Elpy

Elpy is the Emacs Python Development Environment. It aims to provide an easy to install, fully-featured environment for Python development.

Introduction

Overview

Elpy is an extension for the Emacs text editor to work with Python projects. This documentation explains how to use Elpy to work on Python project using Emacs, but it does not aim to be an introduction to either Emacs or Python.

You can read a quick tour of Emacs, or read the built-in tutorial by running C-h t in the editor. That is, you hold down the control key and hit h (the canonical help key in Emacs), release both, and hit t (for tutorial).

For Python, you can read the basic tutorial. If you already know Python, you should check out some best practices.

Once you have these basics, you can go on to install Elpy.

Installation

With use-package

Simply add the following lines to you .emacs (or .emacs.d/init.el, or equivalent):

(use-package elpy
  :ensure t
  :init
  (elpy-enable))

Or if you want to defer Elpy loading:

(use-package elpy
  :ensure t
  :defer t
  :init
  (advice-add 'python-mode :before 'elpy-enable))

From apt (Debian ≥10 an Ubuntu ≥18.10)

Users of Debian ≥10 or Ubuntu ≥18.10 can skip the instructions above this line and may simply install Elpy and all of its recommended dependencies with the following command::

sudo apt install elpa-elpy

Elpy can then be activated by running M-x elpy-enable. This can be made automatic by adding the following to your .emacs:

(elpy-enable)

In order to use all the features (such as navigation with M-.), you’ll need to install some python libraries::

sudo apt install python3-jedi black autopep8 yapf

Manually from Melpa

The main Elpy package is installed via the Emacs package interface, package.el. First, you have to add Elpy’s package archive to your list of archives, though. Add the following code to your .emacs file and restart Emacs:

(require 'package)
(add-to-list 'package-archives
             '("melpa-stable" . "https://stable.melpa.org/packages/"))

Now you can run M-x package-refresh-contents to download a fresh copy of the archive contents, and M-x package-install RET elpy RET to install elpy. If you want to enable Elpy by default, you can simply add the following to your .emacs:

(package-initialize)
(elpy-enable)

Congratulations, Elpy is now successfully installed!

Manually from sources

If you want to test a specific branch or a PR that has not been merged yet, you will have to install Elpy from sources.

  • First, clone the repository containing the branch you are interested in:
git clone https://github.com/USER/elpy.git
cd elpy

Then check out the branch associated to the PR you want to try:

git checkout BRANCH_NAME
  • Now switch to Emacs and add Elpy’s sources to the load path:
(add-to-list 'load-path "/path/to/elpy")
  • Install Elpy’s dependencies:
    • Company
    • highlight-indentation
    • pyvenv
    • yasnippet
    • s
  • Finally, load Elpy files:
(load "elpy")
(load "elpy-rpc")
(load "elpy-shell")
(load "elpy-profile")
(load "elpy-refactor")
(load "elpy-django")

Configuration

You can easily configure Elpy to your own preferences. All options are available through a unified interface, accessible with the following command:

M-x elpy-config

Show the current Elpy configuration, point out possible problems, and provide a quick interface to relevant customization options.

Packages needed by Elpy can be installed and updated from this interface.

Quickstart

Once installed, Elpy will automatically provide code completion, syntax error highlighting and function signature (in the modeline) for python files.

Useful keybindings

Elpy has quite a lot of keybindings, but the following ones should be enough to get you started:

C-c C-c (elpy-shell-send-region-or-buffer)

Evaluate the current script (or region if something is selected) in an interactive python shell. The python shell is automatically displayed aside of your script.

C-RET (elpy-shell-send-statement-and-step)

Evaluate the current statement (current line plus the following nested lines). Useful for evaluating a function or class definition or a for loop.

C-c C-z (elpy-shell-switch-to-shell)

Switch between your script and the interactive shell.

C-c C-d (elpy-doc)

Display documentation for the thing under cursor (function or module). The documentation will pop in a different buffer, that can be closed with q.

Go further

Elpy offers a lot of features, including code navigation, debugging, testing, profiling and support for virtual environments. Feel free to explore the documentation, everything is there !

