The IPython Notebook
The IPython Notebook combines two components:
- A web application, called the IPython Notebook web app, for interactive authoring of literate computations, in which explanatory text, mathematics, computations and rich media output may be combined. Input and output are stored in persistent cells that may be edited in-place.
- Plain text documents, called notebook documents, or notebooks, for recording and distributing the results of the rich computations.
In the documentation, the distinction between the *N*otebook app and *n*otebook documents is made by capitalization.
The Notebook app automatically saves the current state of the computation in the web browser to the corresponding notebook document.
It is also common to refer to the current state of the computation, as represented by the sequence of input cells in the Notebook app, as a notebook. There is no problem with confounding these two concepts, since there is actually a one-to-one correspondence between what you see on the
screen inside the app, and what is stored in the corresponding .ipynb notebook document.
Features of the IPython Notebook web app
Some of the main features of the IPython Notebook app include:
- Display rich data representations (e.g. HTML / LaTeX / SVG) in the browser as a result of computations.
- Compose text cells using Markdown and HTML.
- Include mathematical equations, rendered directly in the browser by MathJax.
- Import standard Python scripts
- In-browser editing, syntax highlighting, tab completion and autoindentation.
- Inline figures rendered by the matplotlib library with publication quality, in a range of formats (SVG / PDF / PNG).
If you have ever used the Mathematica or SAGE notebooks (the latter is also web-based) you should feel right at home. If you have not, you will be able to learn how to use the IPython Notebook in just a few minutes.
Notebook documents
Notebook document files are just standard text files with the extension .ipynb, stored in the working directory on your computer. This file can be easily put under version control and shared with colleagues.
Despite the fact that the notebook documents are plain text files, they use the JSON format in order to store a complete, reproducible copy of the state of the computation as it is inside the Notebook app. That is, they record all computations carried out and the results obtained in a literate way; inputs and outputs of computations can be freely mixed with descriptive text, mathematics, and HTML 5 objects.
Notebooks may easily be exported to a range of static formats, including HTML (for example, for blog posts), PDF and slide shows. Furthermore, any publicly available notebook may be shared via the IPython Notebook Viewer service, which will provide it as a static web page. The results may thus be shared without having to install anything.
See :ref:`our installation documentation <install_index>` for directions on how to install the notebook and its dependencies.
Note
You can start more than one notebook server at the same time, if you want to work on notebooks in different directories. By default the first notebook server starts on port 8888, and later notebook servers search for ports near that one. You can also manually specify the port with the --port option.
Running the IPython Notebook web app
The Notebook web app is started with the command:
$ ipython notebook
The landing page of the notebook server application, the dashboard, shows the notebooks currently available in the working directory (the directory from which the notebook was started). You can create new notebooks from the dashboard with the New Notebook button, or open existing ones by clicking on their name. You can also drag and drop .ipynb notebooks and standard .py Python source code files into the notebook list area.
.py files will be imported into the IPython Notebook as a notebook with the same name, but an .ipynb extension, located in the working directory. The notebook will consist of a single cell containing all the code in the .py file, which you can later manually partition into individual cells.
The IPython Notebook web app is based on a server-client structure. This server uses a two-process kernel architecture based on ZeroMQ, as well as Tornado for serving HTTP requests. Other clients may connect to the same underlying IPython kernel.
When you open or create a new notebook, your browser tab will reflect the name of that notebook, prefixed with "IPy". The URL is currently not meant to be human-readable and is not persistent across invocations of the notebook server; however, this will change in a future version of IPython.
Basic concepts in the Notebook app
When you finally start editing a notebook document in the Notebook, you will be presented with the title of the notebook, a menu bar, a toolbar and an empty input cell.
Notebook title
The title of the notebook document that is currently being edited is displayed at the top of the page, next to the IP[y]: Notebook logo. This title may be edited directly by clicking on it. The title is reflected in the name of the .ipynb notebook document file that is saved.
Toolbar
The tool bar gives handy icons for the most-used operations within the Notebook.
Input cells
Input cells are the core of the functionality of the IPython Notebook. They are regions in the document where you can enter different types of text and commands. These regions are then executed using :kbd:`Shift-Enter`, at which point the Notebook executes the current input cell, displays the resulting output beneath it, and adds a new input cell below.
The notebook consists of a sequence of input cells, providing the means to direct the computational process.
