
     # with a special config file that provides these.
     def magic_guiref(self, arg_s):
         """Show a basic reference about the GUI console."""
         from IPython.core.usage import gui_reference
         page.page(gui_reference, auto_html=True)
 
     def magic_loadpy(self, arg_s):
         """Load a .py python script into the GUI console.
 
         This magic command can either take a local filename or a url::
 
         %loadpy myscript.py
         %loadpy http://www.example.com/myscript.py
         """
         if not arg_s.endswith('.py'):
             raise ValueError('%%load only works with .py files: %s' % arg_s)
         if arg_s.startswith('http'):
             import urllib2
             response = urllib2.urlopen(arg_s)
             content = response.read()
         else:
             content = open(arg_s).read()
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.magic_loadpy', 
             text=content
         )
         self.payload_manager.write_payload(payload)
         
     def magic_Exit(self, parameter_s=''):
         """Exit IPython. If the -k option is provided, the kernel will be left
         running. Otherwise, it will shutdown without prompting.
         """
         opts,args = self.parse_options(parameter_s,'k')
         self.shell.keepkernel_on_exit = opts.has_key('k')
         self.shell.ask_exit()
 
     # Add aliases as magics so all common forms work: exit, quit, Exit, Quit.
     magic_exit = magic_quit = magic_Quit = magic_Exit
 
 InteractiveShellABC.register(ZMQInteractiveShell)

docs/source/development/messaging.txt

@@ -1,873 +1,892 b''
 .. _messaging:
 
 ======================
  Messaging in IPython
 ======================
 
 
 Introduction
 ============
 
 This document explains the basic communications design and messaging
 specification for how the various IPython objects interact over a network
 transport.  The current implementation uses the ZeroMQ_ library for messaging
 within and between hosts.
 
 .. Note::
 
    This document should be considered the authoritative description of the
    IPython messaging protocol, and all developers are strongly encouraged to
    keep it updated as the implementation evolves, so that we have a single
    common reference for all protocol details.
    
 The basic design is explained in the following diagram:
 
 .. image:: frontend-kernel.png
    :width: 450px
    :alt: IPython kernel/frontend messaging architecture.
    :align: center
    :target: ../_images/frontend-kernel.png
 
 A single kernel can be simultaneously connected to one or more frontends.  The
 kernel has three sockets that serve the following functions:
 
 1. REQ: this socket is connected to a *single* frontend at a time, and it allows
    the kernel to request input from a frontend when :func:`raw_input` is called.
    The frontend holding the matching REP socket acts as a 'virtual keyboard'
    for the kernel while this communication is happening (illustrated in the
    figure by the black outline around the central keyboard).  In practice,
    frontends may display such kernel requests using a special input widget or
    otherwise indicating that the user is to type input for the kernel instead
    of normal commands in the frontend.
 
 2. XREP: this single sockets allows multiple incoming connections from
    frontends, and this is the socket where requests for code execution, object
    information, prompts, etc. are made to the kernel by any frontend.  The
    communication on this socket is a sequence of request/reply actions from
    each frontend and the kernel.
 
 3. PUB: this socket is the 'broadcast channel' where the kernel publishes all
    side effects (stdout, stderr, etc.) as well as the requests coming from any
    client over the XREP socket and its own requests on the REP socket.  There
    are a number of actions in Python which generate side effects: :func:`print`
    writes to ``sys.stdout``, errors generate tracebacks, etc.  Additionally, in
    a multi-client scenario, we want all frontends to be able to know what each
    other has sent to the kernel (this can be useful in collaborative scenarios,
    for example).  This socket allows both side effects and the information
    about communications taking place with one client over the XREQ/XREP channel
    to be made available to all clients in a uniform manner.
 
    All messages are tagged with enough information (details below) for clients
    to know which messages come from their own interaction with the kernel and
    which ones are from other clients, so they can display each type
    appropriately.
 
 The actual format of the messages allowed on each of these channels is
 specified below.  Messages are dicts of dicts with string keys and values that
 are reasonably representable in JSON.  Our current implementation uses JSON
 explicitly as its message format, but this shouldn't be considered a permanent
 feature.  As we've discovered that JSON has non-trivial performance issues due
 to excessive copying, we may in the future move to a pure pickle-based raw
 message format.  However, it should be possible to easily convert from the raw
 objects to JSON, since we may have non-python clients (e.g. a web frontend).
 As long as it's easy to make a JSON version of the objects that is a faithful
 representation of all the data, we can communicate with such clients.
 
