update docs to reflect relaxed syntax of argparse

docs/source/config/overview.txt

@@ -1,479 +1,499 b''
 .. _config_overview:
 
 ============================================
 Overview of the IPython configuration system
 ============================================
 
 This section describes the IPython configuration system. Starting with version
 0.11, IPython has a completely new configuration system that is quite
 different from the older :file:`ipythonrc` or :file:`ipy_user_conf.py`
 approaches. The new configuration system was designed from scratch to address
 the particular configuration needs of IPython. While there are many
 other excellent configuration systems out there, we found that none of them
 met our requirements.
 
 .. warning::
 
     If you are upgrading to version 0.11 of IPython, you will need to migrate
     your old :file:`ipythonrc` or :file:`ipy_user_conf.py` configuration files
     to the new system.  Read on for information on how to do this.
 
 The discussion that follows is focused on teaching users how to configure
 IPython to their liking.  Developers who want to know more about how they
 can enable their objects to take advantage of the configuration system
 should consult our :ref:`developer guide <developer_guide>`
 
 The main concepts
 =================
 
 There are a number of abstractions that the IPython configuration system uses.
 Each of these abstractions is represented by a Python class.
 
 Configuration object: :class:`~IPython.config.loader.Config`
     A configuration object is a simple dictionary-like class that holds
     configuration attributes and sub-configuration objects. These classes
     support dotted attribute style access (``Foo.bar``) in addition to the
     regular dictionary style access (``Foo['bar']``). Configuration objects
     are smart. They know how to merge themselves with other configuration
     objects and they automatically create sub-configuration objects.
 
 Application: :class:`~IPython.config.application.Application`
     An application is a process that does a specific job. The most obvious
     application is the :command:`ipython` command line program. Each
     application reads *one or more* configuration files and a single set of
     command line options
     and then produces a master configuration object for the application. This
     configuration object is then passed to the configurable objects that the
     application creates. These configurable objects implement the actual logic
     of the application and know how to configure themselves given the
     configuration object.
     
     Applications always have a `log` attribute that is a configured Logger.
     This allows centralized logging configuration per-application.
 
 Configurable: :class:`~IPython.config.configurable.Configurable`
     A configurable is a regular Python class that serves as a base class for
     all main classes in an application. The
     :class:`~IPython.config.configurable.Configurable` base class is
     lightweight and only does one things.
 
     This :class:`~IPython.config.configurable.Configurable` is a subclass
     of :class:`~IPython.utils.traitlets.HasTraits` that knows how to configure
     itself. Class level traits with the metadata ``config=True`` become
     values that can be configured from the command line and configuration
     files.
     
     Developers create :class:`~IPython.config.configurable.Configurable`
     subclasses that implement all of the logic in the application. Each of
     these subclasses has its own configuration information that controls how
     instances are created.
 
 Singletons: :class:`~IPython.config.configurable.SingletonConfigurable`
     Any object for which there is a single canonical instance. These are
     just like Configurables, except they have a class method 
     :meth:`~IPython.config.configurable.SingletonConfigurable.instance`,
     that returns the current active instance (or creates one if it
     does not exist).  Examples of singletons include
     :class:`~IPython.config.application.Application`s and
     :class:`~IPython.core.interactiveshell.InteractiveShell`.  This lets
     objects easily connect to the current running Application without passing
     objects around everywhere.  For instance, to get the current running 
     Application instance, simply do: ``app = Application.instance()``.
 
 
 .. note::
 
     Singletons are not strictly enforced - you can have many instances
     of a given singleton class, but the :meth:`instance` method will always
     return the same one.
 
 Having described these main concepts, we can now state the main idea in our
 configuration system: *"configuration" allows the default values of class
 attributes to be controlled on a class by class basis*. Thus all instances of
 a given class are configured in the same way. Furthermore, if two instances
 need to be configured differently, they need to be instances of two different
 classes. While this model may seem a bit restrictive, we have found that it
 expresses most things that need to be configured extremely well. However, it
 is possible to create two instances of the same class that have different
 trait values. This is done by overriding the configuration.
 
 Now, we show what our configuration objects and files look like.
 
 Configuration objects and files
 ===============================
 
 A configuration file is simply a pure Python file that sets the attributes
 of a global, pre-created configuration object.  This configuration object is a 
 :class:`~IPython.config.loader.Config` instance.  While in a configuration
 file, to get a reference to this object, simply call the :func:`get_config`
 function.  We inject this function into the global namespace that the 
 configuration file is executed in.
 
 Here is an example of a super simple configuration file that does nothing::
 
     c = get_config()
 
 Once you get a reference to the configuration object, you simply set
 attributes on it.  All you have to know is:
 
 * The name of each attribute.
 * The type of each attribute.
 
 The answers to these two questions are provided by the various
 :class:`~IPython.config.configurable.Configurable` subclasses that an
 application uses. Let's look at how this would work for a simple configurable
 subclass::
 
     # Sample configurable:
     from IPython.config.configurable import Configurable
     from IPython.utils.traitlets import Int, Float, Unicode, Bool
     
     class MyClass(Configurable):
         name = Unicode(u'defaultname', config=True)
         ranking = Int(0, config=True)
         value = Float(99.0)
         # The rest of the class implementation would go here..
 
 In this example, we see that :class:`MyClass` has three attributes, two
 of whom (``name``, ``ranking``) can be configured.  All of the attributes
 are given types and default values.  If a :class:`MyClass` is instantiated,
 but not configured, these default values will be used.  But let's see how
 to configure this class in a configuration file::
 
     # Sample config file
     c = get_config()
     
     c.MyClass.name = 'coolname'
     c.MyClass.ranking = 10
 
 After this configuration file is loaded, the values set in it will override
 the class defaults anytime a :class:`MyClass` is created.  Furthermore,
 these attributes will be type checked and validated anytime they are set.
 This type checking is handled by the :mod:`IPython.utils.traitlets` module,
 which provides the :class:`Unicode`, :class:`Int` and :class:`Float` types.
 In addition to these traitlets, the :mod:`IPython.utils.traitlets` provides
 traitlets for a number of other types.
 
 .. note::
 
     Underneath the hood, the :class:`Configurable` base class is a subclass of
     :class:`IPython.utils.traitlets.HasTraits`. The
     :mod:`IPython.utils.traitlets` module is a lightweight version of
     :mod:`enthought.traits`. Our implementation is a pure Python subset
     (mostly API compatible) of :mod:`enthought.traits` that does not have any
     of the automatic GUI generation capabilities. Our plan is to achieve 100%
     API compatibility to enable the actual :mod:`enthought.traits` to
     eventually be used instead. Currently, we cannot use
     :mod:`enthought.traits` as we are committed to the core of IPython being
     pure Python.
 
 It should be very clear at this point what the naming convention is for 
 configuration attributes::
 
     c.ClassName.attribute_name = attribute_value
 
 Here, ``ClassName`` is the name of the class whose configuration attribute you
 want to set, ``attribute_name`` is the name of the attribute you want to set
 and ``attribute_value`` the the value you want it to have. The ``ClassName``
 attribute of ``c`` is not the actual class, but instead is another
 :class:`~IPython.config.loader.Config` instance.
 
 .. note::
 
     The careful reader may wonder how the ``ClassName`` (``MyClass`` in
     the above example) attribute of the configuration object ``c`` gets
     created. These attributes are created on the fly by the
     :class:`~IPython.config.loader.Config` instance, using a simple naming
     convention. Any attribute of a :class:`~IPython.config.loader.Config`
     instance whose name begins with an uppercase character is assumed to be a
     sub-configuration and a new empty :class:`~IPython.config.loader.Config`
     instance is dynamically created for that attribute. This allows deeply
     hierarchical information created easily (``c.Foo.Bar.value``) on the fly.
 
 Configuration files inheritance
 ===============================
 
 Let's say you want to have different configuration files for various purposes.
 Our configuration system makes it easy for one configuration file to inherit
 the information in another configuration file. The :func:`load_subconfig`
 command can be used in a configuration file for this purpose. Here is a simple
 example that loads all of the values from the file :file:`base_config.py`::
 
     # base_config.py
     c = get_config()
     c.MyClass.name = 'coolname'
     c.MyClass.ranking = 100
 
 into the configuration file :file:`main_config.py`::
 
     # main_config.py
     c = get_config()
     
     # Load everything from base_config.py
     load_subconfig('base_config.py')
     
     # Now override one of the values
     c.MyClass.name = 'bettername'
 
 In a situation like this the :func:`load_subconfig` makes sure that the
 search path for sub-configuration files is inherited from that of the parent.
 Thus, you can typically put the two in the same directory and everything will
 just work.
 
 You can also load configuration files by profile, for instance:
 
 .. sourcecode:: python
 
     load_subconfig('ipython_config.py', profile='default')
 
 to inherit your default configuration as a starting point.
 
 
 Class based configuration inheritance
 =====================================
 
 There is another aspect of configuration where inheritance comes into play.
 Sometimes, your classes will have an inheritance hierarchy that you want
 to be reflected in the configuration system.  Here is a simple example::
 
     from IPython.config.configurable import Configurable
     from IPython.utils.traitlets import Int, Float, Unicode, Bool
     
     class Foo(Configurable):
         name = Unicode(u'fooname', config=True)
         value = Float(100.0, config=True)
     
     class Bar(Foo):
         name = Unicode(u'barname', config=True)
         othervalue = Int(0, config=True)
 
 Now, we can create a configuration file to configure instances of :class:`Foo`
 and :class:`Bar`::
 
     # config file
     c = get_config()
     
     c.Foo.name = u'bestname'
     c.Bar.othervalue = 10
 
 This class hierarchy and configuration file accomplishes the following:
 
 * The default value for :attr:`Foo.name` and :attr:`Bar.name` will be
   'bestname'.  Because :class:`Bar` is a :class:`Foo` subclass it also
   picks up the configuration information for :class:`Foo`.
 * The default value for :attr:`Foo.value` and :attr:`Bar.value` will be
   ``100.0``, which is the value specified as the class default.
 * The default value for :attr:`Bar.othervalue` will be 10 as set in the
   configuration file.  Because :class:`Foo` is the parent of :class:`Bar`
   it doesn't know anything about the :attr:`othervalue` attribute.
 
 
 .. _ipython_dir:
 
 Configuration file location
 ===========================
 
 So where should you put your configuration files? IPython uses "profiles" for
 configuration, and by default, all profiles will be stored in the so called
 "IPython directory". The location of this directory is determined by the
 following algorithm:
 
 * If the ``ipython_dir`` command line flag is given, its value is used.
 
 * If not, the value returned by :func:`IPython.utils.path.get_ipython_dir`
   is used. This function will first look at the :envvar:`IPYTHON_DIR`
   environment variable and then default to a platform-specific default.
 
 On posix systems (Linux, Unix, etc.), IPython respects the ``$XDG_CONFIG_HOME``
 part of the `XDG Base Directory`_ specification. If ``$XDG_CONFIG_HOME`` is
 defined and exists ( ``XDG_CONFIG_HOME`` has a default interpretation of
 :file:`$HOME/.config`), then IPython's config directory will be located in
 :file:`$XDG_CONFIG_HOME/ipython`.  If users still have an IPython directory
 in :file:`$HOME/.ipython`, then that will be used. in preference to the 
 system default.
 
 For most users, the default value will simply be something like 
 :file:`$HOME/.config/ipython` on Linux, or :file:`$HOME/.ipython`
 elsewhere.
 
 Once the location of the IPython directory has been determined, you need to know
 which profile you are using. For users with a single configuration, this will
 simply be 'default', and will be located in 
 :file:`<IPYTHON_DIR>/profile_default`.
 
 The next thing you need to know is what to call your configuration file. The
 basic idea is that each application has its own default configuration filename.
 The default named used by the :command:`ipython` command line program is
 :file:`ipython_config.py`, and *all* IPython applications will use this file.
 Other applications, such as the parallel :command:`ipcluster` scripts or the
 QtConsole will load their own config files *after* :file:`ipython_config.py`. To
 load a particular configuration file instead of the default, the name can be
 overridden by the ``config_file`` command line flag.
 
 To generate the default configuration files, do::
 
     $> ipython profile create
 
 and you will have a default :file:`ipython_config.py` in your IPython directory
 under :file:`profile_default`. If you want the default config files for the
 :mod:`IPython.parallel` applications, add ``--parallel`` to the end of the
 command-line args.
 
 .. _Profiles:
 
 Profiles
 ========
 
 A profile is a directory containing configuration and runtime files, such as
 logs, connection info for the parallel apps, and your IPython command history.
 
 The idea is that users often want to maintain a set of configuration files for
 different purposes: one for doing numerical computing with NumPy and SciPy and
 another for doing symbolic computing with SymPy. Profiles make it easy to keep a
 separate configuration files, logs, and histories for each of these purposes.
 
 Let's start by showing how a profile is used:
 
 .. code-block:: bash
 
     $ ipython --profile=sympy
 
 This tells the :command:`ipython` command line program to get its configuration
 from the "sympy" profile. The file names for various profiles do not change. The
 only difference is that profiles are named in a special way. In the case above,
 the "sympy" profile means looking for :file:`ipython_config.py` in :file:`<IPYTHON_DIR>/profile_sympy`.
 
 The general pattern is this: simply create a new profile with:
 
 .. code-block:: bash
 
     ipython profile create <name>
 
 which adds a directory called ``profile_<name>`` to your IPython directory. Then
 you can load this profile by adding ``--profile=<name>`` to your command line
 options. Profiles are supported by all IPython applications.
 
 IPython ships with some sample profiles in :file:`IPython/config/profile`. If
 you create profiles with the name of one of our shipped profiles, these config
 files will be copied over instead of starting with the automatically generated
 config files.
 
 .. _commandline:
 
 Command-line arguments
 ======================
 
 IPython exposes *all* configurable options on the command-line. The command-line
 arguments are generated from the Configurable traits of the classes associated
 with a given Application.  Configuring IPython from the command-line may look
 very similar to an IPython config file
 
 IPython applications use a parser called
 :class:`~IPython.config.loader.KeyValueLoader` to load values into a Config
 object.  Values are assigned in much the same way as in a config file:
 
 .. code-block:: bash
 
     $> ipython --InteractiveShell.use_readline=False --BaseIPythonApplication.profile='myprofile'
 
 Is the same as adding:
 
 .. sourcecode:: python
 
     c.InteractiveShell.use_readline=False
     c.BaseIPythonApplication.profile='myprofile'
 
 to your config file. Key/Value arguments *always* take a value, separated by '='
 and no spaces.
 
+Common Arguments
+****************
+
+Since the strictness and verbosity of the KVLoader above are not ideal for everyday
+use, common arguments can be specified as flags_ or aliases_.
+
+Flags and Aliases are handled by :mod:`argparse` instead, allowing for more flexible
+parsing. In general, flags and aliases are prefixed by ``--``, except for those
+that are single characters, in which case they can be specified with a single ``-``, e.g.:
+
+.. code-block:: bash
+
+    $> ipython -i -c "import numpy; x=numpy.linspace(0,1)" --profile testing --colors=lightbg
+
 Aliases
 -------
 
-For convenience, applications have a mapping of commonly
-used traits, so you don't have to specify the whole class name. For these **aliases**, the class need not be specified:
+For convenience, applications have a mapping of commonly used traits, so you don't have
+to specify the whole class name:
 
 .. code-block:: bash
 
+    $> ipython --profile myprofile
+    # and
     $> ipython --profile='myprofile'
-    # is equivalent to
+    # are equivalent to
     $> ipython --BaseIPythonApplication.profile='myprofile'
 
 Flags
 -----
 
 Applications can also be passed **flags**. Flags are options that take no
-arguments, and are always prefixed with ``--``. They are simply wrappers for
+arguments. They are simply wrappers for
 setting one or more configurables with predefined values, often True/False.
 
 For instance:
 
 .. code-block:: bash
 
     $> ipcontroller --debug
     # is equivalent to
     $> ipcontroller --Application.log_level=DEBUG
-    # and 
+    # and
     $> ipython --pylab
     # is equivalent to
     $> ipython --pylab=auto
+    # or
+    $> ipython --no-banner
+    # is equivalent to
+    $> ipython --TerminalIPythonApp.display_banner=False
 
 Subcommands
------------
+***********
 
 
 Some IPython applications have **subcommands**. Subcommands are modeled after
 :command:`git`, and are called with the form :command:`command subcommand
 [...args]`.  Currently, the QtConsole is a subcommand of terminal IPython:
 
 .. code-block:: bash
 
     $> ipython qtconsole --profile=myprofile
 
 and :command:`ipcluster` is simply a wrapper for its various subcommands (start,
 stop, engines).
 
 .. code-block:: bash
 
     $> ipcluster start --profile=myprofile --n=4
 
 
 To see a list of the available aliases, flags, and subcommands for an IPython application, simply pass ``-h`` or ``--help``.  And to see the full list of configurable options (*very* long), pass ``--help-all``.
 
 
 Design requirements
 ===================
 
 Here are the main requirements we wanted our configuration system to have:
 
 * Support for hierarchical configuration information.
 
 * Full integration with command line option parsers.  Often, you want to read
   a configuration file, but then override some of the values with command line
   options.  Our configuration system automates this process and allows each
   command line option to be linked to a particular attribute in the 
   configuration hierarchy that it will override.
 
 * Configuration files that are themselves valid Python code. This accomplishes
   many things. First, it becomes possible to put logic in your configuration
   files that sets attributes based on your operating system, network setup,
   Python version, etc. Second, Python has a super simple syntax for accessing
   hierarchical data structures, namely regular attribute access
   (``Foo.Bar.Bam.name``). Third, using Python makes it easy for users to
   import configuration attributes from one configuration file to another.
   Fourth, even though Python is dynamically typed, it does have types that can
   be checked at runtime. Thus, a ``1`` in a config file is the integer '1',
   while a ``'1'`` is a string.
 
 * A fully automated method for getting the configuration information to the
   classes that need it at runtime. Writing code that walks a configuration
   hierarchy to extract a particular attribute is painful. When you have
   complex configuration information with hundreds of attributes, this makes
   you want to cry.
 
 * Type checking and validation that doesn't require the entire configuration
   hierarchy to be specified statically before runtime. Python is a very
   dynamic language and you don't always know everything that needs to be
   configured when a program starts.
 
 
 .. _`XDG Base Directory`: http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html

docs/source/interactive/reference.txt

@@ -1,1310 +1,1310 b''
 =================
 IPython reference
 =================
 
 .. _command_line_options:
 
 Command-line usage
 ==================
 
 You start IPython with the command::
 
     $ ipython [options] files
 
 If invoked with no options, it executes all the files listed in sequence
 and drops you into the interpreter while still acknowledging any options
 you may have set in your ipython_config.py. This behavior is different from
 standard Python, which when called as python -i will only execute one
 file and ignore your configuration setup.
 
 Please note that some of the configuration options are not available at
 the command line, simply because they are not practical here. Look into
 your ipythonrc configuration file for details on those. This file is typically 
 installed in the IPYTHON_DIR directory. For Linux
 users, this will be $HOME/.config/ipython, and for other users it will be
 $HOME/.ipython.  For Windows users, $HOME resolves to C:\\Documents and
 Settings\\YourUserName in most instances.
 
 
 Eventloop integration
 ---------------------
 
 Previously IPython had command line options for controlling GUI event loop
 integration (-gthread, -qthread, -q4thread, -wthread, -pylab). As of IPython
 version 0.11, these have been removed. Please see the new ``%gui``
 magic command or :ref:`this section <gui_support>` for details on the new
 interface, or specify the gui at the commandline::
 
     $ ipython --gui=qt
 
 
 Regular Options
 ---------------
 
 After the above threading options have been given, regular options can
 follow in any order. All options can be abbreviated to their shortest
 non-ambiguous form and are case-sensitive. One or two dashes can be
 used. Some options have an alternate short form, indicated after a ``|``.
 