Basic workflow
The normal workflow in a notebook is, then, quite similar to a standard IPython session, with the difference that you can edit cells in-place multiple times until you obtain the desired results, rather than having to rerun separate scripts with the %run magic command. (Magic commands do, however, also work in the notebook; see below). Typically, you'll work on a problem in pieces, organizing related pieces into cells and moving forward as previous parts work correctly. This is much more convenient for interactive exploration than breaking up a computation into scripts that must be executed together, especially if parts of them take a long time to run
The only significant limitation that the notebook currently has, compared to the Qt console, is that it cannot run any code that expects input from the kernel (such as scripts that call :func:`raw_input`). Very importantly, this means that the %debug magic does not currently work in the notebook! This limitation will be overcome in the future, but in the meantime, there is a way to debug problems in the notebook: you can attach a Qt console to your existing notebook kernel, and run %debug from the Qt console. If your notebook is running on a local computer (i.e. if you are accessing it via your localhost address at 127.0.0.1), you can just type %qtconsole in the notebook and a Qt console will open up, connected to that same kernel.
At certain moments, it may be necessary to interrupt a particularly long calculation, or even to kill the entire computational process. This may be achieved by interrupting or restarting the kernel, respectively. After a restart, all relevant cells must be re-evaluated
Cell types
Each IPython input cell has a cell type. There is a restricted number of possible cell types, which may be set by using the cell type dropdown on the toolbar, or via the following keyboard shortcuts:
- code :kbd:`Ctrl-m y`
- markdown :kbd:`Ctrl-m m`
- raw :kbd:`Ctrl-m t`
- heading :kbd:`Ctrl-m 1` - :kbd:`Ctrl-m 6`
Code cells
Code cells contain code written in some computer language, which is Python by default. When the cell is executed with :kbd:`Shift-Enter`, this code is executed, and the result returned by Python (or the corresponding language) after running the code will be displayed as its output.
Code may be edited inline in the cell, with full syntax highlighting.
Rich text using Markdown
The computational process may be documented in a literate way using rich text. For this purpose, the Notebook provides markdown cells. Text is entered using Markdown syntax, allowing for italics, bold, ordered and unordered lists, etc. This is rendered using Markdown syntax to a rich HTML representation when the cell is executed. In this case, the output replaces the input cell.
Within markdown cells, mathematics can be included in a straightforward manner using LaTeX notation: $...$ for inline math and $$...$$ for displayed math. Standard LaTeX environments, such as \begin{equation}...\end{equation}, also work. New commands may be defined using standard LaTeX commands, placed anywhere in a markdown cell.
Raw cells
Raw cells provide a place to put additional information which is not evaluated by the Notebook. This can be used, for example, for extra information to be used when the notebook is exported to a certain format.
Magic commands
Magic commands, or magics, are one-word commands starting with %%, which give commands to IPython itself (as opposed to standard Python commands which are exported to be run in a Python interpreter).
Magics control different elements of the way that the IPython notebook operates. They are entered into standard code cells and executed as usual with :kbd:`Shift-Enter`.
Some of the main magics are the following:
- %lsmagic Lists all the magic commands available
- %config Configuration of IPython
- %timeit Time a single line of code.
There are also cell magics, which start with %%
Plotting
One major feature of the Notebook is the ability to capture the result of plots as inline output. IPython is designed to work seamlessly together with the %matplotlib plotting library. In order to set this up, the %matplotlib magic command must be run before any plotting takes place.
Note that %matplotlib only sets up IPython to work correctly with matplotlib; it does not actually execute any import commands and does not add anything to the namespace.
There is an alternative magic, %pylab, which, in addition, also executes a sequence of standard import statements required for working with the %matplotlib library. In particular, it automatically imports all names in the numpy and matplotlib packages to the namespace. A less invasive solution is %pylab --no-import-all, which imports just the standard names np for the numpy module and plt for the matplotlib.pyplot module.
- When the default %matplotlib or %pylab magics are used, the output of a plotting command is captured in a separate window. An alternative is to use::
- %matplotlib inline
which captures the output inline within the notebook format. This has the benefit that the resulting plots will be stored in the notebook document.
Exporting a notebook and importing existing scripts
Need to talk about ipython nbconvert
If you want to provide others with a static HTML or PDF view of your notebook, use the Print button. This opens a static view of the document, which you can print to PDF using your operating system's facilities, or save to a file with your web browser's 'Save' option (note that typically, this will create both an html file and a directory called notebook_name_files next to it that contains all the necessary style information, so if you intend to share this, you must send the directory along with the main html file).
The Download button lets you save a notebook file to the Download area configured by your web browser (particularly useful if you are running the notebook server on a remote host and need a file locally). But you can always export the input part of a notebook to a plain python script by choosing Python format in the Download drop list. This removes all output and saves the text cells in comment areas. See ref:below <notebook_format> for more details on the notebook format.
Warning
While in simple cases you can roundtrip a notebook to Python, edit the python file and import it back without loss of main content, this is in general not guaranteed to work at all. First, there is extra metadata saved in the notebook that may not be saved to the .py format. And as the notebook format evolves in complexity, there will be attributes of the notebook that will not survive a roundtrip through the Python form. You should think of the Python format as a way to output a script version of a notebook and the import capabilities as a way to load existing code to get a notebook started. But the Python version is not an alternate notebook format.