 .. Note::
 
    Not all of these have yet been fully fleshed out, but the key ones are, see
    kernel and frontend files for actual implementation details.
 
    
 Python functional API
 =====================
 
 As messages are dicts, they map naturally to a ``func(**kw)`` call form.  We
 should develop, at a few key points, functional forms of all the requests that
 take arguments in this manner and automatically construct the necessary dict
 for sending.
 
 
 General Message Format
 ======================
 
 All messages send or received by any IPython process should have the following
 generic structure::
     
     {
       # The message header contains a pair of unique identifiers for the
       # originating session and the actual message id, in addition to the
       # username for the process that generated the message.  This is useful in
       # collaborative settings where multiple users may be interacting with the
       # same kernel simultaneously, so that frontends can label the various
       # messages in a meaningful way.
       'header' : { 'msg_id' : uuid,
                    'username' : str,
            'session' : uuid
          },
 
       # In a chain of messages, the header from the parent is copied so that
       # clients can track where messages come from.
       'parent_header' : dict,
 
       # All recognized message type strings are listed below.
       'msg_type' : str,
 
       # The actual content of the message must be a dict, whose structure
       # depends on the message type.x
       'content' : dict,
     }
 
 For each message type, the actual content will differ and all existing message
 types are specified in what follows of this document.
 
 
 Messages on the XREP/XREQ socket
 ================================
 
 .. _execute:
 
 Execute
 -------
 
 This message type is used by frontends to ask the kernel to execute code on
 behalf of the user, in a namespace reserved to the user's variables (and thus
 separate from the kernel's own internal code and variables).
 
 Message type: ``execute_request``::
 
     content = {
         # Source code to be executed by the kernel, one or more lines.
     'code' : str,
 
     # A boolean flag which, if True, signals the kernel to execute this
     # code as quietly as possible.  This means that the kernel will compile
     # the code witIPython/core/tests/h 'exec' instead of 'single' (so
     # sys.displayhook will not fire), and will *not*:
     #   - broadcast exceptions on the PUB socket
     #   - do any logging
     #   - populate any history
     #
     # The default is False.
     'silent' : bool,
 
     # A list of variable names from the user's namespace to be retrieved.  What
     # returns is a JSON string of the variable's repr(), not a python object.
     'user_variables' : list,
 
     # Similarly, a dict mapping names to expressions to be evaluated in the
     # user's dict.
     'user_expressions' : dict,
     }
 
 The ``code`` field contains a single string (possibly multiline).  The kernel
 is responsible for splitting this into one or more independent execution blocks
 and deciding whether to compile these in 'single' or 'exec' mode (see below for
 detailed execution semantics).
 
 The ``user_`` fields deserve a detailed explanation.  In the past, IPython had
 the notion of a prompt string that allowed arbitrary code to be evaluated, and
 this was put to good use by many in creating prompts that displayed system
 status, path information, and even more esoteric uses like remote instrument
 status aqcuired over the network.  But now that IPython has a clean separation
 between the kernel and the clients, the kernel has no prompt knowledge; prompts
 are a frontend-side feature, and it should be even possible for different
 frontends to display different prompts while interacting with the same kernel.
 
 The kernel now provides the ability to retrieve data from the user's namespace
 after the execution of the main ``code``, thanks to two fields in the
 ``execute_request`` message:
 
 - ``user_variables``: If only variables from the user's namespace are needed, a
   list of variable names can be passed and a dict with these names as keys and
   their :func:`repr()` as values will be returned.
 
 - ``user_expressions``: For more complex expressions that require function
   evaluations, a dict can be provided with string keys and arbitrary python
   expressions as values.  The return message will contain also a dict with the
   same keys and the :func:`repr()` of the evaluated expressions as value.
 
 With this information, frontends can display any status information they wish
 in the form that best suits each frontend (a status line, a popup, inline for a
 terminal, etc).
 
 .. Note::
 
    In order to obtain the current execution counter for the purposes of
    displaying input prompts, frontends simply make an execution request with an
    empty code string and ``silent=True``.
 