Editing

Emacs Basics

Elpy is an extension to Emacs, and as such the standard bindings in Emacs are available. This manual is not meant to be an introduction to Emacs, but this section will still highlight some features in Emacs that are especially useful for Python editing.

Movement keys in Emacs often use fbnp for forward, backward, next (down) and previous (up). k and backspace (DEL) are for deleting. These are combined with the Control, Meta and Control-Meta modifiers. Control generally refers to the simplest form. C-f moves one character forward. Meta changes this to affect words, that is, consecutive sequences of alphanumeric characters. The Control-Meta combination then affects whole expressions.

In the following table, | refers to the position of point.

Before Key After
|hello_world C-f h|ello_world
|hello_world M-f hello|_world
|hello_world C-M-f hello_world|

Expression-based commands will also work on strings, tuples, dictionaries, or any balanced groups of parentheses. This works for all movement keys (f, b, n, p), with next and previous moving to the next or previous group of parens. It also works with forward and backward deletion (d and DEL/<backspace>, respectively) for character and word groups, but not for expressions. To delete the expression after point, use C-M-k. For the expression before point, you can use C-M-b C-M-k.

If you enable subword-mode, Emacs will also consider CamelCase to be two words instead of one for the purpose of these operations.

In addition to the above, Emacs also supports moving up or down inside nested parentheses groups. C-M-d will move down into the next enclosed group of parentheses, while C-M-u will move up to the directly enclosing group of parentheses.

Finally, a lot of Elpy’s commands change their behavior when the prefix argument is given. That is, hit C-u before the command. In Elpy, the prefix argument often disables any attempt by the command at being smart, in case it would get it wrong.

Moving By Indentation

C-down (elpy-nav-forward-block)
C-up (elpy-nav-backward-block)

These commands are used to navigate between lines with same indentation as the current line. Point should be placed on the first non-whitespace character of the line and then use C-down to move forward or C-up to move backward.

C-left (elpy-nav-backward-indent)
C-right (elpy-nav-forward-indent)

These commands are used to navigate between indentation levels. C-left moves point to previous indent level or over previous word. C-right moves point to next indent level or over the next word.

Moving the Current Region

M-down (elpy-nav-move-line-or-region-down)
M-up (elpy-nav-move-line-or-region-up)
M-left (elpy-nav-indent-shift-left)
M-right (elpy-nav-indent-shift-right)

Elpy can move the selected region (or the current line if no region is selected) by using the cursor keys with meta. Left and right will dedent or indent the code, while up and down will move it line-wise up or down, respectively.

IDE Features

Projects

Elpy supports the notion of projects, a related collection of files under a common directory. This common directory is called the project root. A number of Elpy’s commands work on all files inside the project root.

C-c C-f (elpy-find-file)

Find a file in the current project. This uses a search-as-you-type interface for all files under the project root.

A prefix argument enables “do what I mean” mode. On an import statement, it will try to open the module imported. Elsewhere in a file, it will look for an associated test or implementation file, and if found, open that. If this fails, either way, it will fall back to the normal find file in project behavior.

If the current file is called foo.py, then this will search for a test_foo.py in the same directory, or in a test or tests subdirectory. If the current file is already called test_foo.py, it will try and find a foo.py nearby.

This command uses projectile or find-file-in-project under the hood, so you need one of them to be installed.

C-c C-s (elpy-rgrep-symbol)

Search the files in the current project for a string. By default, this uses the symbol at point. With a prefix argument, it will prompt for a regular expression to search.

This is basically a grep -r through the project.

In addition to these two commands, elpy-check also supports optionally checking all files in the current project.

Elpy’s idea of the project root and which files belong to a project and which don’t can be influenced as well.

M-x elpy-set-project-root

Set the current project root directory. This directory should contain all files related to the current project.

elpy-project-ignored-directories (Customize Option)

When Elpy searches for files in the current project, it will ignore files in directories listed here.

elpy-project-root-finder-functions (Customize Option)

To find the project root, Elpy can utilize a number of heuristics. With this option, you can configure which are used.

To configure Elpy specifically for a single project, you can use Emacs’ Directory Variables. Elpy provides a simple interface to this.

M-x elpy-set-project-variable

Set or change the value of a project-wide variable. With a prefix argument, the value for the variable is removed.