Importing or executing a notebook as a normal Python file
The native format of the notebook, a file with a .ipynb `extension, is a JSON container of all the input and output of the notebook, and therefore not valid Python by itself. This means that by default, you cannot directly import a notebook from Python, nor execute it as a normal python script.
But if you want to be able to use notebooks also as regular Python files, you can start the notebook server with:
ipython notebook --script
or you can set this option permanently in your configuration file with:
c.NotebookManager.save_script=True
This will instruct the notebook server to save the .py export of each notebook, in addition to the .ipynb, at every save. These are standard .py files, and so they can be %run, imported from regular IPython sessions or other notebooks, or executed at the command-line. Since we export the raw code you have typed, for these files to be importable from other code you will have to avoid using syntax such as %magics and other IPython-specific extensions to the language.
In regular practice, the standard way to differentiate importable code from the 'executable' part of a script is to put at the bottom:
if __name__ == '__main__': # rest of the code...
Since all cells in the notebook are run as top-level code, you'll need to similarly protect all cells that you do not want executed when other scripts try to import your notebook. A convenient shortand for this is to define early on:
script = __name__ == '__main__'
and then on any cell that you need to protect, use:
if script: # rest of the cell...
Configuration
The IPython notebook server can be run with a variety of command line arguments. To see a list of available options enter:
$ ipython notebook --help
Defaults for these options can also be set by creating a file named ipython_notebook_config.py in your IPython profile folder. The profile folder is a subfolder of your IPython directory; ipython locate will show you where it is located.
To create a new set of default configuration files, with lots of information on available options, use:
$ ipython profile create
Keyboard shortcuts
All actions in the notebook can be achieved with the mouse, but we have also added keyboard shortcuts for the most common ones, so that productive use of the notebook can be achieved with minimal mouse intervention. The main key bindings you need to remember are:
- :kbd:`Shift-Enter`: execute the current cell (similar to the Qt console), show output (if any) and jump to the next cell below. If :kbd:`Shift-Enter` was invoked on the last input line, a new code cell will also be created. Note that in the notebook, simply using :kbd:`Enter` never forces execution, it simply inserts a new line in the current cell. Therefore, in the notebook you must always use :kbd:`Shift-Enter` to get execution (or use the mouse and click on the Run Selected button).
- :kbd:`Alt-Enter`: this combination is similar to the previous one, with the exception that, if the next cell below is not empty, a new code cell will be added to the notebook, even if the cell execution happens not in the last cell. In this regard, :kbd:`Alt-Enter`: is simply a shortcut for the :kbd:`Shift-Enter`, :kbd:`Ctrl-m a` sequence.
- :kbd:`Ctrl-Enter`: execute the current cell in "terminal mode", where any output is shown but the cursor stays in the current cell, whose input area is flushed empty. This is convenient to do quick in-place experiments or query things like filesystem content without creating additional cells you may not want saved in your notebook.
- :kbd:`Ctrl-m`: this is the prefix for all other keybindings, which consist of an additional single letter. Type :kbd:`Ctrl-m h` (that is, the sole letter :kbd:`h` after :kbd:`Ctrl-m`) and IPython will show you the remaining available keybindings.
Security
You can protect your notebook server with a simple single-password by setting the :attr:`NotebookApp.password` configurable. You can prepare a hashed password using the function :func:`IPython.lib.security.passwd`:
In [1]: from IPython.lib import passwd In [2]: passwd() Enter password: Verify password: Out[2]: 'sha1:67c9e60bb8b6:9ffede0825894254b2e042ea597d771089e11aed'
Note
:func:`~IPython.lib.security.passwd` can also take the password as a string argument. Do not pass it as an argument inside an IPython session, as it will be saved in your input history.
You can then add this to your :file:`ipython_notebook_config.py`, e.g.:
# Password to use for web authentication c.NotebookApp.password = u'sha1:67c9e60bb8b6:9ffede0825894254b2e042ea597d771089e11aed'
When using a password, it is a good idea to also use SSL, so that your password is not sent unencrypted by your browser. You can start the notebook to communicate via a secure protocol mode using a self-signed certificate by typing:
$ ipython notebook --certfile=mycert.pem
Note
A self-signed certificate can be generated with openssl. For example, the following command will create a certificate valid for 365 days with both the key and certificate data written to the same file:
$ openssl req -x509 -nodes -days 365 -newkey rsa:1024 -keyout mycert.pem -out mycert.pem
Your browser will warn you of a dangerous certificate because it is self-signed. If you want to have a fully compliant certificate that will not raise warnings, it is possible (but rather involved) to obtain one for free, as explained in detailed in this tutorial.
Keep in mind that when you enable SSL support, you'll need to access the notebook server over https://, not over plain http://. The startup message from the server prints this, but it's easy to overlook and think the server is for some reason non-responsive.