 Execution semantics
 ~~~~~~~~~~~~~~~~~~~
 
 When the silent flag is false, the execution of use code consists of the
 following phases (in silent mode, only the ``code`` field is executed):
 
 1. Run the ``pre_runcode_hook``.
 
 2. Execute the ``code`` field, see below for details.
 
 3. If #2 succeeds, compute ``user_variables`` and ``user_expressions`` are
    computed.  This ensures that any error in the latter don't harm the main
    code execution.
 
 4. Call any method registered with :meth:`register_post_execute`.
 
 .. warning::
 
    The API for running code before/after the main code block is likely to
    change soon.  Both the ``pre_runcode_hook`` and the
    :meth:`register_post_execute` are susceptible to modification, as we find a
    consistent model for both.
 
 To understand how the ``code`` field is executed, one must know that Python
 code can be compiled in one of three modes (controlled by the ``mode`` argument
 to the :func:`compile` builtin):
 
 *single*
   Valid for a single interactive statement (though the source can contain
   multiple lines, such as a for loop).  When compiled in this mode, the
   generated bytecode contains special instructions that trigger the calling of
   :func:`sys.displayhook` for any expression in the block that returns a value.
   This means that a single statement can actually produce multiple calls to
   :func:`sys.displayhook`, if for example it contains a loop where each
   iteration computes an unassigned expression would generate 10 calls::
 
       for i in range(10):
           i**2
 
 *exec*
   An arbitrary amount of source code, this is how modules are compiled.
   :func:`sys.displayhook` is *never* implicitly called.
 
 *eval*
   A single expression that returns a value.  :func:`sys.displayhook` is *never*
   implicitly called.
 
 
 The ``code`` field is split into individual blocks each of which is valid for
 execution in 'single' mode, and then:
 
 - If there is only a single block: it is executed in 'single' mode.
 
 - If there is more than one block:
 
   * if the last one is a single line long, run all but the last in 'exec' mode
     and the very last one in 'single' mode.  This makes it easy to type simple
     expressions at the end to see computed values.
 
   * if the last one is no more than two lines long, run all but the last in
     'exec' mode and the very last one in 'single' mode.  This makes it easy to
     type simple expressions at the end to see computed values.  - otherwise
     (last one is also multiline), run all in 'exec' mode
 
   * otherwise (last one is also multiline), run all in 'exec' mode as a single
     unit.
 
 Any error in retrieving the ``user_variables`` or evaluating the
 ``user_expressions`` will result in a simple error message in the return fields
 of the form::
 
    [ERROR] ExceptionType: Exception message
 
 The user can simply send the same variable name or expression for evaluation to
 see a regular traceback.
 
 Errors in any registered post_execute functions are also reported similarly,
 and the failing function is removed from the post_execution set so that it does
 not continue triggering failures.
 
 Upon completion of the execution request, the kernel *always* sends a reply,
 with a status code indicating what happened and additional data depending on
 the outcome.  See :ref:`below <execution_results>` for the possible return
 codes and associated data.
 
 
 Execution counter (old prompt number)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The kernel has a single, monotonically increasing counter of all execution
 requests that are made with ``silent=False``.  This counter is used to populate
 the ``In[n]``, ``Out[n]`` and ``_n`` variables, so clients will likely want to
 display it in some form to the user, which will typically (but not necessarily)
 be done in the prompts.  The value of this counter will be returned as the
 ``execution_count`` field of all ``execute_reply`` messages.
 
 .. _execution_results:
 
 Execution results
 ~~~~~~~~~~~~~~~~~
     
 Message type: ``execute_reply``::
 
     content = {
       # One of: 'ok' OR 'error' OR 'abort'
       'status' : str,
 
       # The global kernel counter that increases by one with each non-silent
       # executed request.  This will typically be used by clients to display
       # prompt numbers to the user.  If the request was a silent one, this will
       # be the current value of the counter in the kernel.
       'execution_count' : int,
     }
 
 When status is 'ok', the following extra fields are present::
 
     {
       # The execution payload is a dict with string keys that may have been
       # produced by the code being executed.  It is retrieved by the kernel at
       # the end of the execution and sent back to the front end, which can take
       # action on it as needed.  See main text for further details.
       'payload' : dict,
 
       # Results for the user_variables and user_expressions.
       'user_variables' : dict,
       'user_expressions' : dict,
 
       # The kernel will often transform the input provided to it.  If the
       # '---->' transform had been applied, this is filled, otherwise it's the
       # empty string.  So transformations like magics don't appear here, only
       # autocall ones.
       'transformed_code' : str,
       }
 
 .. admonition:: Execution payloads
     
    The notion of an 'execution payload' is different from a return value of a
    given set of code, which normally is just displayed on the pyout stream
    through the PUB socket.  The idea of a payload is to allow special types of
    code, typically magics, to populate a data container in the IPython kernel
    that will be shipped back to the caller via this channel.  The kernel will
    have an API for this, probably something along the lines of::
 
        ip.exec_payload_add(key, value)
 
    though this API is still in the design stages.  The data returned in this
    payload will allow frontends to present special views of what just happened.
 