This only takes effect in new buffers.

Completion

When you type Python code, Elpy will try and figure out possible completions and provide them in a suggestion window. If Elpy doesn’t do so automatically, you can force it to complete right where you are.

M-TAB (elpy-company-backend)

Provide completion suggestions for a completion at point.

You can use cursor keys or M-n and M-p to scroll through the options, RET to use the selected completion, or TAB to complete the common part.

On any completion option, C-d or <f1> will display a temporary window with documentation. C-w will display a temporary window showing the source code of the completion to get some context.

Elpy uses Company Mode for the completion interface, so its documentation is a good place for further information.

elpy-get-info-from-shell (Customize Option)

If t, use the shell to gather docstrings and completions. Normally elpy provides completion and documentation using static code analysis (from jedi). With this option, elpy will add the completion candidates and the docstrings from the associated python shell; This activates fallback completion candidates for cases when the static code analysis fails.

Interactive Python

Emacs can run a Python interpreter in a special buffer, making it much easier to send code snippets over. Elpy provides additional functionality to seamlessly work with interactive Python in a style similar to ESS.

Interpreter Setup

Elpy uses the Python interpreter setup from the Emacs python package. This section briefly summarizes some common setups; add the one you need to your .emacs file. Note that the code below (and Elpy in general) require at least Emacs 24.4.

Use the Python standard interpreter (default):

(setq python-shell-interpreter "python"
      python-shell-interpreter-args "-i")

Use Jupyter console (recommended for interactive Python):

(setq python-shell-interpreter "jupyter"
      python-shell-interpreter-args "console --simple-prompt"
      python-shell-prompt-detect-failure-warning nil)
(add-to-list 'python-shell-completion-native-disabled-interpreters
             "jupyter")

Use IPython:

(setq python-shell-interpreter "ipython"
      python-shell-interpreter-args "-i --simple-prompt")

Note that various issues with plotting have been reported when running IPython 5 in Emacs under Windows. We recommend using Jupyter console instead.

If you have an older version of IPython and the above code does not work for you, you may also try:

(setenv "IPY_TEST_SIMPLE_PROMPT" "1")
(setq python-shell-interpreter "ipython"
      python-shell-interpreter-args "-i")

As an IPython_ user, you might be interested in the `Emacs IPython
Notebook`_ or an `Elpy layer`_ for Spacemacs_, too.

Note for MacOS users: In some configurations, display artifacts (lines of ^G s) can appear in the shell. This can be fixed by prepending -c exec('__import__(\\'readline\\')') to python-shell-interpeter-args (e.g. (setq python-shell-interpreter-args "-c exec('__import__(\\'readline\\')') -i") for python)

The Shell Buffer

C-c C-z (elpy-shell-switch-to-shell)

Switch to buffer with a Python interpreter running, starting one if necessary.

By default, Elpy tries to find the root directory of the current project (git, svn or hg repository, python package or projectile project) and starts the python interpreter here. This behaviour can be suppressed with the option elpy-shell-starting-directory.

elpy-shell-starting-directory (Customize Option)

Govern the directory in which Python shells will be started. Can be 'project-root (default) to use the current project root, 'current-directory to use the buffer current directory, or a string indicating a specific path.

M-x elpy-shell-toggle-dedicated-shell

By default, python buffers are all attached to a same python shell (that lies in the *Python* buffer), meaning that all buffers and code fragments will be send to this shell. elpy-shell-toggle-dedicated-shell attaches a dedicated python shell (not shared with the other python buffers) to the current python buffer. To make this the default behavior (like the deprecated option elpy-dedicated-shells did), use the following snippet:

(add-hook 'elpy-mode-hook (lambda () (elpy-shell-toggle-dedicated-shell 1)))
M-x elpy-shell-set-local-shell

Attach the current python buffer to a specific python shell (whose name is asked with completion). You can use this function to have one python shell per project, with:

(add-hook 'elpy-mode-hook (lambda () (elpy-shell-set-local-shell (elpy-project-root))))
C-c C-k (elpy-shell-kill)

Kill the associated python shell.

C-c C-K (elpy-shell-kill-all)

Kill all active python shells.