Quick how to's
Connecting to an existing kernel
The notebook server always prints to the terminal the full details of how to connect to each kernel, with lines like:
[IPKernelApp] To connect another client to this kernel, use: [IPKernelApp] --existing kernel-3bb93edd-6b5a-455c-99c8-3b658f45dde5.json
This is the name of a JSON file that contains all the port and validation information necessary to connect to the kernel. You can manually start a Qt console with:
ipython qtconsole --existing kernel-3bb93edd-6b5a-455c-99c8-3b658f45dde5.json
and if you only have a single kernel running, simply typing:
ipython qtconsole --existing
will automatically find it (it will always find the most recently started kernel if there is more than one). You can also request this connection data by typing %connect_info; this will print the same file information as well as the content of the JSON data structure it contains.
Running a public notebook server
If you want to access your notebook server remotely with just a web browser, here is a quick set of instructions. Start by creating a certificate file and a hashed password as explained above. Then, create a custom profile for the notebook. At the command line, type:
ipython profile create nbserver
In the profile directory, edit the file ipython_notebook_config.py. By default the file has all fields commented, the minimum set you need to uncomment and edit is here:
c = get_config() # Kernel config c.IPKernelApp.pylab = 'inline' # if you want plotting support always # Notebook config c.NotebookApp.certfile = u'/absolute/path/to/your/certificate/mycert.pem' c.NotebookApp.ip = '*' c.NotebookApp.open_browser = False c.NotebookApp.password = u'sha1:bcd259ccf...your hashed password here' # It's a good idea to put it on a known, fixed port c.NotebookApp.port = 9999
You can then start the notebook and access it later by pointing your browser to https://your.host.com:9999 with ipython notebook --profile=nbserver.
Running with a different URL prefix
The notebook dashboard (i.e. the default landing page with an overview of all your notebooks) typically lives at a URL path of "http://localhost:8888/". If you want to have it, and the rest of the notebook, live under a sub-directory, e.g. "http://localhost:8888/ipython/", you can do so with configuration options like these (see above for instructions about modifying ipython_notebook_config.py):
c.NotebookApp.base_project_url = '/ipython/' c.NotebookApp.base_kernel_url = '/ipython/' c.NotebookApp.webapp_settings = {'static_url_prefix':'/ipython/static/'}
Using a different notebook store
By default the notebook server stores notebooks as files in the working directory of the notebook server, also known as the notebook_dir. This logic is implemented in the :class:`FileNotebookManager` class. However, the server can be configured to use a different notebook manager class, which can store the notebooks in a different format. Currently, we ship a :class:`AzureNotebookManager` class that stores notebooks in Azure blob storage. This can be used by adding the following lines to your ipython_notebook_config.py file:
c.NotebookApp.notebook_manager_class = 'IPython.html.services.notebooks.azurenbmanager.AzureNotebookManager' c.AzureNotebookManager.account_name = u'paste_your_account_name_here' c.AzureNotebookManager.account_key = u'paste_your_account_key_here' c.AzureNotebookManager.container = u'notebooks'
In addition to providing your Azure Blob Storage account name and key, you will have to provide a container name; you can use multiple containers to organize your Notebooks.
The notebook format
The notebooks themselves are JSON files with an ipynb extension, formatted as legibly as possible with minimal extra indentation and cell content broken across lines to make them reasonably friendly to use in version-control workflows. You should be very careful if you ever edit manually this JSON data, as it is extremely easy to corrupt its internal structure and make the file impossible to load. In general, you should consider the notebook as a file meant only to be edited by IPython itself, not for hand-editing.
Note
Binary data such as figures are directly saved in the JSON file. This provides convenient single-file portability but means the files can be large and diffs of binary data aren't very meaningful. Since the binary blobs are encoded in a single line they only affect one line of the diff output, but they are typically very long lines. You can use the 'ClearAll' button to remove all output from a notebook prior to committing it to version control, if this is a concern.
The notebook server can also generate a pure-python version of your notebook, by clicking on the 'Download' button and selecting py as the format. This file will contain all the code cells from your notebook verbatim, and all text cells prepended with a comment marker. The separation between code and text cells is indicated with special comments and there is a header indicating the format version. All output is stripped out when exporting to python.
Here is an example of a simple notebook with one text cell and one code input cell, when exported to python format:
# <nbformat>2</nbformat> # <markdowncell> # A text cell # <codecell> print "hello IPython"
Known issues
When behind a proxy, especially if your system or browser is set to autodetect the proxy, the html notebook might fail to connect to the server's websockets, and present you with a warning at startup. In this case, you need to configure your system not to use the proxy for the server's address.
In Firefox, for example, go to the Preferences panel, Advanced section, Network tab, click 'Settings...', and add the address of the notebook server to the 'No proxy for' field.