    
 When status is 'error', the following extra fields are present::
 
     {
       'exc_name' : str,   # Exception name, as a string
       'exc_value' : str,  # Exception value, as a string
 
       # The traceback will contain a list of frames, represented each as a
       # string.  For now we'll stick to the existing design of ultraTB, which
       # controls exception level of detail statefully.  But eventually we'll
       # want to grow into a model where more information is collected and
       # packed into the traceback object, with clients deciding how little or
       # how much of it to unpack.  But for now, let's start with a simple list
       # of strings, since that requires only minimal changes to ultratb as
       # written.
       'traceback' : list,
     }
 
 
 When status is 'abort', there are for now no additional data fields.  This
 happens when the kernel was interrupted by a signal.
 
 Kernel attribute access
 -----------------------
 
 .. warning::
 
    This part of the messaging spec is not actually implemented in the kernel
    yet.
  
 While this protocol does not specify full RPC access to arbitrary methods of
 the kernel object, the kernel does allow read (and in some cases write) access
 to certain attributes.
 
 The policy for which attributes can be read is: any attribute of the kernel, or
 its sub-objects, that belongs to a :class:`Configurable` object and has been
 declared at the class-level with Traits validation, is in principle accessible
 as long as its name does not begin with a leading underscore.  The attribute
 itself will have metadata indicating whether it allows remote read and/or write
 access.  The message spec follows for attribute read and write requests.
 
 Message type: ``getattr_request``::
 
     content = {
         # The (possibly dotted) name of the attribute
 	'name' : str,
     }
 
 When a ``getattr_request`` fails, there are two possible error types:
 
 - AttributeError: this type of error was raised when trying to access the
   given name by the kernel itself.  This means that the attribute likely
   doesn't exist.
 
 - AccessError: the attribute exists but its value is not readable remotely.
 
 
 Message type: ``getattr_reply``::
 
     content = {
         # One of ['ok', 'AttributeError', 'AccessError'].
         'status' : str,
 	# If status is 'ok', a JSON object.
 	'value' : object,
     }
 
 Message type: ``setattr_request``::
 
     content = {
         # The (possibly dotted) name of the attribute
 	'name' : str,
 	
 	# A JSON-encoded object, that will be validated by the Traits
 	# information in the kernel
 	'value' : object,
     }
 
 When a ``setattr_request`` fails, there are also two possible error types with
 similar meanings  as those of the ``getattr_request`` case, but for writing.
     
 Message type: ``setattr_reply``::
 
     content = {
         # One of ['ok', 'AttributeError', 'AccessError'].
         'status' : str,
     }
 
 
     
 Object information
 ------------------
 
 One of IPython's most used capabilities is the introspection of Python objects
 in the user's namespace, typically invoked via the ``?`` and ``??`` characters
 (which in reality are shorthands for the ``%pinfo`` magic).  This is used often
 enough that it warrants an explicit message type, especially because frontends
 may want to get object information in response to user keystrokes (like Tab or
 F1) besides from the user explicitly typing code like ``x??``.
 
 Message type: ``object_info_request``::
 
     content = {
         # The (possibly dotted) name of the object to be searched in all
 	# relevant namespaces
         'name' : str,
 
     	# The level of detail desired.  The default (0) is equivalent to typing
 	# 'x?' at the prompt, 1 is equivalent to 'x??'.
 	'detail_level' : int,
     }
 
 The returned information will be a dictionary with keys very similar to the
 field names that IPython prints at the terminal.
     