 Most options can also be set from your ipythonrc configuration file. See
 the provided example for more details on what the options do. Options
 given at the command line override the values set in the ipythonrc file.
 
 All options with a [no] prepended can be specified in negated form
 (--no-option instead of --option) to turn the feature off.
 
     ``-h, --help``  print a help message and exit.
 
     ``--pylab, pylab=<name>``
     See :ref:`Matplotlib support <matplotlib_support>`
     for more details.
 
     ``--autocall=<val>``
 	Make IPython automatically call any callable object even if you
 	didn't type explicit parentheses. For example, 'str 43' becomes
 	'str(43)' automatically. The value can be '0' to disable the feature,
 	'1' for smart autocall, where it is not applied if there are no more
 	arguments on the line, and '2' for full autocall, where all callable
 	objects are automatically called (even if no arguments are
 	present). The default is '1'.
 
     ``--[no-]autoindent``
 	Turn automatic indentation on/off.
 
     ``--[no-]automagic``
 	make magic commands automatic (without needing their first character
 	to be %). Type %magic at the IPython prompt for more information.
 
     ``--[no-]autoedit_syntax``
 	When a syntax error occurs after editing a file, automatically
 	open the file to the trouble causing line for convenient
 	fixing.
 
     ``--[no-]banner``
 	Print the initial information banner (default on).
 
-    ``--c=<command>``
+    ``-c <command>``
 	execute the given command string. This is similar to the -c
 	option in the normal Python interpreter.
 
     ``--cache-size=<n>``
 	size of the output cache (maximum number of entries to hold in
 	memory). The default is 1000, you can change it permanently in your
 	config file. Setting it to 0 completely disables the caching system,
 	and the minimum value accepted is 20 (if you provide a value less than
 	20, it is reset to 0 and a warning is issued) This limit is defined
 	because otherwise you'll spend more time re-flushing a too small cache
 	than working.
 
     ``--classic``
 	Gives IPython a similar feel to the classic Python
 	prompt.
 
     ``--colors=<scheme>``
 	Color scheme for prompts and exception reporting. Currently
 	implemented: NoColor, Linux and LightBG.
 
     ``--[no-]color_info``
 	IPython can display information about objects via a set of functions,
 	and optionally can use colors for this, syntax highlighting source
 	code and various other elements.  However, because this information is
 	passed through a pager (like 'less') and many pagers get confused with
 	color codes, this option is off by default. You can test it and turn
 	it on permanently in your ipythonrc file if it works for you. As a
 	reference, the 'less' pager supplied with Mandrake 8.2 works ok, but
 	that in RedHat 7.2 doesn't.
 
 	Test it and turn it on permanently if it works with your
 	system. The magic function %color_info allows you to toggle this
 	interactively for testing.
 
     ``--[no-]debug``
 	Show information about the loading process. Very useful to pin down
 	problems with your configuration files or to get details about
 	session restores.
 
     ``--[no-]deep_reload``
 	IPython can use the deep_reload module which reloads changes in
 	modules recursively (it replaces the reload() function, so you don't
 	need to change anything to use it).  deep_reload() forces a full
 	reload of modules whose code may have changed, which the default
 	reload() function does not.
 
 	When deep_reload is off, IPython will use the normal reload(),
 	but deep_reload will still be available as dreload(). This
 	feature is off by default [which means that you have both
 	normal reload() and dreload()].
 
     ``--editor=<name>``
 	Which editor to use with the %edit command. By default,
 	IPython will honor your EDITOR environment variable (if not
 	set, vi is the Unix default and notepad the Windows one).
 	Since this editor is invoked on the fly by IPython and is
 	meant for editing small code snippets, you may want to use a
 	small, lightweight editor here (in case your default EDITOR is
 	something like Emacs).
 
     ``--ipython_dir=<name>``
 	name of your IPython configuration directory IPYTHON_DIR. This
 	can also be specified through the environment variable
 	IPYTHON_DIR.
 
     ``--logfile=<name>``
         specify the name of your logfile.
         
         This implies ``%logstart`` at the beginning of your session
         
         generate a log file of all input. The file is named
         ipython_log.py in your current directory (which prevents logs
         from multiple IPython sessions from trampling each other). You
         can use this to later restore a session by loading your
-        logfile with ``ipython --i ipython_log.py``
+        logfile with ``ipython -i ipython_log.py``
 
     ``--logplay=<name>``
     
       NOT AVAILABLE in 0.11
 
       you can replay a previous log. For restoring a session as close as
       possible to the state you left it in, use this option (don't just run
       the logfile). With -logplay, IPython will try to reconstruct the
       previous working environment in full, not just execute the commands in
       the logfile.
 
       When a session is restored, logging is automatically turned on
       again with the name of the logfile it was invoked with (it is
       read from the log header). So once you've turned logging on for
       a session, you can quit IPython and reload it as many times as
       you want and it will continue to log its history and restore
       from the beginning every time.
       
       Caveats: there are limitations in this option. The history
       variables _i*,_* and _dh don't get restored properly. In the
       future we will try to implement full session saving by writing
       and retrieving a 'snapshot' of the memory state of IPython. But
       our first attempts failed because of inherent limitations of
       Python's Pickle module, so this may have to wait.
 
     ``--[no-]messages``
 	Print messages which IPython collects about its startup
 	process (default on).
 
     ``--[no-]pdb``
 	Automatically call the pdb debugger after every uncaught
 	exception. If you are used to debugging using pdb, this puts
 	you automatically inside of it after any call (either in
 	IPython or in code called by it) which triggers an exception
 	which goes uncaught.
 
     ``--[no-]pprint``
 	ipython can optionally use the pprint (pretty printer) module
 	for displaying results. pprint tends to give a nicer display
 	of nested data structures. If you like it, you can turn it on
 	permanently in your config file (default off).
 
     ``--profile=<name>``
 
 	Select the IPython profile by name.
     
 	This is a quick way to keep and load multiple
 	config files for different tasks, especially if you use the
 	include option of config files. You can keep a basic
 	:file:`IPYTHON_DIR/profile_default/ipython_config.py` file 
 	and then have other 'profiles' which
 	include this one and load extra things for particular
 	tasks. For example:
 
 	1. $IPYTHON_DIR/profile_default : load basic things you always want.
 	2. $IPYTHON_DIR/profile_math : load (1) and basic math-related modules.
 	3. $IPYTHON_DIR/profile_numeric : load (1) and Numeric and plotting modules.
 
 	Since it is possible to create an endless loop by having
 	circular file inclusions, IPython will stop if it reaches 15
 	recursive inclusions.
 
     ``InteractiveShell.prompt_in1=<string>``
 
 	Specify the string used for input prompts. Note that if you are using
 	numbered prompts, the number is represented with a '\#' in the
 	string. Don't forget to quote strings with spaces embedded in
 	them. Default: 'In [\#]:'.  The :ref:`prompts section <prompts>`
 	discusses in detail all the available escapes to customize your
 	prompts.
 
     ``InteractiveShell.prompt_in2=<string>``
 	Similar to the previous option, but used for the continuation
 	prompts. The special sequence '\D' is similar to '\#', but
 	with all digits replaced dots (so you can have your
 	continuation prompt aligned with your input prompt).  Default:
 	' .\D.:' (note three spaces at the start for alignment with
 	'In [\#]').
 
     ``InteractiveShell.prompt_out=<string>``
 	String used for output prompts, also uses numbers like
 	prompt_in1. Default: 'Out[\#]:'
 
     ``--quick``
         start in bare bones mode (no config file loaded).
 
     ``config_file=<name>``
 	name of your IPython resource configuration file. Normally
 	IPython loads ipython_config.py (from current directory) or
 	IPYTHON_DIR/profile_default.
 
 	If the loading of your config file fails, IPython starts with
 	a bare bones configuration (no modules loaded at all).
 
     ``--[no-]readline``
 	use the readline library, which is needed to support name
 	completion and command history, among other things.  It is
 	enabled by default, but may cause problems for users of
 	X/Emacs in Python comint or shell buffers.
 
 	Note that X/Emacs 'eterm' buffers (opened with M-x term) support
 	IPython's readline and syntax coloring fine, only 'emacs' (M-x
 	shell and C-c !) buffers do not.
 
     ``--TerminalInteractiveShell.screen_length=<n>``
 	number of lines of your screen. This is used to control
 	printing of very long strings. Strings longer than this number
 	of lines will be sent through a pager instead of directly
 	printed.
 
 	The default value for this is 0, which means IPython will
 	auto-detect your screen size every time it needs to print certain
 	potentially long strings (this doesn't change the behavior of the
 	'print' keyword, it's only triggered internally). If for some
 	reason this isn't working well (it needs curses support), specify
 	it yourself. Otherwise don't change the default.
 
     ``--TerminalInteractiveShell.separate_in=<string>``
 
 	separator before input prompts.
 	Default: '\n'
 
     ``--TerminalInteractiveShell.separate_out=<string>``
         separator before output prompts.
         Default: nothing.
 
     ``--TerminalInteractiveShell.separate_out2=<string>``
 	separator after output prompts.
 	Default: nothing.
 	For these three options, use the value 0 to specify no separator.
 
     ``--nosep``
 	shorthand for setting the above separators to empty strings.
 	
 	Simply removes all input/output separators.
 
     ``--init``
 	allows you to initialize a profile dir for configuration when you
 	install a new version of IPython or want to use a new profile.
 	Since new versions may include new command line options or example
 	files, this copies updated config files. Note that you should probably
 	use %upgrade instead,it's a safer alternative.
 
     ``--version``    print version information and exit.
 
     ``--xmode=<modename>``
 
 	Mode for exception reporting.
 
 	Valid modes: Plain, Context and Verbose.
 
 	* Plain: similar to python's normal traceback printing.
 	* Context: prints 5 lines of context source code around each
 	  line in the traceback.
 	* Verbose: similar to Context, but additionally prints the
 	  variables currently visible where the exception happened
 	  (shortening their strings if too long). This can potentially be
 	  very slow, if you happen to have a huge data structure whose
 	  string representation is complex to compute. Your computer may
 	  appear to freeze for a while with cpu usage at 100%. If this
 	  occurs, you can cancel the traceback with Ctrl-C (maybe hitting it
 	  more than once).
 
 Interactive use
 ===============
 
 IPython is meant to work as a drop-in replacement for the standard interactive
 interpreter. As such, any code which is valid python should execute normally
 under IPython (cases where this is not true should be reported as bugs). It
 does, however, offer many features which are not available at a standard python
 prompt. What follows is a list of these.
 
 
 Caution for Windows users
 -------------------------
 
 Windows, unfortunately, uses the '\\' character as a path separator. This is a
 terrible choice, because '\\' also represents the escape character in most
 modern programming languages, including Python. For this reason, using '/'
 character is recommended if you have problems with ``\``.  However, in Windows
 commands '/' flags options, so you can not use it for the root directory. This
 means that paths beginning at the root must be typed in a contrived manner
 like: ``%copy \opt/foo/bar.txt \tmp``
 
 .. _magic:
 
 Magic command system
 --------------------
 
 IPython will treat any line whose first character is a % as a special
 call to a 'magic' function. These allow you to control the behavior of
 IPython itself, plus a lot of system-type features. They are all
 prefixed with a % character, but parameters are given without
 parentheses or quotes.
 
 Example: typing ``%cd mydir`` changes your working directory to 'mydir', if it
 exists.
 
 If you have 'automagic' enabled (as it by default), you don't need
 to type in the % explicitly. IPython will scan its internal list of
 magic functions and call one if it exists. With automagic on you can
 then just type ``cd mydir`` to go to directory 'mydir'. The automagic
 system has the lowest possible precedence in name searches, so defining
 an identifier with the same name as an existing magic function will
 shadow it for automagic use. You can still access the shadowed magic
 function by explicitly using the % character at the beginning of the line.
 
 An example (with automagic on) should clarify all this:
 
 .. sourcecode:: ipython
 
     In [1]: cd ipython # %cd is called by automagic
 
     /home/fperez/ipython
 
     In [2]: cd=1 # now cd is just a variable
 
     In [3]: cd .. # and doesn't work as a function anymore
 
     ------------------------------
 
         File "<console>", line 1
 
           cd ..
 
               ^
 
     SyntaxError: invalid syntax
 
     In [4]: %cd .. # but %cd always works
 
     /home/fperez
 
     In [5]: del cd # if you remove the cd variable
 
     In [6]: cd ipython # automagic can work again
 
     /home/fperez/ipython
 
 You can define your own magic functions to extend the system. The
 following example defines a new magic command, %impall:
 
 .. sourcecode:: python
 
     ip = get_ipython()
 
     def doimp(self, arg):
 
         ip = self.api
 
         ip.ex("import %s; reload(%s); from %s import *" % (
 
         arg,arg,arg)
 
         )
 
     ip.expose_magic('impall', doimp)
 
 Type `%magic` for more information, including a list of all available magic
 functions at any time and their docstrings. You can also type
 %magic_function_name? (see :ref:`below <dynamic_object_info` for information on
 the '?' system) to get information about any particular magic function you are
 interested in.
 
 The API documentation for the :mod:`IPython.core.magic` module contains the full
 docstrings of all currently available magic commands.
 
 
 Access to the standard Python help
 ----------------------------------
 
 As of Python 2.1, a help system is available with access to object docstrings
 and the Python manuals. Simply type 'help' (no quotes) to access it. You can
 also type help(object) to obtain information about a given object, and
 help('keyword') for information on a keyword. As noted :ref:`here
 <accessing_help>`, you need to properly configure your environment variable
 PYTHONDOCS for this feature to work correctly.
 
 .. _dynamic_object_info:
 
 Dynamic object information
 --------------------------
 
 Typing ``?word`` or ``word?`` prints detailed information about an object. If
 certain strings in the object are too long (docstrings, code, etc.) they get
 snipped in the center for brevity. This system gives access variable types and
 values, full source code for any object (if available), function prototypes and
 other useful information.
 
 Typing ``??word`` or ``word??`` gives access to the full information without
 snipping long strings. Long strings are sent to the screen through the
 less pager if longer than the screen and printed otherwise. On systems
 lacking the less command, IPython uses a very basic internal pager.
 
 The following magic functions are particularly useful for gathering
 information about your working environment. You can get more details by
 typing ``%magic`` or querying them individually (use %function_name? with or
 without the %), this is just a summary:
 
     * **%pdoc <object>**: Print (or run through a pager if too long) the
       docstring for an object. If the given object is a class, it will
       print both the class and the constructor docstrings.
     * **%pdef <object>**: Print the definition header for any callable
       object. If the object is a class, print the constructor information.
     * **%psource <object>**: Print (or run through a pager if too long)
       the source code for an object.
     * **%pfile <object>**: Show the entire source file where an object was
       defined via a pager, opening it at the line where the object
       definition begins.
     * **%who/%whos**: These functions give information about identifiers
       you have defined interactively (not things you loaded or defined
       in your configuration files). %who just prints a list of
       identifiers and %whos prints a table with some basic details about
       each identifier.
 
 Note that the dynamic object information functions (?/??, ``%pdoc``,
 ``%pfile``, ``%pdef``, ``%psource``) give you access to documentation even on
 things which are not really defined as separate identifiers. Try for example
 typing {}.get? or after doing import os, type ``os.path.abspath??``.
 
 .. _readline:
 
 Readline-based features
 -----------------------
 
 These features require the GNU readline library, so they won't work if your
 Python installation lacks readline support. We will first describe the default
 behavior IPython uses, and then how to change it to suit your preferences.
 
 
 Command line completion
 +++++++++++++++++++++++
 
 At any time, hitting TAB will complete any available python commands or
 variable names, and show you a list of the possible completions if
 there's no unambiguous one. It will also complete filenames in the
 current directory if no python names match what you've typed so far.
 
 
 Search command history
 ++++++++++++++++++++++
 
 IPython provides two ways for searching through previous input and thus
 reduce the need for repetitive typing:
 
    1. Start typing, and then use Ctrl-p (previous,up) and Ctrl-n
       (next,down) to search through only the history items that match
       what you've typed so far. If you use Ctrl-p/Ctrl-n at a blank
       prompt, they just behave like normal arrow keys.
    2. Hit Ctrl-r: opens a search prompt. Begin typing and the system
       searches your history for lines that contain what you've typed so
       far, completing as much as it can.
 
 
 Persistent command history across sessions
 ++++++++++++++++++++++++++++++++++++++++++
 
 IPython will save your input history when it leaves and reload it next
 time you restart it. By default, the history file is named
 $IPYTHON_DIR/profile_<name>/history.sqlite. This allows you to keep
 separate histories related to various tasks: commands related to
 numerical work will not be clobbered by a system shell history, for
 example.
 
 
 Autoindent
 ++++++++++
 
 IPython can recognize lines ending in ':' and indent the next line,
 while also un-indenting automatically after 'raise' or 'return'.
 
 This feature uses the readline library, so it will honor your
 :file:`~/.inputrc` configuration (or whatever file your INPUTRC variable points
 to). Adding the following lines to your :file:`.inputrc` file can make
 indenting/unindenting more convenient (M-i indents, M-u unindents)::
 
     $if Python 
     "\M-i": "    "  
     "\M-u": "\d\d\d\d"  
     $endif
 
 Note that there are 4 spaces between the quote marks after "M-i" above.
 
 .. warning::
 
     Setting the above indents will cause problems with unicode text entry in
     the terminal.
 
 .. warning::
 
     Autoindent is ON by default, but it can cause problems with the pasting of
     multi-line indented code (the pasted code gets re-indented on each line). A
     magic function %autoindent allows you to toggle it on/off at runtime. You
     can also disable it permanently on in your :file:`ipython_config.py` file
     (set TerminalInteractiveShell.autoindent=False).
     
     If you want to paste multiple lines, it is recommended that you use
     ``%paste``.
 
 
 Customizing readline behavior
 +++++++++++++++++++++++++++++
 
 All these features are based on the GNU readline library, which has an
 extremely customizable interface. Normally, readline is configured via a
 file which defines the behavior of the library; the details of the
 syntax for this can be found in the readline documentation available
 with your system or on the Internet. IPython doesn't read this file (if
 it exists) directly, but it does support passing to readline valid
 options via a simple interface. In brief, you can customize readline by
 setting the following options in your ipythonrc configuration file (note
 that these options can not be specified at the command line):
 
     * **readline_parse_and_bind**: this option can appear as many times as
       you want, each time defining a string to be executed via a
       readline.parse_and_bind() command. The syntax for valid commands
       of this kind can be found by reading the documentation for the GNU
       readline library, as these commands are of the kind which readline
       accepts in its configuration file.
     * **readline_remove_delims**: a string of characters to be removed
       from the default word-delimiters list used by readline, so that
       completions may be performed on strings which contain them. Do not
       change the default value unless you know what you're doing.
     * **readline_omit__names**: when tab-completion is enabled, hitting
       <tab> after a '.' in a name will complete all attributes of an
       object, including all the special methods whose names include
       double underscores (like __getitem__ or __class__). If you'd
       rather not see these names by default, you can set this option to
       1. Note that even when this option is set, you can still see those
       names by explicitly typing a _ after the period and hitting <tab>:
       'name._<tab>' will always complete attribute names starting with '_'.
       
       This option is off by default so that new users see all
       attributes of any objects they are dealing with.
 
 You will find the default values along with a corresponding detailed
 explanation in your ipythonrc file.
 
 
 Session logging and restoring
 -----------------------------
 
 You can log all input from a session either by starting IPython with the
 command line switch ``--logfile=foo.py`` (see :ref:`here <command_line_options>`)
 or by activating the logging at any moment with the magic function %logstart.
 
 Log files can later be reloaded by running them as scripts and IPython
 will attempt to 'replay' the log by executing all the lines in it, thus
 restoring the state of a previous session. This feature is not quite
 perfect, but can still be useful in many cases.
 
