This documentation is for a development version of IPython. There may be significant differences from the latest stable release.
One of Python’s most useful features is its interactive interpreter. It allows for very fast testing of ideas without the overhead of creating test files as is typical in most programming languages. However, the interpreter supplied with the standard Python distribution is somewhat limited for extended interactive use.
The goal of IPython is to create a comprehensive environment for interactive and exploratory computing. To support this goal, IPython has three main components:
- An enhanced interactive Python shell.
- A decoupled two-process communication model, which allows for multiple clients to connect to a computation kernel, most notably the web-based notebook provided with Jupyter.
- An architecture for interactive parallel computing now part of the
All of IPython is open source (released under the revised BSD license).
Enhanced interactive Python shell¶
IPython’s interactive shell (ipython), has the following goals, amongst others:
- Provide an interactive shell superior to Python’s default. IPython has many features for tab-completion, object introspection, system shell access, command history retrieval across sessions, and its own special command system for adding functionality when working interactively. It tries to be a very efficient environment both for Python code development and for exploration of problems using Python objects (in situations like data analysis).
- Serve as an embeddable, ready to use interpreter for your own programs. An interactive IPython shell can be started with a single call from inside another program, providing access to the current namespace. This can be very useful both for debugging purposes and for situations where a blend of batch-processing and interactive exploration are needed.
- Offer a flexible framework which can be used as the base environment for working with other systems, with Python as the underlying bridge language. Specifically scientific environments like Mathematica, IDL and Matlab inspired its design, but similar ideas can be useful in many fields.
- Allow interactive testing of threaded graphical toolkits. IPython has support for interactive, non-blocking control of GTK, Qt, WX, GLUT, and OS X applications via special threading flags. The normal Python shell can only do this for Tkinter applications.
Main features of the interactive shell¶
- Dynamic object introspection. One can access docstrings, function
definition prototypes, source code, source files and other details
of any object accessible to the interpreter with a single
?, and using
??provides additional detail).
- Searching through modules and namespaces with
*wildcards, both when using the
?system and via the
- Completion in the local namespace, by typing
TABat the prompt. This works for keywords, modules, methods, variables and files in the current directory. This is supported via the
prompt_toolkitlibrary. Custom completers can be implemented easily for different purposes (system commands, magic arguments etc.)
- Numbered input/output prompts with command history (persistent across sessions and tied to each profile), full searching in this history and caching of all input and output.
- User-extensible ‘magic’ commands. A set of commands prefixed with
%%is available for controlling IPython itself and provides directory control, namespace information and many aliases to common system shell commands.
- Alias facility for defining your own system aliases.
- Complete system shell access. Lines starting with
!are passed directly to the system shell, and using
var = !cmdcaptures shell output into python variables for further use.
- The ability to expand python variables when calling the system shell. In a
shell command, any python variable prefixed with
$is expanded. A double
$$allows passing a literal
$to the shell (for access to shell and environment variables like
- Filesystem navigation, via a magic
%cdcommand, along with a persistent bookmark system (using
%bookmark) for fast access to frequently visited directories.
- A lightweight persistence framework via the
%storecommand, which allows you to save arbitrary Python variables. These get restored when you run the
- Automatic indentation and highlighting of code as you type (through the
- Macro system for quickly re-executing multiple lines of previous
input with a single name via the
%macrocommand. Macros can be stored persistently via
%storeand edited via
- Session logging (you can then later use these logs as code in your programs). Logs can optionally timestamp all input, and also store session output (marked as comments, so the log remains valid Python source code).
- Session restoring: logs can be replayed to restore a previous session to the state where you left it.
- Verbose and colored exception traceback printouts. Easier to parse visually, and in verbose mode they produce a lot of useful debugging information (basically a terminal version of the cgitb module).
- Auto-parentheses via the
%autocallcommand: callable objects can be executed without parentheses:
sin 3is automatically converted to
- Auto-quoting: using
;as the first character forces auto-quoting of the rest of the line:
,my_function a bbecomes automatically
;my_function a bbecomes
- Extensible input syntax. You can define filters that pre-process
user input to simplify input in special situations. This allows
for example pasting multi-line code fragments which start with
...such as those from other python sessions or the standard Python documentation.
- Flexible configuration system. It uses a configuration file which allows permanent setting of all command-line options, module loading, code and file execution. The system allows recursive file inclusion, so you can have a base file with defaults and layers which load other customizations for particular projects.