 Message type: ``object_info_reply``::
 
     content = {
     # The name the object was requested under
     'name' : str,
     
     # Boolean flag indicating whether the named object was found or not.  If
     # it's false, all other fields will be empty.
     'found' : bool,
     
     # Flags for magics and system aliases
     'ismagic' : bool,
     'isalias' : bool,
 
     # The name of the namespace where the object was found ('builtin',
     # 'magics', 'alias', 'interactive', etc.)
     'namespace' : str,
 
     # The type name will be type.__name__ for normal Python objects, but it
     # can also be a string like 'Magic function' or 'System alias'
     'type_name' : str,
 
     'string_form' : str,
 
     # For objects with a __class__ attribute this will be set
     'base_class' : str,
 
     # For objects with a __len__ attribute this will be set
     'length' : int,
 
     # If the object is a function, class or method whose file we can find,
     # we give its full path
     'file' : str,
 
     # For pure Python callable objects, we can reconstruct the object
     # definition line which provides its call signature.  For convenience this
     # is returned as a single 'definition' field, but below the raw parts that
     # compose it are also returned as the argspec field.
     'definition' : str,
 
     # The individual parts that together form the definition string.  Clients
     # with rich display capabilities may use this to provide a richer and more
     # precise representation of the definition line (e.g. by highlighting
     # arguments based on the user's cursor position).  For non-callable
     # objects, this field is empty.
     'argspec' : { # The names of all the arguments
                   args : list,
 		  # The name of the varargs (*args), if any
                   varargs : str,
 		  # The name of the varkw (**kw), if any
 		  varkw : str,
 		  # The values (as strings) of all default arguments.  Note
 		  # that these must be matched *in reverse* with the 'args'
 		  # list above, since the first positional args have no default
 		  # value at all.
 		  defaults : list,
 		  },
 
     # For instances, provide the constructor signature (the definition of
     # the __init__ method):
     'init_definition' : str,
     
     # Docstrings: for any object (function, method, module, package) with a
     # docstring, we show it.  But in addition, we may provide additional
     # docstrings.  For example, for instances we will show the constructor
     # and class docstrings as well, if available.
     'docstring' : str,
 
     # For instances, provide the constructor and class docstrings
     'init_docstring' : str,
     'class_docstring' : str,
     
     # If it's a callable object whose call method has a separate docstring and
     # definition line:
     'call_def' : str,
     'call_docstring' : str,
     
     # If detail_level was 1, we also try to find the source code that
     # defines the object, if possible.  The string 'None' will indicate
     # that no source was found.
     'source' : str,
     }
 '
     
 Complete
 --------
 
 Message type: ``complete_request``::
 
     content = {
         # The text to be completed, such as 'a.is'
     'text' : str,
 
     # The full line, such as 'print a.is'.  This allows completers to
     # make decisions that may require information about more than just the
     # current word.
     'line' : str,
 
     # The entire block of text where the line is.  This may be useful in the
     # case of multiline completions where more context may be needed.  Note: if
     # in practice this field proves unnecessary, remove it to lighten the
     # messages.
     
     'block' : str,
 
     # The position of the cursor where the user hit 'TAB' on the line.
     'cursor_pos' : int,
     }
 
 Message type: ``complete_reply``::
 
     content = {
         # The list of all matches to the completion request, such as
     # ['a.isalnum', 'a.isalpha'] for the above example.
     'matches' : list
     }
 
     
 History
 -------
 
 For clients to explicitly request history from a kernel.  The kernel has all
 the actual execution history stored in a single location, so clients can
 request it from the kernel when needed.
 
 Message type: ``history_request``::
 
     content = {
     
       # If True, also return output history in the resulting dict.
       'output' : bool,
 
       # If True, return the raw input history, else the transformed input.
       'raw' : bool,
 
       # This parameter can be one of: A number,  a pair of numbers, None
       # If not given, last 40 are returned.
       #  - number n: return the last n entries.
       #  - pair n1, n2: return entries in the range(n1, n2).
       #  - None: return all history
       'index' : n or (n1, n2) or None,
     }
 
 Message type: ``history_reply``::
 
     content = {
       # A dict with prompt numbers as keys and either (input, output) or input
       # as the value depending on whether output was True or False,
       # respectively.
       'history' : dict,
     }
 
 
 Connect
 -------
 
 When a client connects to the request/reply socket of the kernel, it can issue
 a connect request to get basic information about the kernel, such as the ports
 the other ZeroMQ sockets are listening on. This allows clients to only have
 to know about a single port (the XREQ/XREP channel) to connect to a kernel.
 