 The log files can also be used as a way to have a permanent record of
 any code you wrote while experimenting. Log files are regular text files
 which you can later open in your favorite text editor to extract code or
 to 'clean them up' before using them to replay a session.
 
 The `%logstart` function for activating logging in mid-session is used as
 follows::
 
     %logstart [log_name [log_mode]]
 
 If no name is given, it defaults to a file named 'ipython_log.py' in your
 current working directory, in 'rotate' mode (see below).
 
 '%logstart name' saves to file 'name' in 'backup' mode. It saves your
 history up to that point and then continues logging.
 
 %logstart takes a second optional parameter: logging mode. This can be
 one of (note that the modes are given unquoted):
 
     * [over:] overwrite existing log_name.
     * [backup:] rename (if exists) to log_name~ and start log_name.
     * [append:] well, that says it.
     * [rotate:] create rotating logs log_name.1~, log_name.2~, etc.
 
 The %logoff and %logon functions allow you to temporarily stop and
 resume logging to a file which had previously been started with
 %logstart. They will fail (with an explanation) if you try to use them
 before logging has been started.
 
 .. _system_shell_access:
 
 System shell access
 -------------------
 
 Any input line beginning with a ! character is passed verbatim (minus
 the !, of course) to the underlying operating system. For example,
 typing ``!ls`` will run 'ls' in the current directory.
 
 Manual capture of command output
 --------------------------------
 
 If the input line begins with two exclamation marks, !!, the command is
 executed but its output is captured and returned as a python list, split
 on newlines. Any output sent by the subprocess to standard error is
 printed separately, so that the resulting list only captures standard
 output. The !! syntax is a shorthand for the %sx magic command.
 
 Finally, the %sc magic (short for 'shell capture') is similar to %sx,
 but allowing more fine-grained control of the capture details, and
 storing the result directly into a named variable. The direct use of 
 %sc is now deprecated, and you should ise the ``var = !cmd`` syntax
 instead.
 
 IPython also allows you to expand the value of python variables when
 making system calls. Any python variable or expression which you prepend
 with $ will get expanded before the system call is made::
 
     In [1]: pyvar='Hello world' 
     In [2]: !echo "A python variable: $pyvar"  
     A python variable: Hello world
 
 If you want the shell to actually see a literal $, you need to type it
 twice::
 
     In [3]: !echo "A system variable: $$HOME"  
     A system variable: /home/fperez
 
 You can pass arbitrary expressions, though you'll need to delimit them
 with {} if there is ambiguity as to the extent of the expression::
 
     In [5]: x=10 
     In [6]: y=20 
     In [13]: !echo $x+y  
     10+y  
     In [7]: !echo ${x+y}  
     30
 
 Even object attributes can be expanded::
 
     In [12]: !echo $sys.argv  
     [/home/fperez/usr/bin/ipython]
 
 
 System command aliases
 ----------------------
 
 The %alias magic function and the alias option in the ipythonrc
 configuration file allow you to define magic functions which are in fact
 system shell commands. These aliases can have parameters.
 
 ``%alias alias_name cmd`` defines 'alias_name' as an alias for 'cmd'
 
 Then, typing ``%alias_name params`` will execute the system command 'cmd
 params' (from your underlying operating system).
 
 You can also define aliases with parameters using %s specifiers (one per
 parameter). The following example defines the %parts function as an
 alias to the command 'echo first %s second %s' where each %s will be
 replaced by a positional parameter to the call to %parts::
 
     In [1]: alias parts echo first %s second %s 
     In [2]: %parts A B 
     first A second B  
     In [3]: %parts A  
     Incorrect number of arguments: 2 expected.  
     parts is an alias to: 'echo first %s second %s'
 
 If called with no parameters, %alias prints the table of currently
 defined aliases.
 
 The %rehashx magic allows you to load your entire $PATH as
 ipython aliases. See its docstring for further details.
 
 
 .. _dreload:
 
 Recursive reload
 ----------------
 
 The dreload function does a recursive reload of a module: changes made
 to the module since you imported will actually be available without
 having to exit.
 
 
 Verbose and colored exception traceback printouts
 -------------------------------------------------
 
 IPython provides the option to see very detailed exception tracebacks,
 which can be especially useful when debugging large programs. You can
 run any Python file with the %run function to benefit from these
 detailed tracebacks. Furthermore, both normal and verbose tracebacks can
 be colored (if your terminal supports it) which makes them much easier
 to parse visually.
 
 See the magic xmode and colors functions for details (just type %magic).
 
 These features are basically a terminal version of Ka-Ping Yee's cgitb
 module, now part of the standard Python library.
 
 
 .. _input_caching:
 
 Input caching system
 --------------------
 
 IPython offers numbered prompts (In/Out) with input and output caching
 (also referred to as 'input history'). All input is saved and can be 
 retrieved as variables (besides the usual arrow key recall), in 
 addition to the %rep magic command that brings a history entry
 up for editing on the next  command line.
 
 The following GLOBAL variables always exist (so don't overwrite them!):
 
 * _i, _ii, _iii: store previous, next previous and next-next previous inputs.
 * In, _ih : a list of all inputs; _ih[n] is the input from line n. If you
   overwrite In with a variable of your own, you can remake the assignment to the
   internal list with a simple ``In=_ih``.
 
 Additionally, global variables named _i<n> are dynamically created (<n>
 being the prompt counter), so ``_i<n> == _ih[<n>] == In[<n>]``.
 
 For example, what you typed at prompt 14 is available as _i14, _ih[14]
 and In[14].
 
 This allows you to easily cut and paste multi line interactive prompts
 by printing them out: they print like a clean string, without prompt
 characters. You can also manipulate them like regular variables (they
 are strings), modify or exec them (typing ``exec _i9`` will re-execute the
 contents of input prompt 9.
 
 You can also re-execute multiple lines of input easily by using the
 magic %macro function (which automates the process and allows
 re-execution without having to type 'exec' every time). The macro system
 also allows you to re-execute previous lines which include magic
 function calls (which require special processing). Type %macro? for more details
 on the macro system.
 
 A history function %hist allows you to see any part of your input
 history by printing a range of the _i variables.
 
 You can also search ('grep') through your history by typing 
 ``%hist -g somestring``. This is handy for searching for URLs, IP addresses,
 etc. You can bring history entries listed by '%hist -g' up for editing
 with the %recall command, or run them immediately with %rerun.
 
 .. _output_caching:
 
 Output caching system
 ---------------------
 
 For output that is returned from actions, a system similar to the input
 cache exists but using _ instead of _i. Only actions that produce a
 result (NOT assignments, for example) are cached. If you are familiar
 with Mathematica, IPython's _ variables behave exactly like
 Mathematica's % variables.
 
 The following GLOBAL variables always exist (so don't overwrite them!):
 
     * [_] (a single underscore) : stores previous output, like Python's
       default interpreter.
     * [__] (two underscores): next previous.
     * [___] (three underscores): next-next previous.
 
 Additionally, global variables named _<n> are dynamically created (<n>
 being the prompt counter), such that the result of output <n> is always
 available as _<n> (don't use the angle brackets, just the number, e.g.
 _21).
 
 These global variables are all stored in a global dictionary (not a
 list, since it only has entries for lines which returned a result)
 available under the names _oh and Out (similar to _ih and In). So the
 output from line 12 can be obtained as _12, Out[12] or _oh[12]. If you
 accidentally overwrite the Out variable you can recover it by typing
 'Out=_oh' at the prompt.
 
 This system obviously can potentially put heavy memory demands on your
 system, since it prevents Python's garbage collector from removing any
 previously computed results. You can control how many results are kept
 in memory with the option (at the command line or in your ipythonrc
 file) cache_size. If you set it to 0, the whole system is completely
 disabled and the prompts revert to the classic '>>>' of normal Python.
 
 
 Directory history
 -----------------
 
 Your history of visited directories is kept in the global list _dh, and
 the magic %cd command can be used to go to any entry in that list. The
 %dhist command allows you to view this history. Do ``cd -<TAB>`` to
 conveniently view the directory history.
 
 
 Automatic parentheses and quotes
 --------------------------------
 
 These features were adapted from Nathan Gray's LazyPython. They are
 meant to allow less typing for common situations.
 
 
 Automatic parentheses
 ---------------------
 
 Callable objects (i.e. functions, methods, etc) can be invoked like this
 (notice the commas between the arguments)::
 
     >>> callable_ob arg1, arg2, arg3
 
 and the input will be translated to this::
 
     -> callable_ob(arg1, arg2, arg3)
 
 You can force automatic parentheses by using '/' as the first character
 of a line. For example::
 
     >>> /globals # becomes 'globals()'
 
 Note that the '/' MUST be the first character on the line! This won't work::
 
     >>> print /globals # syntax error
 
 In most cases the automatic algorithm should work, so you should rarely
 need to explicitly invoke /. One notable exception is if you are trying
 to call a function with a list of tuples as arguments (the parenthesis
 will confuse IPython)::
 
     In [1]: zip (1,2,3),(4,5,6) # won't work
 
 but this will work::
 
     In [2]: /zip (1,2,3),(4,5,6) 
     ---> zip ((1,2,3),(4,5,6))  
     Out[2]= [(1, 4), (2, 5), (3, 6)]
 
 IPython tells you that it has altered your command line by displaying
 the new command line preceded by ->. e.g.::
 
     In [18]: callable list 
     ----> callable (list)
 
 
 Automatic quoting
 -----------------
 
 You can force automatic quoting of a function's arguments by using ','
 or ';' as the first character of a line. For example::
 
     >>> ,my_function /home/me # becomes my_function("/home/me")
 
 If you use ';' instead, the whole argument is quoted as a single string
 (while ',' splits on whitespace)::
 
     >>> ,my_function a b c # becomes my_function("a","b","c")
 
     >>> ;my_function a b c # becomes my_function("a b c")
 
 Note that the ',' or ';' MUST be the first character on the line! This
 won't work::
 
     >>> x = ,my_function /home/me # syntax error
 
 IPython as your default Python environment
 ==========================================
 
 Python honors the environment variable PYTHONSTARTUP and will execute at
 startup the file referenced by this variable. If you put at the end of
 this file the following two lines of code::
 
     from IPython.frontend.terminal.ipapp import launch_new_instance
     launch_new_instance()
     raise SystemExit
 
 then IPython will be your working environment anytime you start Python.
 The ``raise SystemExit`` is needed to exit Python when
 it finishes, otherwise you'll be back at the normal Python '>>>'
 prompt.
 
 This is probably useful to developers who manage multiple Python
 versions and don't want to have correspondingly multiple IPython
 versions. Note that in this mode, there is no way to pass IPython any
 command-line options, as those are trapped first by Python itself.
 
 .. _Embedding:
 
 Embedding IPython
 =================
 
 It is possible to start an IPython instance inside your own Python
 programs. This allows you to evaluate dynamically the state of your
 code, operate with your variables, analyze them, etc. Note however that
 any changes you make to values while in the shell do not propagate back
 to the running code, so it is safe to modify your values because you
 won't break your code in bizarre ways by doing so.
 
 This feature allows you to easily have a fully functional python
 environment for doing object introspection anywhere in your code with a
 simple function call. In some cases a simple print statement is enough,
 but if you need to do more detailed analysis of a code fragment this
 feature can be very valuable.
 
 It can also be useful in scientific computing situations where it is
 common to need to do some automatic, computationally intensive part and
 then stop to look at data, plots, etc.
 Opening an IPython instance will give you full access to your data and
 functions, and you can resume program execution once you are done with
 the interactive part (perhaps to stop again later, as many times as
 needed).
 
 The following code snippet is the bare minimum you need to include in
 your Python programs for this to work (detailed examples follow later)::
 
     from IPython import embed
 
     embed() # this call anywhere in your program will start IPython
 
 You can run embedded instances even in code which is itself being run at
 the IPython interactive prompt with '%run <filename>'. Since it's easy
 to get lost as to where you are (in your top-level IPython or in your
 embedded one), it's a good idea in such cases to set the in/out prompts
 to something different for the embedded instances. The code examples
 below illustrate this.
 
 You can also have multiple IPython instances in your program and open
 them separately, for example with different options for data
 presentation. If you close and open the same instance multiple times,
 its prompt counters simply continue from each execution to the next.
 
 Please look at the docstrings in the :mod:`~IPython.frontend.terminal.embed` 
 module for more details on the use of this system.
 
 The following sample file illustrating how to use the embedding
 functionality is provided in the examples directory as example-embed.py.
 It should be fairly self-explanatory:
 
 .. literalinclude:: ../../examples/core/example-embed.py
     :language: python
 
 Once you understand how the system functions, you can use the following
 code fragments in your programs which are ready for cut and paste:
 
 .. literalinclude:: ../../examples/core/example-embed-short.py
     :language: python
 
 Using the Python debugger (pdb)
 ===============================
 
 Running entire programs via pdb
 -------------------------------
 
 pdb, the Python debugger, is a powerful interactive debugger which
 allows you to step through code, set breakpoints, watch variables,
 etc.  IPython makes it very easy to start any script under the control
 of pdb, regardless of whether you have wrapped it into a 'main()'
 function or not. For this, simply type '%run -d myscript' at an
 IPython prompt. See the %run command's documentation (via '%run?' or
 in Sec. magic_ for more details, including how to control where pdb
 will stop execution first.
 
 For more information on the use of the pdb debugger, read the included
 pdb.doc file (part of the standard Python distribution). On a stock
 Linux system it is located at /usr/lib/python2.3/pdb.doc, but the
 easiest way to read it is by using the help() function of the pdb module
 as follows (in an IPython prompt)::
 
     In [1]: import pdb
     In [2]: pdb.help()
 
 This will load the pdb.doc document in a file viewer for you automatically.
 
 
 Automatic invocation of pdb on exceptions
 -----------------------------------------
 
 IPython, if started with the -pdb option (or if the option is set in
 your rc file) can call the Python pdb debugger every time your code
 triggers an uncaught exception. This feature
 can also be toggled at any time with the %pdb magic command. This can be
 extremely useful in order to find the origin of subtle bugs, because pdb
 opens up at the point in your code which triggered the exception, and
 while your program is at this point 'dead', all the data is still
 available and you can walk up and down the stack frame and understand
 the origin of the problem.
 
 Furthermore, you can use these debugging facilities both with the
 embedded IPython mode and without IPython at all. For an embedded shell
 (see sec. Embedding_), simply call the constructor with
 '--pdb' in the argument string and automatically pdb will be called if an
 uncaught exception is triggered by your code.
 
 For stand-alone use of the feature in your programs which do not use
 IPython at all, put the following lines toward the top of your 'main'
 routine::
 
     import sys
     from IPython.core import ultratb
     sys.excepthook = ultratb.FormattedTB(mode='Verbose',
     color_scheme='Linux', call_pdb=1)
 
 The mode keyword can be either 'Verbose' or 'Plain', giving either very
 detailed or normal tracebacks respectively. The color_scheme keyword can
 be one of 'NoColor', 'Linux' (default) or 'LightBG'. These are the same
 options which can be set in IPython with -colors and -xmode.
 
 This will give any of your programs detailed, colored tracebacks with
 automatic invocation of pdb.
 
 
 Extensions for syntax processing
 ================================
 
 This isn't for the faint of heart, because the potential for breaking
 things is quite high. But it can be a very powerful and useful feature.
 In a nutshell, you can redefine the way IPython processes the user input
 line to accept new, special extensions to the syntax without needing to
 change any of IPython's own code.
 
 In the IPython/extensions directory you will find some examples
 supplied, which we will briefly describe now. These can be used 'as is'
 (and both provide very useful functionality), or you can use them as a
 starting point for writing your own extensions.
 
 .. _pasting_with_prompts:
 
 Pasting of code starting with Python or IPython prompts
 -------------------------------------------------------
 
 IPython is smart enough to filter  out input prompts, be they plain Python ones
 (``>>>`` and ``...``) or IPython ones (``In [N]:`` and `` ...:``).  You can
 therefore copy and paste from existing interactive sessions without worry.
 
 The following is a 'screenshot' of how things work, copying an example from the
 standard Python tutorial::
 
     In [1]: >>> # Fibonacci series:
 
     In [2]: ... # the sum of two elements defines the next
 
     In [3]: ... a, b = 0, 1
 
     In [4]: >>> while b < 10:
        ...:     ...     print b
        ...:     ...     a, b = b, a+b
        ...:     
     1
     1
     2
     3
     5
     8
 
 And pasting from IPython sessions works equally well::
 
     In [1]: In [5]: def f(x):
        ...:        ...:     "A simple function"
        ...:        ...:     return x**2
        ...:    ...: 
 
     In [2]: f(3)
     Out[2]: 9
 
 .. _gui_support:
 
 GUI event loop support
 ======================
 
 .. versionadded:: 0.11
     The ``%gui`` magic and :mod:`IPython.lib.inputhook`.
 
 .. warning::
 
    All GUI support with the ``%gui`` magic, described in this section, applies
    only to the plain terminal IPython, *not* to the Qt console.  The Qt console
    currently only supports GUI interaction via the ``--pylab`` flag, as
    explained :ref:`in the matplotlib section <matplotlib_support>`.
 
    We intend to correct this limitation as soon as possible, you can track our
    progress at issue #643_.
 
 .. _643: https://github.com/ipython/ipython/issues/643
     
 IPython has excellent support for working interactively with Graphical User
 Interface (GUI) toolkits, such as wxPython, PyQt4, PyGTK and Tk. This is
 implemented using Python's builtin ``PyOSInputHook`` hook. This implementation
 is extremely robust compared to our previous thread-based version. The
 advantages of this are:
 
 * GUIs can be enabled and disabled dynamically at runtime.
 * The active GUI can be switched dynamically at runtime.
 * In some cases, multiple GUIs can run simultaneously with no problems.
 * There is a developer API in :mod:`IPython.lib.inputhook` for customizing 
   all of these things.
 
 For users, enabling GUI event loop integration is simple.  You simple use the
 ``%gui`` magic as follows::
 
     %gui [GUINAME]
 
 With no arguments, ``%gui`` removes all GUI support.  Valid ``GUINAME``
 arguments are ``wx``, ``qt4``, ``gtk`` and ``tk``.
 
 Thus, to use wxPython interactively and create a running :class:`wx.App`
 object, do::
 
     %gui wx
 
 For information on IPython's Matplotlib integration (and the ``pylab`` mode)
 see :ref:`this section <matplotlib_support>`.
 
 For developers that want to use IPython's GUI event loop integration in the
 form of a library, these capabilities are exposed in library form in the
 :mod:`IPython.lib.inputhook` and :mod:`IPython.lib.guisupport` modules.
 Interested developers should see the module docstrings for more information,
 but there are a few points that should be mentioned here.
 
 First, the ``PyOSInputHook`` approach only works in command line settings 
 where readline is activated.  As indicated in the warning above, we plan on
 improving the integration of GUI event loops with the standalone kernel used by
 the Qt console and other frontends (issue 643_).
 
 Second, when using the ``PyOSInputHook`` approach, a GUI application should
 *not* start its event loop. Instead all of this is handled by the
 ``PyOSInputHook``. This means that applications that are meant to be used both
 in IPython and as standalone apps need to have special code to detects how the
 application is being run. We highly recommend using IPython's support for this.
 Since the details vary slightly between toolkits, we point you to the various
 examples in our source directory :file:`docs/examples/lib` that demonstrate
 these capabilities.
 
 .. warning::
 
    The WX version of this is currently broken.  While ``--pylab=wx`` works
    fine, standalone WX apps do not.  See
    https://github.com/ipython/ipython/issues/645 for details of our progress on
    this issue.
 
 
 Third, unlike previous versions of IPython, we no longer "hijack" (replace
 them with no-ops) the event loops. This is done to allow applications that
 actually need to run the real event loops to do so. This is often needed to
 process pending events at critical points.
 