- Embeddable. You can call IPython as a python shell inside your own python programs. This can be used both for debugging code or for providing interactive abilities to your programs with knowledge about the local namespaces (very useful in debugging and data analysis situations).
- Easy debugger access. You can set IPython to call up an enhanced version of
the Python debugger (pdb) every time there is an uncaught exception. This
drops you inside the code which triggered the exception with all the data
live and it is possible to navigate the stack to rapidly isolate the source
of a bug. The
%runmagic command (with the
-doption) can run any script under pdb’s control, automatically setting initial breakpoints for you. This version of pdb has IPython-specific improvements, including tab-completion and traceback coloring support. For even easier debugger access, try
%debugafter seeing an exception.
- Profiler support. You can run single statements (similar to
profile.run()) or complete programs under the profiler’s control. While this is possible with standard cProfile or profile modules, IPython wraps this functionality with magic commands (see
%run -p) convenient for rapid interactive work.
- Simple timing information. You can use the
%timeitcommand to get the execution time of a Python statement or expression. This machinery is intelligent enough to do more repetitions for commands that finish very quickly in order to get a better estimate of their running time.
In : %timeit 1+1 10000000 loops, best of 3: 25.5 ns per loop In : %timeit [math.sin(x) for x in range(5000)] 1000 loops, best of 3: 719 µs per loop
To get the timing information for more than one expression, use the
%%timeitcell magic command.
- Doctest support. The special
%doctest_modecommand toggles a mode to use doctest-compatible prompts, so you can use IPython sessions as doctest code. By default, IPython also allows you to paste existing doctests, and strips out the leading
...prompts in them.
Decoupled two-process model¶
IPython has abstracted and extended the notion of a traditional Read-Evaluate-Print Loop (REPL) environment by decoupling the evaluation into its own process. We call this process a kernel: it receives execution instructions from clients and communicates the results back to them.
This decoupling allows us to have several clients connected to the same
kernel, and even allows clients and kernels to live on different machines.
With the exclusion of the traditional single process terminal-based IPython
(what you start if you run
ipython without any subcommands), all
other IPython machinery uses this two-process model. Most of this is now part
Jupyter project, whis includes
As an example, this means that when you start
jupyter qtconsole, you’re
really starting two processes, a kernel and a Qt-based client can send
commands to and receive results from that kernel. If there is already a kernel
running that you want to connect to, you can pass the
which will skip initiating a new kernel and connect to the most recent kernel,
instead. To connect to a specific kernel once you have several kernels
running, use the
%connect_info magic to get the unique connection file,
which will be something like
--existing kernel-19732.json but with
different numbers which correspond to the Process ID of the kernel.
Frontend/Kernel Model example notebook
Interactive parallel computing¶
This functionality is optional and now part of the ipyparallel project.
Increasingly, parallel computer hardware, such as multicore CPUs, clusters and supercomputers, is becoming ubiquitous. Over the last several years, we have developed an architecture within IPython that allows such hardware to be used quickly and easily from Python. Moreover, this architecture is designed to support interactive and collaborative parallel computing.
The main features of this system are:
- Quickly parallelize Python code from an interactive Python/IPython session.
- A flexible and dynamic process model that be deployed on anything from multicore workstations to supercomputers.
- An architecture that supports many different styles of parallelism, from message passing to task farming. And all of these styles can be handled interactively.
- Both blocking and fully asynchronous interfaces.
- High level APIs that enable many things to be parallelized in a few lines of code.
- Write parallel code that will run unchanged on everything from multicore workstations to supercomputers.
- Full integration with Message Passing libraries (MPI).
- Capabilities based security model with full encryption of network connections.
- Share live parallel jobs with other users securely. We call this collaborative parallel computing.
- Dynamically load balanced task farming system.
- Robust error handling. Python exceptions raised in parallel execution are gathered and presented to the top-level code.
For more information, see our overview of using IPython for parallel computing.
Portability and Python requirements¶
As of the 2.0 release, IPython works with Python 2.7 and 3.3 or above. Version 1.0 additionally worked with Python 2.6 and 3.2. Version 0.12 was the first version to fully support Python 3.
IPython is known to work on the following operating systems:
- Most other Unix-like OSs (AIX, Solaris, BSD, etc.)
- Mac OS X
- Windows (CygWin, XP, Vista, etc.)
See here for instructions on how to install IPython.