 Message type: ``connect_request``::
 
     content = {
     }
 
 Message type: ``connect_reply``::
 
     content = {
         'xrep_port' : int  # The port the XREP socket is listening on.
         'pub_port' : int   # The port the PUB socket is listening on.
         'req_port' : int   # The port the REQ socket is listening on.
         'hb_port' : int    # The port the heartbeat socket is listening on.
     }
 
 
 
 Kernel shutdown
 ---------------
 
 The clients can request the kernel to shut itself down; this is used in
 multiple cases:
 
 - when the user chooses to close the client application via a menu or window
   control.
 - when the user types 'exit' or 'quit' (or their uppercase magic equivalents).
 - when the user chooses a GUI method (like the 'Ctrl-C' shortcut in the
   IPythonQt client) to force a kernel restart to get a clean kernel without
   losing client-side state like history or inlined figures.
 
 The client sends a shutdown request to the kernel, and once it receives the
 reply message (which is otherwise empty), it can assume that the kernel has
 completed shutdown safely.
 
 Upon their own shutdown, client applications will typically execute a last
 minute sanity check and forcefully terminate any kernel that is still alive, to
 avoid leaving stray processes in the user's machine.
 
 For both shutdown request and reply, there is no actual content that needs to
 be sent, so the content dict is empty.
 
 Message type: ``shutdown_request``::
 
     content = {
         'restart' : bool # whether the shutdown is final, or precedes a restart
     }
 
 Message type: ``shutdown_reply``::
 
     content = {
         'restart' : bool # whether the shutdown is final, or precedes a restart
     }
 
 .. Note::
 
    When the clients detect a dead kernel thanks to inactivity on the heartbeat
    socket, they simply send a forceful process termination signal, since a dead
    process is unlikely to respond in any useful way to messages.
     
 
 Messages on the PUB/SUB socket
 ==============================
 
 Streams (stdout,  stderr, etc)
 ------------------------------
 
 Message type: ``stream``::
 
     content = {
         # The name of the stream is one of 'stdin', 'stdout', 'stderr'
     'name' : str,
     
     # The data is an arbitrary string to be written to that stream
     'data' : str,
     }
 
 When a kernel receives a raw_input call, it should also broadcast it on the pub
 socket with the names 'stdin' and 'stdin_reply'.  This will allow other clients
 to monitor/display kernel interactions and possibly replay them to their user
 or otherwise expose them.
 
 Python inputs
 -------------
 
 These messages are the re-broadcast of the ``execute_request``.
 
 Message type: ``pyin``::
 
     content = {
         # Source code to be executed, one or more lines
     'code' : str
     }
 
 Python outputs
 --------------
 
 When Python produces output from code that has been compiled in with the
 'single' flag to :func:`compile`, any expression that produces a value (such as
 ``1+1``) is passed to ``sys.displayhook``, which is a callable that can do with
 this value whatever it wants.  The default behavior of ``sys.displayhook`` in
 the Python interactive prompt is to print to ``sys.stdout`` the :func:`repr` of
 the value as long as it is not ``None`` (which isn't printed at all).  In our
 case, the kernel instantiates as ``sys.displayhook`` an object which has
 similar behavior, but which instead of printing to stdout, broadcasts these
 values as ``pyout`` messages for clients to display appropriately.
 
+IPython's displayhook can handle multiple simultaneous formats depending on its
+configuration. The default pretty-printed repr text is always given with the
+``data`` entry in this message. Any other formats are provided in the
+``extra_formats`` list. Frontends are free to display any or all of these
+according to its capabilities. ``extra_formats`` list contains 3-tuples of an ID
+string, a type string, and the data. The ID is unique to the formatter
+implementation that created the data. Frontends will typically ignore the ID
+unless if it has requested a particular formatter. The type string tells the
+frontend how to interpret the data. It is often, but not always a MIME type.
+Frontends should ignore types that it does not understand. The data itself is
+any JSON object and depends on the format. It is often, but not always a string.
+
 Message type: ``pyout``::
 
     content = {
-        # The data is typically the repr() of the object.
-    'data' : str,
+        # The data is typically the repr() of the object. It should be displayed
+        # as monospaced text.
+        'data' : str,
     
-    # The counter for this execution is also provided so that clients can
-    # display it, since IPython automatically creates variables called _N (for
-    # prompt N).
-    'execution_count' : int,
+        # The counter for this execution is also provided so that clients can
+        # display it, since IPython automatically creates variables called _N
+        # (for prompt N).
+        'execution_count' : int,
+
+        # Any extra formats.
+        # The tuples are of the form (ID, type, data).
+        'extra_formats' : [
+           [str, str, object]
+        ]
     }
     
 Python errors
 -------------
 
 When an error occurs during code execution
 
 Message type: ``pyerr``::
 
     content = {
        # Similar content to the execute_reply messages for the 'error' case,
        # except the 'status' field is omitted.
     }
 
 Kernel status
 -------------
 
 This message type is used by frontends to monitor the status of the kernel.
 