 Finally, we also have a number of examples in our source directory
 :file:`docs/examples/lib` that demonstrate these capabilities.
 
 PyQt and PySide
 ---------------
 
 .. attempt at explanation of the complete mess that is Qt support
 
 When you use ``--gui=qt`` or ``--pylab=qt``, IPython can work with either
 PyQt4 or PySide.  There are three options for configuration here, because
 PyQt4 has two APIs for QString and QVariant - v1, which is the default on
 Python 2, and the more natural v2, which is the only API supported by PySide.
 v2 is also the default for PyQt4 on Python 3.  IPython's code for the QtConsole
 uses v2, but you can still use any interface in your code, since the
 Qt frontend is in a different process.
 
 The default will be to import PyQt4 without configuration of the APIs, thus
 matching what most applications would expect. It will fall back of PySide if
 PyQt4 is unavailable.
 
 If specified, IPython will respect the environment variable ``QT_API`` used
 by ETS.  ETS 4.0 also works with both PyQt4 and PySide, but it requires
 PyQt4 to use its v2 API.  So if ``QT_API=pyside`` PySide will be used,
 and if ``QT_API=pyqt`` then PyQt4 will be used *with the v2 API* for
 QString and QVariant, so ETS codes like MayaVi will also work with IPython.
 
 If you launch IPython in pylab mode with ``ipython --pylab=qt``, then IPython
 will ask matplotlib which Qt library to use (only if QT_API is *not set*), via
 the 'backend.qt4' rcParam.  If matplotlib is version 1.0.1 or older, then
 IPython will always use PyQt4 without setting the v2 APIs, since neither v2
 PyQt nor PySide work.
 
 .. warning::
 
     Note that this means for ETS 4 to work with PyQt4, ``QT_API`` *must* be set
     to work with IPython's qt integration, because otherwise PyQt4 will be
     loaded in an incompatible mode.
     
     It also means that you must *not* have ``QT_API`` set if you want to
     use ``--gui=qt`` with code that requires PyQt4 API v1.
 
 
 .. _matplotlib_support:
 
 Plotting with matplotlib
 ========================
 
 `Matplotlib`_ provides high quality 2D and 3D plotting for Python. Matplotlib
 can produce plots on screen using a variety of GUI toolkits, including Tk,
 PyGTK, PyQt4 and wxPython. It also provides a number of commands useful for
 scientific computing, all with a syntax compatible with that of the popular
 Matlab program.
 
 To start IPython with matplotlib support, use the ``--pylab`` switch.  If no
 arguments are given, IPython will automatically detect your choice of
 matplotlib backend.  You can also request a specific backend with
 ``--pylab=backend``, where ``backend`` must be one of: 'tk', 'qt', 'wx', 'gtk',
 'osx'.
 
 .. _Matplotlib: http://matplotlib.sourceforge.net
 
 .. _interactive_demos:
 
 Interactive demos with IPython
 ==============================
 
 IPython ships with a basic system for running scripts interactively in
 sections, useful when presenting code to audiences. A few tags embedded
 in comments (so that the script remains valid Python code) divide a file
 into separate blocks, and the demo can be run one block at a time, with
 IPython printing (with syntax highlighting) the block before executing
 it, and returning to the interactive prompt after each block. The
 interactive namespace is updated after each block is run with the
 contents of the demo's namespace.
 
 This allows you to show a piece of code, run it and then execute
 interactively commands based on the variables just created. Once you
 want to continue, you simply execute the next block of the demo. The
 following listing shows the markup necessary for dividing a script into
 sections for execution as a demo:
 
 .. literalinclude:: ../../examples/lib/example-demo.py
     :language: python
 
 In order to run a file as a demo, you must first make a Demo object out
 of it. If the file is named myscript.py, the following code will make a
 demo::
 
     from IPython.lib.demo import Demo
 
     mydemo = Demo('myscript.py')
 
 This creates the mydemo object, whose blocks you run one at a time by
 simply calling the object with no arguments. If you have autocall active
 in IPython (the default), all you need to do is type::
 
     mydemo 
 
 and IPython will call it, executing each block. Demo objects can be
 restarted, you can move forward or back skipping blocks, re-execute the
 last block, etc. Simply use the Tab key on a demo object to see its
 methods, and call '?' on them to see their docstrings for more usage
 details. In addition, the demo module itself contains a comprehensive
 docstring, which you can access via::
 
     from IPython.lib import demo
 
     demo?
 
 Limitations: It is important to note that these demos are limited to
 fairly simple uses. In particular, you can not put division marks in
 indented code (loops, if statements, function definitions, etc.)
 Supporting something like this would basically require tracking the
 internal execution state of the Python interpreter, so only top-level
 divisions are allowed. If you want to be able to open an IPython
 instance at an arbitrary point in a program, you can use IPython's
 embedding facilities, see :func:`IPython.embed` for details.
 

docs/source/parallel/parallel_intro.txt

@@ -1,263 +1,263 b''
 .. _parallel_overview:
 
 ============================
 Overview and getting started
 ============================
 
 Introduction
 ============
 
 This section gives an overview of IPython's sophisticated and powerful
 architecture for parallel and distributed computing. This architecture
 abstracts out parallelism in a very general way, which enables IPython to
 support many different styles of parallelism including:
 
 * Single program, multiple data (SPMD) parallelism.
 * Multiple program, multiple data (MPMD) parallelism.
 * Message passing using MPI.
 * Task farming.
 * Data parallel.
 * Combinations of these approaches.
 * Custom user defined approaches.
 
 Most importantly, IPython enables all types of parallel applications to
 be developed, executed, debugged and monitored *interactively*. Hence,
 the ``I`` in IPython.  The following are some example usage cases for IPython:
 
 * Quickly parallelize algorithms that are embarrassingly parallel
   using a number of simple approaches.  Many simple things can be
   parallelized interactively in one or two lines of code.
 
 * Steer traditional MPI applications on a supercomputer from an
   IPython session on your laptop.
 
 * Analyze and visualize large datasets (that could be remote and/or
   distributed) interactively using IPython and tools like
   matplotlib/TVTK.
 
 * Develop, test and debug new parallel algorithms 
   (that may use MPI) interactively.
 
 * Tie together multiple MPI jobs running on different systems into
   one giant distributed and parallel system.
 
 * Start a parallel job on your cluster and then have a remote 	
   collaborator connect to it and pull back data into their 
   local IPython session for plotting and analysis.
 
 * Run a set of tasks on a set of CPUs using dynamic load balancing.
 
 .. tip::
 
    At the SciPy 2011 conference in Austin, Min Ragan-Kelley presented a
    complete 4-hour tutorial on the use of these features, and all the materials
    for the tutorial are now `available online`__.  That tutorial provides an
    excellent, hands-on oriented complement to the reference documentation
    presented here.
 
 .. __: http://minrk.github.com/scipy-tutorial-2011
 
 Architecture overview
 =====================
 
 The IPython architecture consists of four components:
 
 * The IPython engine.
 * The IPython hub.
 * The IPython schedulers.
 * The controller client.
 
 These components live in the :mod:`IPython.parallel` package and are
 installed with IPython.  They do, however, have additional dependencies
 that must be installed.  For more information, see our
 :ref:`installation documentation <install_index>`.
 
 .. TODO: include zmq in install_index
 
 IPython engine
 ---------------
 
 The IPython engine is a Python instance that takes Python commands over a
 network connection. Eventually, the IPython engine will be a full IPython
 interpreter, but for now, it is a regular Python interpreter. The engine
 can also handle incoming and outgoing Python objects sent over a network
 connection.  When multiple engines are started, parallel and distributed
 computing becomes possible. An important feature of an IPython engine is
 that it blocks while user code is being executed. Read on for how the
 IPython controller solves this problem to expose a clean asynchronous API
 to the user.
 
 IPython controller
 ------------------
 
 The IPython controller processes provide an interface for working with a set of engines.
 At a general level, the controller is a collection of processes to which IPython engines
 and clients can connect. The controller is composed of a :class:`Hub` and a collection of
 :class:`Schedulers`. These Schedulers are typically run in separate processes but on the
 same machine as the Hub, but can be run anywhere from local threads or on remote machines.
 
 The controller also provides a single point of contact for users who wish to
 utilize the engines connected to the controller. There are different ways of
 working with a controller. In IPython, all of these models are implemented via
 the client's :meth:`.View.apply` method, with various arguments, or
 constructing :class:`.View` objects to represent subsets of engines. The two
 primary models for interacting with engines are:
 
 * A **Direct** interface, where engines are addressed explicitly.
 * A **LoadBalanced** interface, where the Scheduler is trusted with assigning work to
   appropriate engines.
 
 Advanced users can readily extend the View models to enable other
 styles of parallelism. 
 
 .. note:: 
 
     A single controller and set of engines can be used with multiple models
     simultaneously. This opens the door for lots of interesting things.
 
 
 The Hub
 *******
 
 The center of an IPython cluster is the Hub. This is the process that keeps
 track of engine connections, schedulers, clients, as well as all task requests and
 results. The primary role of the Hub is to facilitate queries of the cluster state, and
 minimize the necessary information required to establish the many connections involved in
 connecting new clients and engines.
 
 
 Schedulers
 **********
 
 All actions that can be performed on the engine go through a Scheduler. While the engines
 themselves block when user code is run, the schedulers hide that from the user to provide
 a fully asynchronous interface to a set of engines.
 
 
 IPython client and views
 ------------------------
 
 There is one primary object, the :class:`~.parallel.Client`, for connecting to a cluster.
 For each execution model, there is a corresponding :class:`~.parallel.View`. These views
 allow users to interact with a set of engines through the interface. Here are the two default
 views:
 
 * The :class:`DirectView` class for explicit addressing.
 * The :class:`LoadBalancedView` class for destination-agnostic scheduling.
 
 Security
 --------
 
 IPython uses ZeroMQ for networking, which has provided many advantages, but
 one of the setbacks is its utter lack of security [ZeroMQ]_. By default, no IPython
 connections are encrypted, but open ports only listen on localhost. The only
 source of security for IPython is via ssh-tunnel. IPython supports both shell
 (`openssh`) and `paramiko` based tunnels for connections.  There is a key necessary
 to submit requests, but due to the lack of encryption, it does not provide
 significant security if loopback traffic is compromised.
 
 In our architecture, the controller is the only process that listens on
 network ports, and is thus the main point of vulnerability. The standard model
 for secure connections is to designate that the controller listen on
 localhost, and use ssh-tunnels to connect clients and/or
 engines.
 
 To connect and authenticate to the controller an engine or client needs
 some information that the controller has stored in a JSON file.
 Thus, the JSON files need to be copied to a location where
 the clients and engines can find them. Typically, this is the
 :file:`~/.ipython/profile_default/security` directory on the host where the 
 client/engine is running (which could be a different host than the controller). 
 Once the JSON files are copied over, everything should work fine.
 
 Currently, there are two JSON files that the controller creates:
 
 ipcontroller-engine.json
     This JSON file has the information necessary for an engine to connect
     to a controller.
 
 ipcontroller-client.json
     The client's connection information.  This may not differ from the engine's,
     but since the controller may listen on different ports for clients and
     engines, it is stored separately.
 
 More details of how these JSON files are used are given below.
 
 A detailed description of the security model and its implementation in IPython
 can be found :ref:`here <parallelsecurity>`.
 
 .. warning::
 
     Even at its most secure, the Controller listens on ports on localhost, and
     every time you make a tunnel, you open a localhost port on the connecting
     machine that points to the Controller. If localhost on the Controller's
     machine, or the machine of any client or engine, is untrusted, then your
     Controller is insecure. There is no way around this with ZeroMQ.
 
 
 
 Getting Started
 ===============
 
 To use IPython for parallel computing, you need to start one instance of 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 your
 localhost, just do::
 
-    $ ipcluster start --n=4
+    $ ipcluster start -n 4
 
 More details about starting the IPython controller and engines can be found
 :ref:`here <parallel_process>`
 
 Once you have started the IPython controller and one or more engines, you
 are ready to use the engines to do something useful. To make sure
 everything is working correctly, try the following commands:
 
 .. sourcecode:: ipython
 
 	In [1]: from IPython.parallel import Client
 	
 	In [2]: c = Client()
 	
 	In [4]: c.ids
 	Out[4]: set([0, 1, 2, 3])
 	
 	In [5]: c[:].apply_sync(lambda : "Hello, World")
 	Out[5]: [ 'Hello, World', 'Hello, World', 'Hello, World', 'Hello, World' ]
 
 
 When a client is created with no arguments, the client tries to find the corresponding JSON file
 in the local `~/.ipython/profile_default/security` directory. Or if you specified a profile,
 you can use that with the Client.  This should cover most cases:
 
 .. sourcecode:: ipython
 
     In [2]: c = Client(profile='myprofile')
 
 If you have put the JSON file in a different location or it has a different name, create the
 client like this:
 
 .. sourcecode:: ipython
 
     In [2]: c = Client('/path/to/my/ipcontroller-client.json')
 
 Remember, a client needs to be able to see the Hub's ports to connect. So if they are on a
 different machine, you may need to use an ssh server to tunnel access to that machine,
 then you would connect to it with:
 
 .. sourcecode:: ipython
 
     In [2]: c = Client(sshserver='myhub.example.com')
 
 Where 'myhub.example.com' is the url or IP address of the machine on
 which the Hub process is running (or another machine that has direct access to the Hub's ports).
 
 The SSH server may already be specified in ipcontroller-client.json, if the controller was
 instructed at its launch time.
 
 You are now ready to learn more about the :ref:`Direct
 <parallel_multiengine>` and :ref:`LoadBalanced <parallel_task>` interfaces to the
 controller.
 
 .. [ZeroMQ] ZeroMQ.  http://www.zeromq.org

docs/source/parallel/parallel_mpi.txt

@@ -1,151 +1,151 b''
 .. _parallelmpi:
 
 =======================
 Using MPI with IPython
 =======================
 
 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.
 
 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.
 
 Additional installation requirements
 ====================================
 
 If you want to use MPI with IPython, you will need to install:
 
 * A standard MPI implementation such as OpenMPI [OpenMPI]_ or MPICH.
 * The mpi4py [mpi4py]_ package.
 
 .. note::
 
     The mpi4py package is not a strict requirement. However, you need to
     have *some* way of calling MPI from Python. You also need some way of
     making sure that :func:`MPI_Init` is called when the IPython engines start
     up. There are a number of ways of doing this and a good number of
     associated subtleties. We highly recommend just using mpi4py as it
     takes care of most of these problems. If you want to do something
     different, let us know and we can help you get started.
 
 Starting the engines with MPI enabled
 =====================================
 
 To use code that calls MPI, there are typically two things that MPI requires.
 
 1. The process that wants to call MPI must be started using
    :command:`mpiexec` or a batch system (like PBS) that has MPI support.
 2. Once the process starts, it must call :func:`MPI_Init`.
 
 There are a couple of ways that you can start the IPython engines and get
 these things to happen.
 
 Automatic starting using :command:`mpiexec` and :command:`ipcluster`
 --------------------------------------------------------------------
 
 The easiest approach is to use the `MPIExec` Launchers in :command:`ipcluster`,
 which will first start a controller and then a set of engines using
 :command:`mpiexec`::
 
-    $ ipcluster start --n=4 --elauncher=MPIExecEngineSetLauncher
+    $ ipcluster start -n 4 --elauncher=MPIExecEngineSetLauncher
 
 This approach is best as interrupting :command:`ipcluster` will automatically
 stop and clean up the controller and engines.
 
 Manual starting using :command:`mpiexec`
 ----------------------------------------
 
 If you want to start the IPython engines using the :command:`mpiexec`, just
 do::
 
     $ mpiexec n=4 ipengine --mpi=mpi4py
 
 This requires that you already have a controller running and that the FURL
 files for the engines are in place. We also have built in support for
 PyTrilinos [PyTrilinos]_, which can be used (assuming is installed) by
 starting the engines with::
 
     $ mpiexec n=4 ipengine --mpi=pytrilinos
 
 Automatic starting using PBS and :command:`ipcluster`
 ------------------------------------------------------
 
 The :command:`ipcluster` command also has built-in integration with PBS. For
 more information on this approach, see our documentation on :ref:`ipcluster
 <parallel_process>`.
 
 Actually using MPI
 ==================
 
 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]_ version 1.1.0 or later.
 
 First, lets define a simply function that uses MPI to calculate the sum of a
 distributed array. Save the following text in a file called :file:`psum.py`:
 
 .. sourcecode:: python
 
     from mpi4py import MPI
     import numpy as np
 
     def psum(a):
         s = np.sum(a)
         rcvBuf = np.array(0.0,'d')
         MPI.COMM_WORLD.Allreduce([s, MPI.DOUBLE],
             [rcvBuf, MPI.DOUBLE],
             op=MPI.SUM)
         return rcvBuf
 
 Now, start an IPython cluster::
 
-    $ ipcluster start --profile=mpi --n=4
+    $ ipcluster start --profile=mpi -n 4
     
 .. note::
 
     It is assumed here that the mpi profile has been set up, as described :ref:`here
     <parallel_process>`.
 
 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:
 
 .. sourcecode:: ipython
 
     In [1]: from IPython.parallel import Client
     
     In [2]: %load_ext parallel_magic
     
     In [3]: c = Client(profile='mpi')
     
     In [4]: view = c[:]
     
     In [5]: view.activate()
     
     # run the contents of the file on each engine:
     In [6]: view.run('psum.py')
     
     In [6]: px a = np.random.rand(100)
     Parallel execution on engines: [0,1,2,3]
     
     In [8]: px s = psum(a)
     Parallel execution on engines: [0,1,2,3]
     
     In [9]: view['s']
     Out[9]: [187.451545803,187.451545803,187.451545803,187.451545803]
 
 Any Python code that makes calls to MPI can be used in this manner, including
 compiled C, C++ and Fortran libraries that have been exposed to Python.
 
 .. [MPI] Message Passing Interface.  http://www-unix.mcs.anl.gov/mpi/
 .. [mpi4py] MPI for Python. mpi4py: http://mpi4py.scipy.org/
 .. [OpenMPI] Open MPI. http://www.open-mpi.org/
 .. [PyTrilinos] PyTrilinos. http://trilinos.sandia.gov/packages/pytrilinos/
 
 

docs/source/parallel/parallel_multiengine.txt

@@ -1,847 +1,847 b''
 .. _parallel_multiengine:
 
 ==========================
 IPython's Direct interface
 ==========================
 
 The direct, or multiengine, interface represents one possible way of working with a set of
 IPython engines. The basic idea behind the multiengine interface is that the
 capabilities of each engine are directly and explicitly exposed to the user.
 Thus, in the multiengine interface, each engine is given an id that is used to
 identify the engine and give it work to do. This interface is very intuitive
 and is designed with interactive usage in mind, and is the best place for
 new users of IPython to begin.
 
 Starting the IPython controller and engines
 ===========================================
 
 To follow along with this tutorial, you will need to start the IPython
 controller and four IPython engines. The simplest way of doing this is to use
 the :command:`ipcluster` command::
 
-    $ ipcluster start --n=4
+    $ ipcluster start -n 4
     
 For more detailed information about starting the controller and engines, see
 our :ref:`introduction <parallel_overview>` to using IPython for parallel computing.
 
 Creating a ``Client`` instance
 ==============================
 
 The first step is to import the IPython :mod:`IPython.parallel`
 module and then create a :class:`.Client` instance:
 
 .. sourcecode:: ipython
 
     In [1]: from IPython.parallel import Client
     
     In [2]: rc = Client()
 
 This form assumes that the default connection information (stored in
 :file:`ipcontroller-client.json` found in :file:`IPYTHON_DIR/profile_default/security`) is
 accurate. If the controller was started on a remote machine, you must copy that connection
 file to the client machine, or enter its contents as arguments to the Client constructor:
 
 .. sourcecode:: ipython
 
     # If you have copied the json connector file from the controller:
     In [2]: rc = Client('/path/to/ipcontroller-client.json')
     # or to connect with a specific profile you have set up:
     In [3]: rc = Client(profile='mpi')
     
 
 To make sure there are engines connected to the controller, users can get a list
 of engine ids:
 
 .. sourcecode:: ipython
 
     In [3]: rc.ids
     Out[3]: [0, 1, 2, 3]
 
 Here we see that there are four engines ready to do work for us.
 