 Message type: ``status``::
 
     content = {
         # When the kernel starts to execute code, it will enter the 'busy'
         # state and when it finishes, it will enter the 'idle' state.
         execution_state : ('busy', 'idle')
     }
 
 Kernel crashes
 --------------
 
 When the kernel has an unexpected exception, caught by the last-resort
 sys.excepthook, we should broadcast the crash handler's output before exiting.
 This will allow clients to notice that a kernel died, inform the user and
 propose further actions.
 
 Message type: ``crash``::
 
     content = {
        # Similarly to the 'error' case for execute_reply messages, this will
        # contain exc_name, exc_type and traceback fields.
 
        # An additional field with supplementary information such as where to
        # send the crash message
        'info' : str,
     }
 
 
 Future ideas
 ------------
     
 Other potential message types, currently unimplemented, listed below as ideas.
     
 Message type: ``file``::
 
     content = {
     'path' : 'cool.jpg',
     'mimetype' : str,
     'data' : str,
     }
 
     
 Messages on the REQ/REP socket
 ==============================
 
 This is a socket that goes in the opposite direction: from the kernel to a
 *single* frontend, and its purpose is to allow ``raw_input`` and similar
 operations that read from ``sys.stdin`` on the kernel to be fulfilled by the
 client.  For now we will keep these messages as simple as possible, since they
 basically only mean to convey the ``raw_input(prompt)`` call.
 
 Message type: ``input_request``::
 
     content = { 'prompt' : str }
 
 Message type: ``input_reply``::
 
     content = { 'value' : str }
     
 .. Note::
 
    We do not explicitly try to forward the raw ``sys.stdin`` object, because in
    practice the kernel should behave like an interactive program.  When a
    program is opened on the console, the keyboard effectively takes over the
    ``stdin`` file descriptor, and it can't be used for raw reading anymore.
    Since the IPython kernel effectively behaves like a console program (albeit
    one whose "keyboard" is actually living in a separate process and
    transported over the zmq connection), raw ``stdin`` isn't expected to be
    available.
 
    
 Heartbeat for kernels
 =====================
 
 Initially we had considered using messages like those above over ZMQ for a
 kernel 'heartbeat' (a way to detect quickly and reliably whether a kernel is
 alive at all, even if it may be busy executing user code).  But this has the
 problem that if the kernel is locked inside extension code, it wouldn't execute
 the python heartbeat code.  But it turns out that we can implement a basic
 heartbeat with pure ZMQ, without using any Python messaging at all.
 
 The monitor sends out a single zmq message (right now, it is a str of the
 monitor's lifetime in seconds), and gets the same message right back, prefixed
 with the zmq identity of the XREQ socket in the heartbeat process. This can be
 a uuid, or even a full message, but there doesn't seem to be a need for packing
 up a message when the sender and receiver are the exact same Python object.
 
 The model is this::
 
     monitor.send(str(self.lifetime)) # '1.2345678910'
 
 and the monitor receives some number of messages of the form::
 
     ['uuid-abcd-dead-beef', '1.2345678910']
 
 where the first part is the zmq.IDENTITY of the heart's XREQ on the engine, and
 the rest is the message sent by the monitor.  No Python code ever has any
 access to the message between the monitor's send, and the monitor's recv.
 
 
 ToDo
 ====
 
 Missing things include:
 
 * Important: finish thinking through the payload concept and API.
 
 * Important: ensure that we have a good solution for magics like %edit.  It's
   likely that with the payload concept we can build a full solution, but not
   100% clear yet.
 
 * Finishing the details of the heartbeat protocol.
 
 * Signal handling: specify what kind of information kernel should broadcast (or
   not) when it receives signals.
 
 .. include:: ../links.rst