 For direct execution, we will make use of a :class:`DirectView` object, which can be
 constructed via list-access to the client:
 
 .. sourcecode:: ipython
 
     In [4]: dview = rc[:] # use all engines
 
 .. seealso::
 
     For more information, see the in-depth explanation of :ref:`Views <parallel_details>`.
 
 
 Quick and easy parallelism
 ==========================
 
 In many cases, you simply want to apply a Python function to a sequence of
 objects, but *in parallel*. The client interface provides a simple way
 of accomplishing this: using the DirectView's :meth:`~DirectView.map` method.
 
 Parallel map
 ------------
 
 Python's builtin :func:`map` functions allows a function to be applied to a
 sequence element-by-element. This type of code is typically trivial to
 parallelize. In fact, since IPython's interface is all about functions anyway,
 you can just use the builtin :func:`map` with a :class:`RemoteFunction`, or a 
 DirectView's :meth:`map` method:
 
 .. sourcecode:: ipython
 
     In [62]: serial_result = map(lambda x:x**10, range(32))
     
     In [63]: parallel_result = dview.map_sync(lambda x: x**10, range(32))
 
     In [67]: serial_result==parallel_result
     Out[67]: True
 
 
 .. note::
 
     The :class:`DirectView`'s version of :meth:`map` does
     not do dynamic load balancing. For a load balanced version, use a
     :class:`LoadBalancedView`.
 
 .. seealso::
     
     :meth:`map` is implemented via :class:`ParallelFunction`.
 
 Remote function decorators
 --------------------------
 
 Remote functions are just like normal functions, but when they are called,
 they execute on one or more engines, rather than locally. IPython provides
 two decorators:
 
 .. sourcecode:: ipython
 
     In [10]: @dview.remote(block=True)
         ...: def getpid():
         ...:     import os
         ...:     return os.getpid()
         ...: 
 
     In [11]: getpid()
     Out[11]: [12345, 12346, 12347, 12348]
 
 The ``@parallel`` decorator creates parallel functions, that break up an element-wise
 operations and distribute them, reconstructing the result.
 
 .. sourcecode:: ipython
 
     In [12]: import numpy as np
     
     In [13]: A = np.random.random((64,48))
     
     In [14]: @dview.parallel(block=True)
         ...: def pmul(A,B):
         ...:     return A*B
     
     In [15]: C_local = A*A
     
     In [16]: C_remote = pmul(A,A)
     
     In [17]: (C_local == C_remote).all()
     Out[17]: True
 
 .. seealso::
 
     See the docstrings for the :func:`parallel` and :func:`remote` decorators for
     options.
 
 Calling Python functions
 ========================
 
 The most basic type of operation that can be performed on the engines is to
 execute Python code or call Python functions. Executing Python code can be
 done in blocking or non-blocking mode (non-blocking is default) using the
 :meth:`.View.execute` method, and calling functions can be done via the
 :meth:`.View.apply` method.
 
 apply
 -----
 
 The main method for doing remote execution (in fact, all methods that
 communicate with the engines are built on top of it), is :meth:`View.apply`.
 
 We strive to provide the cleanest interface we can, so `apply` has the following
 signature:
 
 .. sourcecode:: python
 
     view.apply(f, *args, **kwargs)
 
 There are various ways to call functions with IPython, and these flags are set as
 attributes of the View.  The ``DirectView`` has just two of these flags:
 
 dv.block : bool
     whether to wait for the result, or return an :class:`AsyncResult` object
     immediately
 dv.track : bool
     whether to instruct pyzmq to track when 
     This is primarily useful for non-copying sends of numpy arrays that you plan to
     edit in-place.  You need to know when it becomes safe to edit the buffer
     without corrupting the message.
 
 
 Creating a view is simple: index-access on a client creates a :class:`.DirectView`.
 
 .. sourcecode:: ipython
 
     In [4]: view = rc[1:3]
     Out[4]: <DirectView [1, 2]>
 
     In [5]: view.apply<tab>
     view.apply  view.apply_async  view.apply_sync
 
 For convenience, you can set block temporarily for a single call with the extra sync/async methods.
 
 Blocking execution
 ------------------
 
 In blocking mode, the :class:`.DirectView` object (called ``dview`` in
 these examples) submits the command to the controller, which places the
 command in the engines' queues for execution. The :meth:`apply` call then
 blocks until the engines are done executing the command:
 
 .. sourcecode:: ipython
 
     In [2]: dview = rc[:] # A DirectView of all engines
     In [3]: dview.block=True
     In [4]: dview['a'] = 5
 
     In [5]: dview['b'] = 10
 
     In [6]: dview.apply(lambda x: a+b+x, 27)
     Out[6]: [42, 42, 42, 42]
 
 You can also select blocking execution on a call-by-call basis with the :meth:`apply_sync`
 method:
 
     In [7]: dview.block=False
 
     In [8]: dview.apply_sync(lambda x: a+b+x, 27)
     Out[8]: [42, 42, 42, 42]
 
 Python commands can be executed as strings on specific engines by using a View's ``execute``
 method:
 
 .. sourcecode:: ipython
 
     In [6]: rc[::2].execute('c=a+b')
 
     In [7]: rc[1::2].execute('c=a-b')
 
     In [8]: dview['c'] # shorthand for dview.pull('c', block=True)
     Out[8]: [15, -5, 15, -5]
 
 
 Non-blocking execution
 ----------------------
 
 In non-blocking mode, :meth:`apply` submits the command to be executed and
 then returns a :class:`AsyncResult` object immediately. The
 :class:`AsyncResult` object gives you a way of getting a result at a later
 time through its :meth:`get` method.
 
 .. Note::
 
     The :class:`AsyncResult` object provides a superset of the interface in 
     :py:class:`multiprocessing.pool.AsyncResult`.  See the 
     `official Python documentation <http://docs.python.org/library/multiprocessing#multiprocessing.pool.AsyncResult>`_
     for more.
 
 
 This allows you to quickly submit long running commands without blocking your
 local Python/IPython session:
 
 .. sourcecode:: ipython
     
     # define our function
     In [6]: def wait(t):
        ...:     import time
        ...:     tic = time.time()
        ...:     time.sleep(t)
        ...:     return time.time()-tic
     
     # In non-blocking mode
     In [7]: ar = dview.apply_async(wait, 2)
 
     # Now block for the result
     In [8]: ar.get()
     Out[8]: [2.0006198883056641, 1.9997570514678955, 1.9996809959411621, 2.0003249645233154]
 
     # Again in non-blocking mode
     In [9]: ar = dview.apply_async(wait, 10)
 
     # Poll to see if the result is ready
     In [10]: ar.ready()
     Out[10]: False
     
     # ask for the result, but wait a maximum of 1 second:
     In [45]: ar.get(1)
     ---------------------------------------------------------------------------
     TimeoutError                              Traceback (most recent call last)
     /home/you/<ipython-input-45-7cd858bbb8e0> in <module>()
     ----> 1 ar.get(1)
 
     /path/to/site-packages/IPython/parallel/asyncresult.pyc in get(self, timeout)
          62                 raise self._exception
          63         else:
     ---> 64             raise error.TimeoutError("Result not ready.")
          65 
          66     def ready(self):
 
     TimeoutError: Result not ready.
 
 .. Note::
 
     Note the import inside the function. This is a common model, to ensure
     that the appropriate modules are imported where the task is run. You can
     also manually import modules into the engine(s) namespace(s) via 
     :meth:`view.execute('import numpy')`.
 
 Often, it is desirable to wait until a set of :class:`AsyncResult` objects
 are done. For this, there is a the method :meth:`wait`. This method takes a
 tuple of :class:`AsyncResult` objects (or `msg_ids` or indices to the client's History),
 and blocks until all of the associated results are ready:
 
 .. sourcecode:: ipython
 
     In [72]: dview.block=False
 
     # A trivial list of AsyncResults objects
     In [73]: pr_list = [dview.apply_async(wait, 3) for i in range(10)]
 
     # Wait until all of them are done
     In [74]: dview.wait(pr_list)
 
     # Then, their results are ready using get() or the `.r` attribute
     In [75]: pr_list[0].get()
     Out[75]: [2.9982571601867676, 2.9982588291168213, 2.9987530708312988, 2.9990990161895752]
 
 
 
 The ``block`` and ``targets`` keyword arguments and attributes
 --------------------------------------------------------------
 
 Most DirectView methods (excluding :meth:`apply` and :meth:`map`) accept ``block`` and
 ``targets`` as keyword arguments. As we have seen above, these keyword arguments control the
 blocking mode and which engines the command is applied to. The :class:`View` class also has
 :attr:`block` and :attr:`targets` attributes that control the default behavior when the keyword
 arguments are not provided. Thus the following logic is used for :attr:`block` and :attr:`targets`:
 
 * If no keyword argument is provided, the instance attributes are used.
 * Keyword argument, if provided override the instance attributes for
   the duration of a single call.
   
 The following examples demonstrate how to use the instance attributes:
 
 .. sourcecode:: ipython
 
     In [16]: dview.targets = [0,2]
     
     In [17]: dview.block = False
 
     In [18]: ar = dview.apply(lambda : 10)
 
     In [19]: ar.get()
     Out[19]: [10, 10]
 
     In [16]: dview.targets = v.client.ids # all engines (4)
     
     In [21]: dview.block = True
 
     In [22]: dview.apply(lambda : 42)
     Out[22]: [42, 42, 42, 42]
 
 The :attr:`block` and :attr:`targets` instance attributes of the
 :class:`.DirectView` also determine the behavior of the parallel magic commands.
 
 Parallel magic commands
 -----------------------
 
 .. warning::
 
     The magics have not been changed to work with the zeromq system. The
     magics do work, but *do not* print stdin/out like they used to in IPython.kernel.
 
 We provide a few IPython magic commands (``%px``, ``%autopx`` and ``%result``)
 that make it more pleasant to execute Python commands on the engines
 interactively. These are simply shortcuts to :meth:`execute` and
 :meth:`get_result` of the :class:`DirectView`. The ``%px`` magic executes a single 
 Python command on the engines specified by the :attr:`targets` attribute of the
 :class:`DirectView` instance:
 
 .. sourcecode:: ipython
 
     # load the parallel magic extension:
     In [21]: %load_ext parallelmagic
     
     # Create a DirectView for all targets
     In [22]: dv = rc[:]
     
     # Make this DirectView active for parallel magic commands
     In [23]: dv.activate()
 
     In [24]: dv.block=True
 
     In [25]: import numpy
 
     In [26]: %px import numpy
     Parallel execution on engines: [0, 1, 2, 3]
 
     In [27]: %px a = numpy.random.rand(2,2)
     Parallel execution on engines: [0, 1, 2, 3]
 
     In [28]: %px ev = numpy.linalg.eigvals(a)
     Parallel execution on engines: [0, 1, 2, 3]
 
     In [28]: dv['ev']
     Out[28]: [ array([ 1.09522024, -0.09645227]), 
                array([ 1.21435496, -0.35546712]), 
                array([ 0.72180653,  0.07133042]), 
                array([  1.46384341e+00,   1.04353244e-04])
              ]
 
 The ``%result`` magic gets the most recent result, or takes an argument
 specifying the index of the result to be requested. It is simply a shortcut to the
 :meth:`get_result` method:
 
 .. sourcecode:: ipython
     
     In [29]: dv.apply_async(lambda : ev)
     
     In [30]: %result
     Out[30]: [ [ 1.28167017  0.14197338],
                 [-0.14093616  1.27877273],
                 [-0.37023573  1.06779409],
                 [ 0.83664764 -0.25602658] ]
 
 The ``%autopx`` magic switches to a mode where everything you type is executed
 on the engines given by the :attr:`targets` attribute:
 
 .. sourcecode:: ipython
 
     In [30]: dv.block=False
 
     In [31]: %autopx
     Auto Parallel Enabled
     Type %autopx to disable
 
     In [32]: max_evals = []
     <IPython.parallel.AsyncResult object at 0x17b8a70>
 
     In [33]: for i in range(100):
        ....:     a = numpy.random.rand(10,10)
        ....:     a = a+a.transpose()
        ....:     evals = numpy.linalg.eigvals(a)
        ....:     max_evals.append(evals[0].real)
        ....:
        ....:
     <IPython.parallel.AsyncResult object at 0x17af8f0>
 
     In [34]: %autopx
     Auto Parallel Disabled
 
     In [35]: dv.block=True
 
     In [36]: px ans= "Average max eigenvalue is: %f"%(sum(max_evals)/len(max_evals))
     Parallel execution on engines: [0, 1, 2, 3]
     
     In [37]: dv['ans']
     Out[37]: [ 'Average max eigenvalue is:  10.1387247332',
                'Average max eigenvalue is:  10.2076902286',
                'Average max eigenvalue is:  10.1891484655',
                'Average max eigenvalue is:  10.1158837784',]
 
 
 Moving Python objects around
 ============================
 
 In addition to calling functions and executing code on engines, you can
 transfer Python objects to and from your IPython session and the engines. In
 IPython, these operations are called :meth:`push` (sending an object to the
 engines) and :meth:`pull` (getting an object from the engines).
 
 Basic push and pull
 -------------------
 
 Here are some examples of how you use :meth:`push` and :meth:`pull`:
 
 .. sourcecode:: ipython
 
     In [38]: dview.push(dict(a=1.03234,b=3453))
     Out[38]: [None,None,None,None]
 
     In [39]: dview.pull('a')
     Out[39]: [ 1.03234, 1.03234, 1.03234, 1.03234]
 
     In [40]: dview.pull('b', targets=0)
     Out[40]: 3453
 
     In [41]: dview.pull(('a','b'))
     Out[41]: [ [1.03234, 3453], [1.03234, 3453], [1.03234, 3453], [1.03234, 3453] ]
     
     In [43]: dview.push(dict(c='speed'))
     Out[43]: [None,None,None,None]
 
 In non-blocking mode :meth:`push` and :meth:`pull` also return
 :class:`AsyncResult` objects:
 
 .. sourcecode:: ipython
 
     In [48]: ar = dview.pull('a', block=False)
 
     In [49]: ar.get()
     Out[49]: [1.03234, 1.03234, 1.03234, 1.03234]
 
 
 Dictionary interface
 --------------------
 
 Since a Python namespace is just a :class:`dict`, :class:`DirectView` objects provide
 dictionary-style access by key and methods such as :meth:`get` and
 :meth:`update` for convenience. This make the remote namespaces of the engines
 appear as a local dictionary. Underneath, these methods call :meth:`apply`:
 
 .. sourcecode:: ipython
 
     In [51]: dview['a']=['foo','bar']
 
     In [52]: dview['a']
     Out[52]: [ ['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar'], ['foo', 'bar'] ]
 
 Scatter and gather
 ------------------
 
 Sometimes it is useful to partition a sequence and push the partitions to
 different engines. In MPI language, this is know as scatter/gather and we
 follow that terminology. However, it is important to remember that in
 IPython's :class:`Client` class, :meth:`scatter` is from the
 interactive IPython session to the engines and :meth:`gather` is from the
 engines back to the interactive IPython session. For scatter/gather operations
 between engines, MPI should be used:
 
 .. sourcecode:: ipython
 
     In [58]: dview.scatter('a',range(16))
     Out[58]: [None,None,None,None]
 
     In [59]: dview['a']
     Out[59]: [ [0, 1, 2, 3], [4, 5, 6, 7], [8, 9, 10, 11], [12, 13, 14, 15] ]
 
     In [60]: dview.gather('a')
     Out[60]: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]
 
 Other things to look at
 =======================
 
 How to do parallel list comprehensions
 --------------------------------------
 
 In many cases list comprehensions are nicer than using the map function. While
 we don't have fully parallel list comprehensions, it is simple to get the
 basic effect using :meth:`scatter` and :meth:`gather`:
 
 .. sourcecode:: ipython
 
     In [66]: dview.scatter('x',range(64))
 
     In [67]: %px y = [i**10 for i in x]
     Parallel execution on engines: [0, 1, 2, 3]
     Out[67]:
 
     In [68]: y = dview.gather('y')
 
     In [69]: print y
     [0, 1, 1024, 59049, 1048576, 9765625, 60466176, 282475249, 1073741824,...]
 
 Remote imports
 --------------
 
 Sometimes you will want to import packages both in your interactive session
 and on your remote engines.  This can be done with the :class:`ContextManager`
 created by a DirectView's :meth:`sync_imports` method:
 
 .. sourcecode:: ipython
 
     In [69]: with dview.sync_imports():
         ...:     import numpy
     importing numpy on engine(s)
 
 Any imports made inside the block will also be performed on the view's engines.
 sync_imports also takes a `local` boolean flag that defaults to True, which specifies
 whether the local imports should also be performed.  However, support for `local=False`
 has not been implemented, so only packages that can be imported locally will work
 this way.
 
 You can also specify imports via the ``@require`` decorator.  This is a decorator
 designed for use in Dependencies, but can be used to handle remote imports as well.
 Modules or module names passed to ``@require`` will be imported before the decorated
 function is called.  If they cannot be imported, the decorated function will never
 execution, and will fail with an UnmetDependencyError.
 
 .. sourcecode:: ipython
 
     In [69]: from IPython.parallel import require
 
     In [70]: @require('re'):
         ...: def findall(pat, x):
         ...:     # re is guaranteed to be available
         ...:     return re.findall(pat, x)
           
     # you can also pass modules themselves, that you already have locally:
     In [71]: @require(time):
         ...: def wait(t):
         ...:     time.sleep(t)
         ...:     return t
 
 .. _parallel_exceptions:
 
 Parallel exceptions
 -------------------
 
 In the multiengine interface, parallel commands can raise Python exceptions,
 just like serial commands. But, it is a little subtle, because a single
 parallel command can actually raise multiple exceptions (one for each engine
 the command was run on). To express this idea, we have a
 :exc:`CompositeError` exception class that will be raised in most cases. The
 :exc:`CompositeError` class is a special type of exception that wraps one or
 more other types of exceptions. Here is how it works:
 
 .. sourcecode:: ipython
 
     In [76]: dview.block=True
 
     In [77]: dview.execute('1/0')
     ---------------------------------------------------------------------------
     CompositeError                            Traceback (most recent call last)
     /home/user/<ipython-input-10-5d56b303a66c> in <module>()
     ----> 1 dview.execute('1/0')
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in execute(self, code, targets, block)
         591                 default: self.block
         592         """
     --> 593         return self._really_apply(util._execute, args=(code,), block=block, targets=targets)
         594 
         595     def run(self, filename, targets=None, block=None):
 
     /home/user/<string> in _really_apply(self, f, args, kwargs, targets, block, track)
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in sync_results(f, self, *args, **kwargs)
          55 def sync_results(f, self, *args, **kwargs):
          56     """sync relevant results from self.client to our results attribute."""
     ---> 57     ret = f(self, *args, **kwargs)
          58     delta = self.outstanding.difference(self.client.outstanding)
          59     completed = self.outstanding.intersection(delta)
 
     /home/user/<string> in _really_apply(self, f, args, kwargs, targets, block, track)
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in save_ids(f, self, *args, **kwargs)
          44     n_previous = len(self.client.history)
          45     try:
     ---> 46         ret = f(self, *args, **kwargs)
          47     finally:
          48         nmsgs = len(self.client.history) - n_previous
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in _really_apply(self, f, args, kwargs, targets, block, track)
         529         if block:
         530             try:
     --> 531                 return ar.get()
         532             except KeyboardInterrupt:
         533                 pass
 
     /path/to/site-packages/IPython/parallel/client/asyncresult.pyc in get(self, timeout)
         101                 return self._result
         102             else:
     --> 103                 raise self._exception
         104         else:
         105             raise error.TimeoutError("Result not ready.")
 
     CompositeError: one or more exceptions from call to method: _execute
     [0:apply]: ZeroDivisionError: integer division or modulo by zero
     [1:apply]: ZeroDivisionError: integer division or modulo by zero
     [2:apply]: ZeroDivisionError: integer division or modulo by zero
     [3:apply]: ZeroDivisionError: integer division or modulo by zero
 
 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:
 
 .. sourcecode:: ipython
 
     In [80]: try:
        ....:     dview.execute('1/0')
        ....: except parallel.error.CompositeError, e:
        ....:     e.raise_exception()
        ....:     
        ....:     
     ---------------------------------------------------------------------------
     RemoteError                               Traceback (most recent call last)
     /home/user/<ipython-input-17-8597e7e39858> in <module>()
           2     dview.execute('1/0')
           3 except CompositeError as e:
     ----> 4     e.raise_exception()
 
     /path/to/site-packages/IPython/parallel/error.pyc in raise_exception(self, excid)
         266             raise IndexError("an exception with index %i does not exist"%excid)
         267         else:
     --> 268             raise RemoteError(en, ev, etb, ei)
         269 
         270 
 
     RemoteError: ZeroDivisionError(integer division or modulo by zero)
     Traceback (most recent call last):
       File "/path/to/site-packages/IPython/parallel/engine/streamkernel.py", line 330, in apply_request
         exec code in working,working
       File "<string>", line 1, in <module>
       File "/path/to/site-packages/IPython/parallel/util.py", line 354, in _execute
         exec code in globals()
       File "<string>", line 1, in <module>
     ZeroDivisionError: integer division or modulo by zero
 
 If you are working in IPython, you can simple type ``%debug`` after one of
 these :exc:`CompositeError` exceptions is raised, and inspect the exception
 instance:
 
 .. sourcecode:: ipython
 
     In [81]: dview.execute('1/0')
     ---------------------------------------------------------------------------
     CompositeError                            Traceback (most recent call last)
     /home/user/<ipython-input-10-5d56b303a66c> in <module>()
     ----> 1 dview.execute('1/0')
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in execute(self, code, targets, block)
         591                 default: self.block
         592         """
     --> 593         return self._really_apply(util._execute, args=(code,), block=block, targets=targets)
         594 
         595     def run(self, filename, targets=None, block=None):
 
     /home/user/<string> in _really_apply(self, f, args, kwargs, targets, block, track)
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in sync_results(f, self, *args, **kwargs)
          55 def sync_results(f, self, *args, **kwargs):
          56     """sync relevant results from self.client to our results attribute."""
     ---> 57     ret = f(self, *args, **kwargs)
          58     delta = self.outstanding.difference(self.client.outstanding)
          59     completed = self.outstanding.intersection(delta)
 
     /home/user/<string> in _really_apply(self, f, args, kwargs, targets, block, track)
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in save_ids(f, self, *args, **kwargs)
          44     n_previous = len(self.client.history)
          45     try:
     ---> 46         ret = f(self, *args, **kwargs)
          47     finally:
          48         nmsgs = len(self.client.history) - n_previous
 
     /path/to/site-packages/IPython/parallel/client/view.pyc in _really_apply(self, f, args, kwargs, targets, block, track)
         529         if block:
         530             try:
     --> 531                 return ar.get()
         532             except KeyboardInterrupt:
         533                 pass
 
     /path/to/site-packages/IPython/parallel/client/asyncresult.pyc in get(self, timeout)
         101                 return self._result
         102             else:
     --> 103                 raise self._exception
         104         else:
         105             raise error.TimeoutError("Result not ready.")
 
     CompositeError: one or more exceptions from call to method: _execute
     [0:apply]: ZeroDivisionError: integer division or modulo by zero
     [1:apply]: ZeroDivisionError: integer division or modulo by zero
     [2:apply]: ZeroDivisionError: integer division or modulo by zero
     [3:apply]: ZeroDivisionError: integer division or modulo by zero
     
     In [82]: %debug
     > /path/to/site-packages/IPython/parallel/client/asyncresult.py(103)get()
         102             else:
     --> 103                 raise self._exception
         104         else:
     
     # With the debugger running, self._exception is the exceptions instance.  We can tab complete
     # on it and see the extra methods that are available.
     ipdb> self._exception.<tab>
     e.__class__         e.__getitem__       e.__new__           e.__setstate__      e.args
     e.__delattr__       e.__getslice__      e.__reduce__        e.__str__           e.elist
     e.__dict__          e.__hash__          e.__reduce_ex__     e.__weakref__       e.message
     e.__doc__           e.__init__          e.__repr__          e._get_engine_str   e.print_tracebacks
     e.__getattribute__  e.__module__        e.__setattr__       e._get_traceback    e.raise_exception
     ipdb> self._exception.print_tracebacks()
     [0:apply]: 
     Traceback (most recent call last):
       File "/path/to/site-packages/IPython/parallel/engine/streamkernel.py", line 330, in apply_request
         exec code in working,working
       File "<string>", line 1, in <module>
       File "/path/to/site-packages/IPython/parallel/util.py", line 354, in _execute
         exec code in globals()
       File "<string>", line 1, in <module>
     ZeroDivisionError: integer division or modulo by zero
 
 
     [1:apply]: 
     Traceback (most recent call last):
       File "/path/to/site-packages/IPython/parallel/engine/streamkernel.py", line 330, in apply_request
         exec code in working,working
       File "<string>", line 1, in <module>
       File "/path/to/site-packages/IPython/parallel/util.py", line 354, in _execute
         exec code in globals()
       File "<string>", line 1, in <module>
     ZeroDivisionError: integer division or modulo by zero
 
 
     [2:apply]: 
     Traceback (most recent call last):
       File "/path/to/site-packages/IPython/parallel/engine/streamkernel.py", line 330, in apply_request
         exec code in working,working
       File "<string>", line 1, in <module>
       File "/path/to/site-packages/IPython/parallel/util.py", line 354, in _execute
         exec code in globals()
       File "<string>", line 1, in <module>
     ZeroDivisionError: integer division or modulo by zero
 
 
     [3:apply]: 
     Traceback (most recent call last):
       File "/path/to/site-packages/IPython/parallel/engine/streamkernel.py", line 330, in apply_request
         exec code in working,working
       File "<string>", line 1, in <module>
       File "/path/to/site-packages/IPython/parallel/util.py", line 354, in _execute
         exec code in globals()
       File "<string>", line 1, in <module>
     ZeroDivisionError: integer division or modulo by zero
 
 
 All of this same error handling magic even works in non-blocking mode:
 
 .. sourcecode:: ipython
 
     In [83]: dview.block=False
 
     In [84]: ar = dview.execute('1/0')
 
     In [85]: ar.get()
     ---------------------------------------------------------------------------
     CompositeError                            Traceback (most recent call last)
     /home/user/<ipython-input-21-8531eb3d26fb> in <module>()
     ----> 1 ar.get()
 
     /path/to/site-packages/IPython/parallel/client/asyncresult.pyc in get(self, timeout)
         101                 return self._result
         102             else:
     --> 103                 raise self._exception
         104         else:
         105             raise error.TimeoutError("Result not ready.")
 
     CompositeError: one or more exceptions from call to method: _execute
     [0:apply]: ZeroDivisionError: integer division or modulo by zero
     [1:apply]: ZeroDivisionError: integer division or modulo by zero
     [2:apply]: ZeroDivisionError: integer division or modulo by zero
     [3:apply]: ZeroDivisionError: integer division or modulo by zero
 

docs/source/parallel/parallel_process.txt

@@ -1,691 +1,691 b''
 .. _parallel_process:
 
 ===========================================
 Starting the IPython controller and engines
 ===========================================
 
 To use IPython for parallel computing, you need to start one instance of
 the controller and one or more instances of the engine. The controller
 and each engine can run on different machines or on the same machine.
 Because of this, there are many different possibilities.
 
 Broadly speaking, there are two ways of going about starting a controller and engines:
 
 * In an automated manner using the :command:`ipcluster` command.
 * In a more manual way using the :command:`ipcontroller` and 
   :command:`ipengine` commands.
 
 This document describes both of these methods. We recommend that new users
 start with the :command:`ipcluster` command as it simplifies many common usage
 cases.
 
 General considerations
 ======================
 
 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.
 
 If you are running engines on multiple machines, you will likely need to instruct the
 controller to listen for connections on an external interface. This can be done by specifying
 the ``ip`` argument on the command-line, or the ``HubFactory.ip`` configurable in
 :file:`ipcontroller_config.py`.
 
 If your machines are on a trusted network, you can safely instruct the controller to listen
 on all public interfaces with::
 
     $> ipcontroller --ip=*
 
 Or you can set the same behavior as the default by adding the following line to your :file:`ipcontroller_config.py`:
 
 .. sourcecode:: python
 
     c.HubFactory.ip = '*'
 
 .. note::
 
     Due to the lack of security in ZeroMQ, the controller will only listen for connections on
     localhost by default. If you see Timeout errors on engines or clients, then the first
     thing you should check is the ip address the controller is listening on, and make sure
     that it is visible from the timing out machine.
     
 .. seealso::
 
     Our `notes <parallel_security>`_ on security in the new parallel computing code.
 
 Let's say that you want to start the controller on ``host0`` and engines on
 hosts ``host1``-``hostn``. The following steps are then required:
 
 1. Start the controller on ``host0`` by running :command:`ipcontroller` on
    ``host0``.  The controller must be instructed to listen on an interface visible
    to the engine machines, via the ``ip`` command-line argument or ``HubFactory.ip``
    in :file:`ipcontroller_config.py`.
 2. Move the JSON file (:file:`ipcontroller-engine.json`) created by the 
    controller from ``host0`` to hosts ``host1``-``hostn``.
 3. Start the engines on hosts ``host1``-``hostn`` by running 
    :command:`ipengine`.  This command has to be told where the JSON file 
    (:file:`ipcontroller-engine.json`) is located.
 
 At this point, the controller and engines will be connected. By default, the JSON files
 created by the controller are put into the :file:`~/.ipython/profile_default/security`
 directory. If the engines share a filesystem with the controller, step 2 can be skipped as
 the engines will automatically look at that location.
 
 The final step required to actually use the running controller from a client is to move
 the JSON file :file:`ipcontroller-client.json` from ``host0`` to any host where clients
 will be run. If these file are put into the :file:`~/.ipython/profile_default/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.
 
 Using :command:`ipcluster`
 ===========================
 
 The :command:`ipcluster` command provides a simple way of starting a
 controller and engines in the following situations:
 
 1. When the controller and engines are all run on localhost. This is useful
    for testing or running on a multicore computer.
 2. When engines are started using the :command:`mpiexec` command that comes 
    with most MPI [MPI]_ implementations
 3. When engines are started using the PBS [PBS]_ batch system 
    (or other `qsub` systems, such as SGE).
 4. When the controller is started on localhost and the engines are started on 
    remote nodes using :command:`ssh`.
 5. When engines are started using the Windows HPC Server batch system.
 
 .. note::
 
     Currently :command:`ipcluster` requires that the
     :file:`~/.ipython/profile_<name>/security` directory live on a shared filesystem that is
     seen by both the controller and engines. If you don't have a shared file
     system you will need to use :command:`ipcontroller` and
     :command:`ipengine` directly.
 
 Under the hood, :command:`ipcluster` just uses :command:`ipcontroller`
 and :command:`ipengine` to perform the steps described above.
 
 The simplest way to use ipcluster requires no configuration, and will
 launch a controller and a number of engines on the local machine. For instance, 
 to start one controller and 4 engines on localhost, just do::
 
-    $ ipcluster start --n=4
+    $ ipcluster start -n 4
 
 To see other command line options, do::
 
     $ ipcluster -h
 
 
 Configuring an IPython cluster
 ==============================
 
 Cluster configurations are stored as `profiles`.  You can create a new profile with::
 
     $ ipython profile create --parallel --profile=myprofile
 
 This will create the directory :file:`IPYTHONDIR/profile_myprofile`, and populate it
 with the default configuration files for the three IPython cluster commands. Once
 you edit those files, you can continue to call ipcluster/ipcontroller/ipengine
 with no arguments beyond ``profile=myprofile``, and any configuration will be maintained.
 
 There is no limit to the number of profiles you can have, so you can maintain a profile for each
 of your common use cases. The default profile will be used whenever the
 profile argument is not specified, so edit :file:`IPYTHONDIR/profile_default/*_config.py` to
 represent your most common use case.
 
 The configuration files are loaded with commented-out settings and explanations,
 which should cover most of the available possibilities.
 
 Using various batch systems with :command:`ipcluster`
 -----------------------------------------------------
 
 :command:`ipcluster` has a notion of Launchers that can start controllers
 and engines with various remote execution schemes.  Currently supported
 models include :command:`ssh`, :command:`mpiexec`, PBS-style (Torque, SGE),
 and Windows HPC Server.
 
 .. note::
 
     The Launchers and configuration are designed in such a way that advanced
     users can subclass and configure them to fit their own system that we
     have not yet supported (such as Condor)
 
 Using :command:`ipcluster` in mpiexec/mpirun mode
 --------------------------------------------------
 
 
 The mpiexec/mpirun mode is useful if you:
 
 1. Have MPI installed.
 2. Your systems are configured to use the :command:`mpiexec` or
    :command:`mpirun` commands to start MPI processes.
 
 If these are satisfied, you can create a new profile::
 
     $ ipython profile create --parallel --profile=mpi
 
 and edit the file :file:`IPYTHONDIR/profile_mpi/ipcluster_config.py`.
 
 There, instruct ipcluster to use the MPIExec launchers by adding the lines:
 
 .. sourcecode:: python
 
     c.IPClusterEngines.engine_launcher = 'IPython.parallel.apps.launcher.MPIExecEngineSetLauncher'
 
 If the default MPI configuration is correct, then you can now start your cluster, with::
 
-    $ ipcluster start --n=4 --profile=mpi
+    $ ipcluster start -n 4 --profile=mpi
 
 This does the following:
 
 1. Starts the IPython controller on current host.
 2. Uses :command:`mpiexec` to start 4 engines.
 
 If you have a reason to also start the Controller with mpi, you can specify:
 
 .. sourcecode:: python
 
     c.IPClusterStart.controller_launcher = 'IPython.parallel.apps.launcher.MPIExecControllerLauncher'
 
 .. note::
 
     The Controller *will not* be in the same MPI universe as the engines, so there is not
     much reason to do this unless sysadmins demand it.
 
 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 specify the ``c.MPI.use`` option in :file:`ipengine_config.py`:
 
 .. sourcecode:: python
 
     c.MPI.use = 'mpi4py'
 
 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.
 
 More details on using MPI with IPython can be found :ref:`here <parallelmpi>`.
 
 
 Using :command:`ipcluster` in PBS mode
 ---------------------------------------
 
 The PBS mode uses the Portable Batch System (PBS) to start the engines. 
 
 As usual, we will start by creating a fresh profile::
 
     $ ipython profile create --parallel --profile=pbs
 
 And in :file:`ipcluster_config.py`, we will select the PBS launchers for the controller
 and engines:
 
 .. sourcecode:: python
 
     c.IPClusterStart.controller_launcher = \
         'IPython.parallel.apps.launcher.PBSControllerLauncher'
     c.IPClusterEngines.engine_launcher = \
         'IPython.parallel.apps.launcher.PBSEngineSetLauncher'
 
 .. note::
 
     Note that the configurable is IPClusterEngines for the engine launcher, and
     IPClusterStart for the controller launcher. This is because the start command is a
     subclass of the engine command, adding a controller launcher. Since it is a subclass,
     any configuration made in IPClusterEngines is inherited by IPClusterStart unless it is
     overridden.
 
 IPython does provide simple default batch templates for PBS and SGE, but you may need
 to specify your own. Here is a sample PBS script template:
 
 .. sourcecode:: bash
 
     #PBS -N ipython
     #PBS -j oe
     #PBS -l walltime=00:10:00
     #PBS -l nodes={n/4}:ppn=4
     #PBS -q {queue}
 
     cd $PBS_O_WORKDIR
     export PATH=$HOME/usr/local/bin
     export PYTHONPATH=$HOME/usr/local/lib/python2.7/site-packages
     /usr/local/bin/mpiexec -n {n} ipengine --profile-dir={profile_dir}
 
 There are a few important points about this template:
 
 1. This template will be rendered at runtime using IPython's :class:`EvalFormatter`.
    This is simply a subclass of :class:`string.Formatter` that allows simple expressions
    on keys.
 
 2. Instead of putting in the actual number of engines, use the notation
    ``{n}`` to indicate the number of engines to be started. You can also use
    expressions like ``{n/4}`` in the template to indicate the number of nodes.
    There will always be ``{n}`` and ``{profile_dir}`` variables passed to the formatter.
    These allow the batch system to know how many engines, and where the configuration
    files reside. The same is true for the batch queue, with the template variable
    ``{queue}``.
 
 3. Any options to :command:`ipengine` can be given in the batch script
    template, or in :file:`ipengine_config.py`.
 
 4. Depending on the configuration of you system, you may have to set
    environment variables in the script template.
 
 The controller template should be similar, but simpler:
 
 .. sourcecode:: bash
 
     #PBS -N ipython
     #PBS -j oe
     #PBS -l walltime=00:10:00
     #PBS -l nodes=1:ppn=4
     #PBS -q {queue}
 
     cd $PBS_O_WORKDIR
     export PATH=$HOME/usr/local/bin
     export PYTHONPATH=$HOME/usr/local/lib/python2.7/site-packages
     ipcontroller --profile-dir={profile_dir}
 
 
 Once you have created these scripts, save them with names like
 :file:`pbs.engine.template`. Now you can load them into the :file:`ipcluster_config` with:
 
 .. sourcecode:: python
 
     c.PBSEngineSetLauncher.batch_template_file = "pbs.engine.template"
     
     c.PBSControllerLauncher.batch_template_file = "pbs.controller.template"
         
 
 Alternately, you can just define the templates as strings inside :file:`ipcluster_config`.
 
 Whether you are using your own templates or our defaults, the extra configurables available are
 the number of engines to launch (``{n}``, and the batch system queue to which the jobs are to be
 submitted (``{queue}``)). These are configurables, and can be specified in
 :file:`ipcluster_config`:
 
 .. sourcecode:: python
 
     c.PBSLauncher.queue = 'veryshort.q'
     c.IPClusterEngines.n = 64
 
 Note that assuming you are running PBS on a multi-node cluster, the Controller's default behavior
 of listening only on localhost is likely too restrictive.  In this case, also assuming the
 nodes are safely behind a firewall, you can simply instruct the Controller to listen for
 connections on all its interfaces, by adding in :file:`ipcontroller_config`:
 
 .. sourcecode:: python
 
     c.HubFactory.ip = '*'
 
 You can now run the cluster with::
 
-    $ ipcluster start --profile=pbs --n=128
+    $ ipcluster start --profile=pbs -n 128
 
 Additional configuration options can be found in the PBS section of :file:`ipcluster_config`.
 
 .. note::
 
     Due to the flexibility of configuration, the PBS launchers work with simple changes
     to the template for other :command:`qsub`-using systems, such as Sun Grid Engine,
     and with further configuration in similar batch systems like Condor.
 
 
 Using :command:`ipcluster` in SSH mode
 ---------------------------------------
 
 
 The SSH mode uses :command:`ssh` to execute :command:`ipengine` on remote
 nodes and :command:`ipcontroller` can be run remotely as well, or on localhost.
 
 .. note::
 
     When using this mode it highly recommended that you have set up SSH keys
     and are using ssh-agent [SSH]_ for password-less logins.
 
 As usual, we start by creating a clean profile::
 
     $ ipython profile create --parallel --profile=ssh
 
 To use this mode, select the SSH launchers in :file:`ipcluster_config.py`:
 
 .. sourcecode:: python
 
     c.IPClusterEngines.engine_launcher = \
         'IPython.parallel.apps.launcher.SSHEngineSetLauncher'
     # and if the Controller is also to be remote:
     c.IPClusterStart.controller_launcher = \
         'IPython.parallel.apps.launcher.SSHControllerLauncher'
     
 
 The controller's remote location and configuration can be specified:
 
 .. sourcecode:: python
 
     # Set the user and hostname for the controller
     # c.SSHControllerLauncher.hostname = 'controller.example.com'
     # c.SSHControllerLauncher.user = os.environ.get('USER','username')
 
     # Set the arguments to be passed to ipcontroller
     # note that remotely launched ipcontroller will not get the contents of 
     # the local ipcontroller_config.py unless it resides on the *remote host*
     # in the location specified by the `profile-dir` argument.
     # c.SSHControllerLauncher.program_args = ['--reuse', '--ip=*', '--profile-dir=/path/to/cd']
 
 .. note::
 
     SSH mode does not do any file movement, so you will need to distribute configuration
     files manually.  To aid in this, the `reuse_files` flag defaults to True for ssh-launched
     Controllers, so you will only need to do this once, unless you override this flag back
     to False.
 
 Engines are specified in a dictionary, by hostname and the number of engines to be run
 on that host.
 
 .. sourcecode:: python
 
     c.SSHEngineSetLauncher.engines = { 'host1.example.com' : 2,
                 'host2.example.com' : 5,
                 'host3.example.com' : (1, ['--profile-dir=/home/different/location']),
                 'host4.example.com' : 8 }
 
 * The `engines` dict, where the keys are the host we want to run engines on and
   the value is the number of engines to run on that host.
 * on host3, the value is a tuple, where the number of engines is first, and the arguments
   to be passed to :command:`ipengine` are the second element.
 
 For engines without explicitly specified arguments, the default arguments are set in
 a single location:
 
 .. sourcecode:: python
 
     c.SSHEngineSetLauncher.engine_args = ['--profile-dir=/path/to/profile_ssh']
 
 Current limitations of the SSH mode of :command:`ipcluster` are:
 
 * Untested on Windows. Would require a working :command:`ssh` on Windows.
   Also, we are using shell scripts to setup and execute commands on remote
   hosts.
 * No file movement - This is a regression from 0.10, which moved connection files
   around with scp. This will be improved, but not before 0.11 release.
 
 Using the :command:`ipcontroller` and :command:`ipengine` commands
 ====================================================================
 
 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.
 
 Starting the controller and engine on your local machine
 --------------------------------------------------------
 
 To use :command:`ipcontroller` and :command:`ipengine` to start things on your
 local machine, do the following.
 
 First start the controller::
 
     $ ipcontroller
     
 Next, start however many instances of the engine you want using (repeatedly)
 the command::
 
     $ ipengine
 
 The engines should start and automatically connect to the controller using the
 JSON files in :file:`~/.ipython/profile_default/security`. You are now ready to use the
 controller and engines from IPython.
 
 .. warning:: 
     
     The order of the above operations may be important.  You *must*
     start the controller before the engines, unless you are reusing connection
     information (via ``--reuse``), in which case ordering is not important.
 
 .. note::
 
     On some platforms (OS X), to put the controller and engine into the
     background you may need to give these commands in the form ``(ipcontroller
     &)`` and ``(ipengine &)`` (with the parentheses) for them to work
     properly.
 
 Starting the controller and engines on different hosts
 ------------------------------------------------------
 
 When the controller and engines are running on different hosts, things are
 slightly more complicated, but the underlying ideas are the same:
 
 1. Start the controller on a host using :command:`ipcontroller`. The controller must be
    instructed to listen on an interface visible to the engine machines, via the ``ip``
    command-line argument or ``HubFactory.ip`` in :file:`ipcontroller_config.py`.
 2. Copy :file:`ipcontroller-engine.json` from :file:`~/.ipython/profile_<name>/security` on
    the controller's host to the host where the engines will run.
 3. Use :command:`ipengine` on the engine's hosts to start the engines.
 
 The only thing you have to be careful of is to tell :command:`ipengine` where
 the :file:`ipcontroller-engine.json` file is located. There are two ways you
 can do this:
 
 * Put :file:`ipcontroller-engine.json` in the :file:`~/.ipython/profile_<name>/security`
   directory on the engine's host, where it will be found automatically.
 * Call :command:`ipengine` with the ``--file=full_path_to_the_file``
   flag.
 
 The ``file`` flag works like this::
 
     $ ipengine --file=/path/to/my/ipcontroller-engine.json
 
 .. note::
     
     If the controller's and engine's hosts all have a shared file system  
     (:file:`~/.ipython/profile_<name>/security` is the same on all of them), then things
     will just work!
 
 Make JSON files persistent
 --------------------------
 
 At fist glance it may seem that that managing the JSON files is a bit
 annoying. Going back to the house and key analogy, copying the JSON around
 each time you start the controller is like having to make a new key every time
 you want to unlock the door and enter your house. As with your house, you want
 to be able to create the key (or JSON file) once, and then simply use it at
 any point in the future.
 
 To do this, the only thing you have to do is specify the `--reuse` flag, so that
 the connection information in the JSON files remains accurate::
 
     $ ipcontroller --reuse
 
 Then, just copy the JSON files over the first time and you are set. You can
 start and stop the controller and engines any many times as you want in the
 future, just make sure to tell the controller to reuse the file.
 
 .. note::
 
     You may ask the question: what ports does the controller listen on if you
     don't tell is to use specific ones? The default is to use high random port
     numbers. We do this for two reasons: i) to increase security through
     obscurity and ii) to multiple controllers on a given host to start and
     automatically use different ports.
 
 Log files
 ---------
 
 All of the components of IPython have log files associated with them.
 These log files can be extremely useful in debugging problems with
 IPython and can be found in the directory :file:`~/.ipython/profile_<name>/log`.
 Sending the log files to us will often help us to debug any problems.
 
 
 Configuring `ipcontroller`
 ---------------------------
 
 The IPython Controller takes its configuration from the file :file:`ipcontroller_config.py`
 in the active profile directory.
 
 Ports and addresses
 *******************
 
 In many cases, you will want to configure the Controller's network identity.  By default,
 the Controller listens only on loopback, which is the most secure but often impractical.
 To instruct the controller to listen on a specific interface, you can set the 
 :attr:`HubFactory.ip` trait.  To listen on all interfaces, simply specify:
 
 .. sourcecode:: python
 
     c.HubFactory.ip = '*'
 
 When connecting to a Controller that is listening on loopback or behind a firewall, it may
 be necessary to specify an SSH server to use for tunnels, and the external IP of the
 Controller. If you specified that the HubFactory listen on loopback, or all interfaces,
 then IPython will try to guess the external IP. If you are on a system with VM network
 devices, or many interfaces, this guess may be incorrect. In these cases, you will want
 to specify the 'location' of the Controller. This is the IP of the machine the Controller
 is on, as seen by the clients, engines, or the SSH server used to tunnel connections.
 
 For example, to set up a cluster with a Controller on a work node, using ssh tunnels
 through the login node, an example :file:`ipcontroller_config.py` might contain:
 
 .. sourcecode:: python
 
     # allow connections on all interfaces from engines
     # engines on the same node will use loopback, while engines
     # from other nodes will use an external IP
     c.HubFactory.ip = '*'
     
     # you typically only need to specify the location when there are extra
     # interfaces that may not be visible to peer nodes (e.g. VM interfaces)
     c.HubFactory.location = '10.0.1.5'
     # or to get an automatic value, try this:
     import socket
     ex_ip = socket.gethostbyname_ex(socket.gethostname())[-1][0]
     c.HubFactory.location = ex_ip
     
     # now instruct clients to use the login node for SSH tunnels:
     c.HubFactory.ssh_server = 'login.mycluster.net'
 
 After doing this, your :file:`ipcontroller-client.json` file will look something like this:
 
 .. this can be Python, despite the fact that it's actually JSON, because it's
 .. still valid Python
 
 .. sourcecode:: python
 
     {
       "url":"tcp:\/\/*:43447",
       "exec_key":"9c7779e4-d08a-4c3b-ba8e-db1f80b562c1",
       "ssh":"login.mycluster.net",
       "location":"10.0.1.5"
     }
 
 Then this file will be all you need for a client to connect to the controller, tunneling
 SSH connections through login.mycluster.net.
 
 Database Backend
 ****************
 
 The Hub stores all messages and results passed between Clients and Engines.
 For large and/or long-running clusters, it would be unreasonable to keep all
 of this information in memory. For this reason, we have two database backends:
 [MongoDB]_ via PyMongo_, and SQLite with the stdlib :py:mod:`sqlite`.
 
 MongoDB is our design target, and the dict-like model it uses has driven our design. As far
 as we are concerned, BSON can be considered essentially the same as JSON, adding support
 for binary data and datetime objects, and any new database backend must support the same
 data types.
 
 .. seealso::
 
     MongoDB `BSON doc <http://www.mongodb.org/display/DOCS/BSON>`_
 
 To use one of these backends, you must set the :attr:`HubFactory.db_class` trait:
 
 .. sourcecode:: python
 
     # for a simple dict-based in-memory implementation, use dictdb
     # This is the default and the fastest, since it doesn't involve the filesystem
     c.HubFactory.db_class = 'IPython.parallel.controller.dictdb.DictDB'
     
     # To use MongoDB:
     c.HubFactory.db_class = 'IPython.parallel.controller.mongodb.MongoDB'
     
     # and SQLite:
     c.HubFactory.db_class = 'IPython.parallel.controller.sqlitedb.SQLiteDB'
 
 When using the proper databases, you can actually allow for tasks to persist from
 one session to the next by specifying the MongoDB database or SQLite table in
 which tasks are to be stored.  The default is to use a table named for the Hub's Session,
 which is a UUID, and thus different every time.
 
 .. sourcecode:: python
 
     # To keep persistant task history in MongoDB:
     c.MongoDB.database = 'tasks'
     
     # and in SQLite:
     c.SQLiteDB.table = 'tasks'
 
 
 Since MongoDB servers can be running remotely or configured to listen on a particular port,
 you can specify any arguments you may need to the PyMongo `Connection 
 <http://api.mongodb.org/python/1.9/api/pymongo/connection.html#pymongo.connection.Connection>`_:
 
 .. sourcecode:: python
 
     # positional args to pymongo.Connection
     c.MongoDB.connection_args = []
     
     # keyword args to pymongo.Connection
     c.MongoDB.connection_kwargs = {}
 
 .. _MongoDB: http://www.mongodb.org
 .. _PyMongo: http://api.mongodb.org/python/1.9/
 
 Configuring `ipengine`
 -----------------------
 
 The IPython Engine takes its configuration from the file :file:`ipengine_config.py`
 
 The Engine itself also has some amount of configuration. Most of this
 has to do with initializing MPI or connecting to the controller.
 
 To instruct the Engine to initialize with an MPI environment set up by
 mpi4py, add:
 
 .. sourcecode:: python
 
     c.MPI.use = 'mpi4py'
 
 In this case, the Engine will use our default mpi4py init script to set up
 the MPI environment prior to exection.  We have default init scripts for
 mpi4py and pytrilinos.  If you want to specify your own code to be run
 at the beginning, specify `c.MPI.init_script`.
 
 You can also specify a file or python command to be run at startup of the
 Engine:
 
 .. sourcecode:: python
 
     c.IPEngineApp.startup_script = u'/path/to/my/startup.py'
     
     c.IPEngineApp.startup_command = 'import numpy, scipy, mpi4py'
 
 These commands/files will be run again, after each
 
 It's also useful on systems with shared filesystems to run the engines
 in some scratch directory.  This can be set with:
 
 .. sourcecode:: python
 
     c.IPEngineApp.work_dir = u'/path/to/scratch/'
 
 
 
 .. [MongoDB] MongoDB database http://www.mongodb.org
 
 .. [PBS] Portable Batch System http://www.openpbs.org
 
 .. [SSH] SSH-Agent http://en.wikipedia.org/wiki/ssh-agent

docs/source/parallel/parallel_task.txt

@@ -1,442 +1,442 b''
 .. _parallel_task:
 
 ==========================
 The IPython task interface
 ==========================
 
 The task interface to the cluster presents the engines as a fault tolerant,
 dynamic load-balanced system of workers. Unlike the multiengine interface, in
 the task interface the user have no direct access to individual engines. By
 allowing the IPython scheduler to assign work, this interface is simultaneously
 simpler and more powerful.
 
 Best of all, the user can use both of these interfaces running at the same time
 to take advantage of their respective 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.
 
 Starting the IPython controller and engines
 ===========================================
 
 To follow along with this tutorial, you will need to start the IPython
 controller and four IPython engines. The simplest way of doing this is to use
 the :command:`ipcluster` command::
 
-	$ ipcluster start --n=4
+	$ ipcluster start -n 4
 	
 For more detailed information about starting the controller and engines, see
 our :ref:`introduction <parallel_overview>` to using IPython for parallel computing.
 
 Creating a ``Client`` instance
 ==============================
 
 The first step is to import the IPython :mod:`IPython.parallel`
 module and then create a :class:`.Client` instance, and we will also be using
 a :class:`LoadBalancedView`, here called `lview`:
 
 .. sourcecode:: ipython
 
     In [1]: from IPython.parallel import Client
 
     In [2]: rc = Client()
     
 
 This form assumes that the controller was started on localhost with default
 configuration. If not, the location of the controller must be given as an
 argument to the constructor:
 
 .. sourcecode:: ipython
 
     # for a visible LAN controller listening on an external port:
     In [2]: rc = Client('tcp://192.168.1.16:10101')
     # or to connect with a specific profile you have set up:
     In [3]: rc = Client(profile='mpi')
 
 For load-balanced execution, we will make use of a :class:`LoadBalancedView` object, which can
 be constructed via the client's :meth:`load_balanced_view` method:
 
 .. sourcecode:: ipython
 
     In [4]: lview = rc.load_balanced_view() # default load-balanced view
 
 .. seealso::
 
     For more information, see the in-depth explanation of :ref:`Views <parallel_details>`.
 
 
 Quick and easy parallelism
 ==========================
 
 In many cases, you simply want to apply a Python function to a sequence of
 objects, but *in parallel*. Like the multiengine interface, these can be
 implemented via the task interface. The exact same tools can perform these
 actions in load-balanced ways as well as multiplexed ways: a parallel version
 of :func:`map` and :func:`@parallel` function decorator. If one specifies the
 argument `balanced=True`, then they are dynamically load balanced. Thus, if the
 execution time per item varies significantly, you should use the versions in
 the task interface.
 
 Parallel map
 ------------
 
 To load-balance :meth:`map`,simply use a LoadBalancedView:
 
 .. sourcecode:: ipython
     
     In [62]: lview.block = True
     
 	In [63]: serial_result = map(lambda x:x**10, range(32))
 
 	In [64]: parallel_result = lview.map(lambda x:x**10, range(32))
 
 	In [65]: serial_result==parallel_result
 	Out[65]: True
 
 Parallel function decorator
 ---------------------------
 
 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:
 
 .. sourcecode:: ipython
 
     In [10]: @lview.parallel()
        ....: def f(x):
        ....:     return 10.0*x**4
        ....:
 
     In [11]: f.map(range(32))    # this is done in parallel
     Out[11]: [0.0,10.0,160.0,...]
 
 .. _parallel_dependencies:
 
 Dependencies
 ============
 
 Often, pure atomic load-balancing is too primitive for your work. In these cases, you
 may want to associate some kind of `Dependency` that describes when, where, or whether
 a task can be run.  In IPython, we provide two types of dependencies: 
 `Functional Dependencies`_ and `Graph Dependencies`_
 
 .. note::
 
     It is important to note that the pure ZeroMQ scheduler does not support dependencies,
     and you will see errors or warnings if you try to use dependencies with the pure
     scheduler.
 
 Functional Dependencies
 -----------------------
 
 Functional dependencies are used to determine whether a given engine is capable of running
 a particular task.  This is implemented via a special :class:`Exception` class, 
 :class:`UnmetDependency`, found in `IPython.parallel.error`.  Its use is very simple: 
 if a task fails with an UnmetDependency exception, then the scheduler, instead of relaying
 the error up to the client like any other error, catches the error, and submits the task
 to a different engine.  This will repeat indefinitely, and a task will never be submitted
 to a given engine a second time.
 
 You can manually raise the :class:`UnmetDependency` yourself, but IPython has provided 
 some decorators for facilitating this behavior.
 
 There are two decorators and a class used for functional dependencies:
 
 .. sourcecode:: ipython
 
     In [9]: from IPython.parallel import depend, require, dependent
 
 @require
 ********
 
 The simplest sort of dependency is requiring that a Python module is available. The
 ``@require`` decorator lets you define a function that will only run on engines where names
 you specify are importable:
 
 .. sourcecode:: ipython
 
     In [10]: @require('numpy', 'zmq')
         ...: def myfunc():
         ...:     return dostuff()
 
 Now, any time you apply :func:`myfunc`, the task will only run on a machine that has
 numpy and pyzmq available, and when :func:`myfunc` is called, numpy and zmq will be imported.
 
 @depend
 *******
 
 The ``@depend`` decorator lets you decorate any function with any *other* function to
 evaluate the dependency. The dependency function will be called at the start of the task,
 and if it returns ``False``, then the dependency will be considered unmet, and the task
 will be assigned to another engine. If the dependency returns *anything other than
 ``False``*, the rest of the task will continue.
 
 .. sourcecode:: ipython
 
     In [10]: def platform_specific(plat):
         ...:    import sys
         ...:    return sys.platform == plat
     
     In [11]: @depend(platform_specific, 'darwin')
         ...: def mactask():
         ...:    do_mac_stuff()
 
     In [12]: @depend(platform_specific, 'nt')
         ...: def wintask():
         ...:    do_windows_stuff()
     
 In this case, any time you apply ``mytask``, it will only run on an OSX machine.
 ``@depend`` is just like ``apply``, in that it has a ``@depend(f,*args,**kwargs)``
 signature.
 
 dependents
 **********
 
 You don't have to use the decorators on your tasks, if for instance you may want
 to run tasks with a single function but varying dependencies, you can directly construct
 the :class:`dependent` object that the decorators use:
 
 .. sourcecode::ipython
 
     In [13]: def mytask(*args):
         ...:    dostuff()
     
     In [14]: mactask = dependent(mytask, platform_specific, 'darwin')
     # this is the same as decorating the declaration of mytask with @depend
     # but you can do it again:
     
     In [15]: wintask = dependent(mytask, platform_specific, 'nt')
     
     # in general:
     In [16]: t = dependent(f, g, *dargs, **dkwargs)
     
     # is equivalent to:
     In [17]: @depend(g, *dargs, **dkwargs)
         ...: def t(a,b,c):
         ...:     # contents of f
 
 Graph Dependencies
 ------------------
 
 Sometimes you want to restrict the time and/or location to run a given task as a function
 of the time and/or location of other tasks. This is implemented via a subclass of
 :class:`set`, called a :class:`Dependency`. A Dependency is just a set of `msg_ids`
 corresponding to tasks, and a few attributes to guide how to decide when the Dependency
 has been met.
 
 The switches we provide for interpreting whether a given dependency set has been met:
 
 any|all
     Whether the dependency is considered met if *any* of the dependencies are done, or
     only after *all* of them have finished.  This is set by a Dependency's :attr:`all`
     boolean attribute, which defaults to ``True``.
 
 success [default: True]
     Whether to consider tasks that succeeded as fulfilling dependencies.
 
 failure [default : False]
     Whether to consider tasks that failed as fulfilling dependencies.
     using `failure=True,success=False` is useful for setting up cleanup tasks, to be run
     only when tasks have failed.
 
 Sometimes you want to run a task after another, but only if that task succeeded. In this case,
 ``success`` should be ``True`` and ``failure`` should be ``False``. However sometimes you may
 not care whether the task succeeds, and always want the second task to run, in which case you
 should use `success=failure=True`. The default behavior is to only use successes.
 
 There are other switches for interpretation that are made at the *task* level.  These are
 specified via keyword arguments to the client's :meth:`apply` method.
 
 after,follow
     You may want to run a task *after* a given set of dependencies have been run and/or
     run it *where* another set of dependencies are met. To support this, every task has an
     `after` dependency to restrict time, and a `follow` dependency to restrict
     destination.
 
 timeout
     You may also want to set a time-limit for how long the scheduler should wait before a
     task's dependencies are met. This is done via a `timeout`, which defaults to 0, which
     indicates that the task should never timeout. If the timeout is reached, and the
     scheduler still hasn't been able to assign the task to an engine, the task will fail
     with a :class:`DependencyTimeout`.
 
 .. note::
 
     Dependencies only work within the task scheduler. You cannot instruct a load-balanced
     task to run after a job submitted via the MUX interface.
 
 The simplest form of Dependencies is with `all=True,success=True,failure=False`. In these cases,
 you can skip using Dependency objects, and just pass msg_ids or AsyncResult objects as the
 `follow` and `after` keywords to :meth:`client.apply`:
 
 .. sourcecode:: ipython
 
     In [14]: client.block=False
 
     In [15]: ar = lview.apply(f, args, kwargs)
 
     In [16]: ar2 = lview.apply(f2)
 
     In [17]: ar3 = lview.apply_with_flags(f3, after=[ar,ar2])
 
     In [17]: ar4 = lview.apply_with_flags(f3, follow=[ar], timeout=2.5)
 
 
 .. seealso::
 
     Some parallel workloads can be described as a `Directed Acyclic Graph
     <http://en.wikipedia.org/wiki/Directed_acyclic_graph>`_, or DAG. See :ref:`DAG
     Dependencies <dag_dependencies>` for an example demonstrating how to use map a NetworkX DAG
     onto task dependencies.
 
 
 
 
 Impossible Dependencies
 ***********************
 
 The schedulers do perform some analysis on graph dependencies to determine whether they
 are not possible to be met. If the scheduler does discover that a dependency cannot be
 met, then the task will fail with an :class:`ImpossibleDependency` error. This way, if the
 scheduler realized that a task can never be run, it won't sit indefinitely in the
 scheduler clogging the pipeline.
 
 The basic cases that are checked:
 
 * depending on nonexistent messages
 * `follow` dependencies were run on more than one machine and `all=True`
 * any dependencies failed and `all=True,success=True,failures=False`
 * all dependencies failed and `all=False,success=True,failure=False`
 
 .. warning::
 
     This analysis has not been proven to be rigorous, so it is likely possible for tasks
     to become impossible to run in obscure situations, so a timeout may be a good choice.
 
 
 Retries and Resubmit
 ====================
 
 Retries
 -------
 
 Another flag for tasks is `retries`.  This is an integer, specifying how many times
 a task should be resubmitted after failure.  This is useful for tasks that should still run
 if their engine was shutdown, or may have some statistical chance of failing.  The default
 is to not retry tasks.
 
 Resubmit
 --------
 
 Sometimes you may want to re-run a task. This could be because it failed for some reason, and
 you have fixed the error, or because you want to restore the cluster to an interrupted state.
 For this, the :class:`Client` has a :meth:`rc.resubmit` method.  This simply takes one or more
 msg_ids, and returns an :class:`AsyncHubResult` for the result(s).  You cannot resubmit
 a task that is pending - only those that have finished, either successful or unsuccessful.
 
 .. _parallel_schedulers:
 
 Schedulers
 ==========
 
 There are a variety of valid ways to determine where jobs should be assigned in a 
 load-balancing situation.  In IPython, we support several standard schemes, and
 even make it easy to define your own.  The scheme can be selected via the ``scheme``
 argument to :command:`ipcontroller`, or in the :attr:`TaskScheduler.schemename` attribute
 of a controller config object.
 
 The built-in routing schemes:
 
 To select one of these schemes, simply do::
 
     $ ipcontroller --scheme=<schemename>
     for instance:
     $ ipcontroller --scheme=lru
 
 lru: Least Recently Used
 
     Always assign work to the least-recently-used engine.  A close relative of
     round-robin, it will be fair with respect to the number of tasks, agnostic 
     with respect to runtime of each task.
 
 plainrandom: Plain Random
 
     Randomly picks an engine on which to run.
 
 twobin: Two-Bin Random
 
     **Requires numpy**
     
     Pick two engines at random, and use the LRU of the two. This is known to be better
     than plain random in many cases, but requires a small amount of computation.
 
 leastload: Least Load
 
     **This is the default scheme**
 
     Always assign tasks to the engine with the fewest outstanding tasks (LRU breaks tie).
 
 weighted: Weighted Two-Bin Random
 
     **Requires numpy**
     
     Pick two engines at random using the number of outstanding tasks as inverse weights,
     and use the one with the lower load.
 
 
 Pure ZMQ Scheduler
 ------------------
 
 For maximum throughput, the 'pure' scheme is not Python at all, but a C-level
 :class:`MonitoredQueue` from PyZMQ, which uses a ZeroMQ ``XREQ`` socket to perform all
 load-balancing. This scheduler does not support any of the advanced features of the Python
 :class:`.Scheduler`.
 
 Disabled features when using the ZMQ Scheduler:
 
 * Engine unregistration
     Task farming will be disabled if an engine unregisters.  
     Further, if an engine is unregistered during computation, the scheduler may not recover.
 * Dependencies
     Since there is no Python logic inside the Scheduler, routing decisions cannot be made
     based on message content.
 * Early destination notification
     The Python schedulers know which engine gets which task, and notify the Hub.  This 
     allows graceful handling of Engines coming and going.  There is no way to know
     where ZeroMQ messages have gone, so there is no way to know what tasks are on which
     engine until they *finish*.  This makes recovery from engine shutdown very difficult.
 
 
 .. note::
 
     TODO: performance comparisons
 
 
 
 
 More details
 ============
 
 The :class:`LoadBalancedView` has many more powerful features that allow quite a bit
 of flexibility in how tasks are defined and run. The next places to look are
 in the following classes:
 
 * :class:`~IPython.parallel.client.view.LoadBalancedView`
 * :class:`~IPython.parallel.client.asyncresult.AsyncResult`
 * :meth:`~IPython.parallel.client.view.LoadBalancedView.apply`
 * :mod:`~IPython.parallel.controller.dependency`
 
 The following is an overview of how to use these classes together:
 
 1. Create a :class:`Client` and :class:`LoadBalancedView`
 2. Define some functions to be run as tasks
 3. Submit your tasks to using the :meth:`apply` method of your
    :class:`LoadBalancedView` instance.
 4. Use :meth:`Client.get_result` to get the results of the
    tasks, or use the :meth:`AsyncResult.get` method of the results to wait
    for and then receive the results.
 
 .. seealso::
 
     A demo of :ref:`DAG Dependencies <dag_dependencies>` with NetworkX and IPython.

docs/source/parallel/parallel_winhpc.txt

@@ -1,334 +1,334 b''
 ============================================
 Getting started with Windows HPC Server 2008
 ============================================
 
 .. note::
 
     Not adapted to zmq yet
 
 Introduction
 ============
 
 The Python programming language is an increasingly popular language for
 numerical computing. This is due to a unique combination of factors. First,
 Python is a high-level and *interactive* language that is well matched to
 interactive numerical work. Second, it is easy (often times trivial) to
 integrate legacy C/C++/Fortran code into Python. Third, a large number of
 high-quality open source projects provide all the needed building blocks for
 numerical computing: numerical arrays (NumPy), algorithms (SciPy), 2D/3D
 Visualization (Matplotlib, Mayavi, Chaco), Symbolic Mathematics (Sage, Sympy)
 and others.
 
 The IPython project is a core part of this open-source toolchain and is
 focused on creating a comprehensive environment for interactive and
 exploratory computing in the Python programming language. It enables all of
 the above tools to be used interactively and consists of two main components:
 
 * An enhanced interactive Python shell with support for interactive plotting
   and visualization.
 * An architecture for interactive parallel computing.
 
 With these components, it is possible to perform all aspects of a parallel
 computation interactively. This type of workflow is particularly relevant in
 scientific and numerical computing where algorithms, code and data are
 continually evolving as the user/developer explores a problem. The broad
 treads in computing (commodity clusters, multicore, cloud computing, etc.)
 make these capabilities of IPython particularly relevant.
 
 While IPython is a cross platform tool, it has particularly strong support for
 Windows based compute clusters running Windows HPC Server 2008. This document
 describes how to get started with IPython on Windows HPC Server 2008. The
 content and emphasis here is practical: installing IPython, configuring
 IPython to use the Windows job scheduler and running example parallel programs
 interactively. A more complete description of IPython's parallel computing
 capabilities can be found in IPython's online documentation
 (http://ipython.org/documentation.html).
 
 Setting up your Windows cluster
 ===============================
 
 This document assumes that you already have a cluster running Windows
 HPC Server 2008. Here is a broad overview of what is involved with setting up
 such a cluster:
 
 1. Install Windows Server 2008 on the head and compute nodes in the cluster.
 2. Setup the network configuration on each host. Each host should have a 
    static IP address.
 3. On the head node, activate the "Active Directory Domain Services" role 
    and make the head node the domain controller.
 4. Join the compute nodes to the newly created Active Directory (AD) domain.
 5. Setup user accounts in the domain with shared home directories.
 6. Install the HPC Pack 2008 on the head node to create a cluster.
 7. Install the HPC Pack 2008 on the compute nodes.
 
 More details about installing and configuring Windows HPC Server 2008 can be
 found on the Windows HPC Home Page (http://www.microsoft.com/hpc). Regardless
 of what steps you follow to set up your cluster, the remainder of this
 document will assume that:
 
 * There are domain users that can log on to the AD domain and submit jobs
   to the cluster scheduler.
 * These domain users have shared home directories. While shared home
   directories are not required to use IPython, they make it much easier to
   use IPython.
 
 Installation of IPython and its dependencies
 ============================================
 
 IPython and all of its dependencies are freely available and open source.
 These packages provide a powerful and cost-effective approach to numerical and
 scientific computing on Windows. The following dependencies are needed to run
 IPython on Windows:
 
 * Python 2.6 or 2.7 (http://www.python.org)
 * pywin32 (http://sourceforge.net/projects/pywin32/)
 * PyReadline (https://launchpad.net/pyreadline)
 * pyzmq (http://github.com/zeromq/pyzmq/downloads)
 * IPython (http://ipython.org)
 
 In addition, the following dependencies are needed to run the demos described
 in this document.
 
 * NumPy and SciPy (http://www.scipy.org)
 * Matplotlib (http://matplotlib.sourceforge.net/)
 
 The easiest way of obtaining these dependencies is through the Enthought
 Python Distribution (EPD) (http://www.enthought.com/products/epd.php). EPD is
 produced by Enthought, Inc. and contains all of these packages and others in a
 single installer and is available free for academic users. While it is also
 possible to download and install each package individually, this is a tedious
 process. Thus, we highly recommend using EPD to install these packages on
 Windows.
 
 Regardless of how you install the dependencies, here are the steps you will
 need to follow:
 
 1. Install all of the packages listed above, either individually or using EPD
    on the head node, compute nodes and user workstations.
 
 2. Make sure that :file:`C:\\Python27` and :file:`C:\\Python27\\Scripts` are
    in the system :envvar:`%PATH%` variable on each node.
 
 3. Install the latest development version of IPython. This can be done by
    downloading the the development version from the IPython website
    (http://ipython.org) and following the installation instructions.
 
 Further details about installing IPython or its dependencies can be found in
 the online IPython documentation (http://ipython.org/documentation.html)
 Once you are finished with the installation, you can try IPython out by
 opening a Windows Command Prompt and typing ``ipython``. This will
 start IPython's interactive shell and you should see something like the
 following screenshot:
 
 .. image:: ipython_shell.*
 
 Starting an IPython cluster
 ===========================
 
 To use IPython's parallel computing capabilities, you will need to start an
 IPython cluster. An IPython cluster consists of one controller and multiple
 engines:
 
 IPython controller
     The IPython controller manages the engines and acts as a gateway between 
     the engines and the client, which runs in the user's interactive IPython
     session. The controller is started using the :command:`ipcontroller`
     command.
 
 IPython engine
     IPython engines run a user's Python code in parallel on the compute nodes. 
     Engines are starting using the :command:`ipengine` command.
 
 Once these processes are started, a user can run Python code interactively and
 in parallel on the engines from within the IPython shell using an appropriate
 client. This includes the ability to interact with, plot and visualize data
 from the engines.
 
 IPython has a command line program called :command:`ipcluster` that automates
 all aspects of starting the controller and engines on the compute nodes.
 :command:`ipcluster` has full support for the Windows HPC job scheduler,
 meaning that :command:`ipcluster` can use this job scheduler to start the
 controller and engines. In our experience, the Windows HPC job scheduler is
 particularly well suited for interactive applications, such as IPython. Once
 :command:`ipcluster` is configured properly, a user can start an IPython
 cluster from their local workstation almost instantly, without having to log
 on to the head node (as is typically required by Unix based job schedulers).
 This enables a user to move seamlessly between serial and parallel
 computations.
 
 In this section we show how to use :command:`ipcluster` to start an IPython
 cluster using the Windows HPC Server 2008 job scheduler. To make sure that
 :command:`ipcluster` is installed and working properly, you should first try
 to start an IPython cluster on your local host. To do this, open a Windows
 Command Prompt and type the following command::
 
     ipcluster start n=2
 
 You should see a number of messages printed to the screen, ending with 
 "IPython cluster: started". The result should look something like the following
 screenshot:
 
 .. image:: ipcluster_start.*
 
 At this point, the controller and two engines are running on your local host.
 This configuration is useful for testing and for situations where you want to
 take advantage of multiple cores on your local computer. 
 
 Now that we have confirmed that :command:`ipcluster` is working properly, we
 describe how to configure and run an IPython cluster on an actual compute
 cluster running Windows HPC Server 2008. Here is an outline of the needed
 steps:
 
 1. Create a cluster profile using: ``ipython profile create --parallel profile=mycluster``
 
 2. Edit configuration files in the directory :file:`.ipython\\cluster_mycluster`
 
 3. Start the cluster using: ``ipcluser start profile=mycluster n=32``
 
 Creating a cluster profile
 --------------------------
 
 In most cases, you will have to create a cluster profile to use IPython on a
 cluster. A cluster profile is a name (like "mycluster") that is associated
 with a particular cluster configuration. The profile name is used by
 :command:`ipcluster` when working with the cluster.
 
 Associated with each cluster profile is a cluster directory. This cluster
 directory is a specially named directory (typically located in the
 :file:`.ipython` subdirectory of your home directory) that contains the
 configuration files for a particular cluster profile, as well as log files and
 security keys. The naming convention for cluster directories is:
 :file:`profile_<profile name>`. Thus, the cluster directory for a profile named
 "foo" would be :file:`.ipython\\cluster_foo`.
 
 To create a new cluster profile (named "mycluster") and the associated cluster
 directory, type the following command at the Windows Command Prompt::
 
     ipython profile create --parallel --profile=mycluster
 
 The output of this command is shown in the screenshot below. Notice how
 :command:`ipcluster` prints out the location of the newly created cluster
 directory.
 
 .. image:: ipcluster_create.*
 
 Configuring a cluster profile
 -----------------------------
 
 Next, you will need to configure the newly created cluster profile by editing
 the following configuration files in the cluster directory:
 
 * :file:`ipcluster_config.py`
 * :file:`ipcontroller_config.py`
 * :file:`ipengine_config.py`
 
 When :command:`ipcluster` is run, these configuration files are used to 
 determine how the engines and controller will be started. In most cases,
 you will only have to set a few of the attributes in these files.
 
 To configure :command:`ipcluster` to use the Windows HPC job scheduler, you
 will need to edit the following attributes in the file
 :file:`ipcluster_config.py`::
 
     # Set these at the top of the file to tell ipcluster to use the
     # Windows HPC job scheduler.
     c.IPClusterStart.controller_launcher = \
         'IPython.parallel.apps.launcher.WindowsHPCControllerLauncher'
     c.IPClusterEngines.engine_launcher = \
         'IPython.parallel.apps.launcher.WindowsHPCEngineSetLauncher'
 
     # Set these to the host name of the scheduler (head node) of your cluster.
     c.WindowsHPCControllerLauncher.scheduler = 'HEADNODE'
     c.WindowsHPCEngineSetLauncher.scheduler = 'HEADNODE'
 
 There are a number of other configuration attributes that can be set, but
 in most cases these will be sufficient to get you started.
 
 .. warning::
     If any of your configuration attributes involve specifying the location
     of shared directories or files, you must make sure that you use UNC paths
     like :file:`\\\\host\\share`. It is also important that you specify
     these paths using raw Python strings: ``r'\\host\share'`` to make sure
     that the backslashes are properly escaped.
 
 Starting the cluster profile
 ----------------------------
 
 Once a cluster profile has been configured, starting an IPython cluster using
 the profile is simple::
 
-    ipcluster start --profile=mycluster --n=32
+    ipcluster start --profile=mycluster -n 32
 
 The ``-n`` option tells :command:`ipcluster` how many engines to start (in
 this case 32). Stopping the cluster is as simple as typing Control-C.
 
 Using the HPC Job Manager
 -------------------------
 
 When ``ipcluster start`` is run the first time, :command:`ipcluster` creates
 two XML job description files in the cluster directory:
 
 * :file:`ipcontroller_job.xml`
 * :file:`ipengineset_job.xml`
 
 Once these files have been created, they can be imported into the HPC Job
 Manager application. Then, the controller and engines for that profile can be
 started using the HPC Job Manager directly, without using :command:`ipcluster`.
 However, anytime the cluster profile is re-configured, ``ipcluster start``
 must be run again to regenerate the XML job description files. The
 following screenshot shows what the HPC Job Manager interface looks like
 with a running IPython cluster.
 
 .. image:: hpc_job_manager.*
 
 Performing a simple interactive parallel computation
 ====================================================
 
 Once you have started your IPython cluster, you can start to use it. To do
 this, open up a new Windows Command Prompt and start up IPython's interactive
 shell by typing::
 
     ipython
 
 Then you can create a :class:`MultiEngineClient` instance for your profile and
 use the resulting instance to do a simple interactive parallel computation. In
 the code and screenshot that follows, we take a simple Python function and
 apply it to each element of an array of integers in parallel using the
 :meth:`MultiEngineClient.map` method:
 
 .. sourcecode:: ipython
 
     In [1]: from IPython.parallel import *
 
     In [2]: c = MultiEngineClient(profile='mycluster')
 
     In [3]: mec.get_ids()
     Out[3]: [0, 1, 2, 3, 4, 5, 67, 8, 9, 10, 11, 12, 13, 14]
 
     In [4]: def f(x):
        ...:     return x**10
 
     In [5]: mec.map(f, range(15))  # f is applied in parallel
     Out[5]:
     [0,
      1,
      1024,
      59049,
      1048576,
      9765625,
      60466176,
      282475249,
      1073741824,
      3486784401L,
      10000000000L,
      25937424601L,
      61917364224L,
      137858491849L,
      289254654976L]
 
 The :meth:`map` method has the same signature as Python's builtin :func:`map`
 function, but runs the calculation in parallel. More involved examples of using
 :class:`MultiEngineClient` are provided in the examples that follow.
 
 .. image:: mec_simple.*