Evaluating code fragments

Elpy provides commands to send the current Python statement (e), function definition (f), class definition (c), top-level statement (s), group of Python statements (g), cell (w), region (r), or buffer (b) to the Python shell for evaluation. These commands are bound to prefix C-c C-y, followed by the single character indicating what to send; e.g., C-c C-y e sends the Python statement at point.

Each of the commands to send code fragments to the shell has four variants, one for each combination of: whether or not the point should move after sending (“step”), and whether or not the Python shell should be focused after sending (“go”). Step is activated by C-, go by S-. For example:

C-c C-y e (elpy-shell-send-statement)

Send the current statement to the Python shell and keep point position. Here statement refers to the Python statement the point is on, including potentially nested statements. If point is on an if/elif/else clause send the entire if statement (with all its elif/else clauses). If point is on a decorated function, send the decorator as well.

C-c C-y C-e (elpy-shell-send-statement-and-step)

Send the current statement to the Python shell and move point to first subsequent statement.

Also bound to C-RET.

C-c C-y E (elpy-shell-send-statement-and-go)

Send the current statement to the Python shell, keeping point position, and switch focus to the Python shell buffer.

C-c C-y C-S-E (elpy-shell-send-statement-and-step-and-go)

Send the current statement to the Python shell, move point to first subsequent statement, and switch focus to the Python shell buffer.

Elpy provides support for sending multiple statements to the shell.

C-c C-y O (elpy-shell-send-group-and-step)

Send the current or next group of top-level statements to the Python shell and step. A sequence of top-level statements is a group if they are not separated by empty lines. Empty lines within each top-level statement are ignored.

If the point is within a statement, send the group around this statement. Otherwise, go to the top-level statement below point and send the group around this statement.

C-c C-y W (elpy-shell-send-codecell-and-step)

Send the current code cell to the Python shell and step. A code cell is a piece of code surrounded by special separator lines; see below. For example, you can insert two lines starting with ## to quickly send the code in-between.

elpy-shell-codecell-beginning-regexp (Customize Option)

Regular expression for matching a line indicating the beginning of a code cell. By default, ##.* is treated as a beginning of a code cell, as are the code cell beginnings in Python files exported from IPython or Jupyter notebooks (e.g., # <codecell> or # In[1]:).

elpy-shell-cell-boundary-regexp (Customize Option)

Regular expression for matching a line indicating the boundary of a cell (beginning or ending). By default, ##.* is treated as a cell boundary, as are the boundaries in Python files exported from IPython or Jupyter notebooks (e.g., # <markdowncell>, # In[1]:, or # Out[1]:).

Note that elpy-shell-codecell-beginning-regexp must also match the cell boundaries defined here.

The functions for sending the entire buffer have special support for avoiding accidental code execution, e.g.:

C-c C-y r (elpy-shell-send-region-or-buffer)

Send the the active region (if any) or the entire buffer (otherwise) to the Python shell and keep point position.

When sending the whole buffer, this command will also escape any uses of the if __name__ == '__main__' idiom, to prevent accidental execution of a script. If you want this to be evaluated, pass a prefix argument with C-u.

Also bound to C-c C-c.

elpy-shell-add-to-shell-history (Customize Option)

If t, Elpy will make the code sent available in the shell history. This allows to use comint-previous-input (C-up) in the python shell to get back the pieces of code sent by Elpy.

The list of remaining commands to send code fragments is:

C-c C-y s (elpy-shell-send-top-statement)
C-c C-y S (elpy-shell-send-top-statement-and-go)
C-c C-y f (elpy-shell-send-defun)
C-c C-y F (elpy-shell-send-defun-and-go)
C-c C-y c (elpy-shell-send-defclass)
C-c C-y C (elpy-shell-send-defclass-and-go)
C-c C-y o (elpy-shell-send-group)
C-c C-y O (elpy-shell-send-group-and-go)
C-c C-y w (elpy-shell-send-codecell)
C-c C-y W (elpy-shell-send-codecell-and-go)
C-c C-y R (elpy-shell-send-region-or-buffer-and-go)
C-c C-y b (elpy-shell-send-buffer)
C-c C-y B (elpy-shell-send-buffer-and-go)
C-c C-y C-s (elpy-shell-send-top-statement-and-step)
C-c C-y C-S-S (elpy-shell-send-top-statement-and-step-and-go)
C-c C-y C-f (elpy-shell-send-defun-and-step)
C-c C-y C-S-F (elpy-shell-send-defun-and-step-and-go)
C-c C-y C-c (elpy-shell-send-defclass-and-step)
C-c C-y C-S-C (elpy-shell-send-defclass-and-step-and-go)
C-c C-y C-S-O (elpy-shell-send-group-and-step-and-go)
C-c C-y C-W (elpy-shell-send-codecell-and-step-and-go)
C-c C-y C-r (elpy-shell-send-region-or-buffer-and-step)
C-c C-y C-S-R (elpy-shell-send-region-or-buffer-and-step-and-go)
C-c C-y C-b (elpy-shell-send-buffer-and-step)
C-c C-y C-S-B (elpy-shell-send-buffer-and-step-and-go)

Shell feedback

When package eval-sexp-fu is loaded and eval-sexp-fu-flash-mode is active, the statements sent to the shell are briefly flashed after running an evaluation command, thereby providing visual feedback.

elpy-shell-echo-input (Customize Option)

Whenever a code fragment is sent to the Python shell, Elpy prints it in the Python shell buffer (i.e., it looks as if it was actually typed into the shell). This behavior can be turned on and off via the custom variable elpy-shell-echo-input and further customized via elpy-shell-echo-input-cont-prompt (whether to show continuation prompts for multi-line inputs) and elpy-shell-echo-input-lines-head / elpy-shell-echo-input-lines-tail (how much to cut when input is long).

elpy-shell-echo-output (Customize Option)

Elpy shows the output produced by a code fragment sent to the shell in the echo area when the shell buffer is currently invisible. This behavior can be controlled via elpy-shell-echo-output (never, always, or only when shell invisible). Output echoing is particularly useful if the custom variable elpy-shell-display-buffer-after-send is set to nil (the default value). Then, no window is needed to display the shell (thereby saving screen real estate), but the outputs can still be seen in the echo area.

elpy-shell-display-buffer-after-send (Customize Option)

Whether to display the Python shell after sending something to it (default nil).

Syntax Checking

Whenever you save a file, Elpy will run a syntax check and highlight possible errors or warnings inline.

C-c C-n (elpy-flymake-next-error)
C-c C-p (elpy-flymake-previous-error)

You can navigate between any error messages with these keys. The current error will be shown in the minibuffer.

Elpy uses the built-in Flymake library to find syntax errors on the fly, so see there for more configuration options.

C-c C-v (elpy-check)

Alternatively, you can run a syntax check on the current file where the output is displayed in a new buffer, giving you an overview and allowing you to jump to the errors from there.

With a prefix argument, this will run the syntax check on all files in the current project.

python-check-command (Customize Option)

To change which command is used for syntax checks, you can customize this option. By default, Elpy uses the flake8 program, which you have to install separately. The elpy-config command will prompt you to do this if Elpy can’t find the program.

It is possible to create a single virtual env for the sole purpose of installing flake8 in there, and then simply link the command script to a directory inside your PATH, meaning you do not need to install the program in every virtual env separately.

Documentation

Elpy provides a single interface to documentation.

C-c C-d (elpy-doc)

When point is on a symbol, Elpy will try and find the documentation for that object, and display that. If it can’t find the documentation for whatever reason, it will try and look up the symbol at point in pydoc. If it’s not there, either, it will prompt the user for a string to look up in pydoc.

With a prefix argument, Elpy will skip all the guessing and just prompt the user for a string to look up in pydoc.

If the autodoc module is enabled (not by default) the documentation is automatically updated with the symbol at point or the currently selected company candidate.

elpy-autodoc-delay (Customize Option)

The idle delay in seconds until documentation is updated automatically.

Snippets

Elpy uses yasnippet to provide code templates that helps writing common pieces of code faster. You can access a template by typing a “template key” and hitting TAB to expand it. You may then be asked to fill some fields in the template, just hit TAB when you are done to proceed to the next field.

The sequel presents a list of templates provided by Elpy.

Special methods

Elpy provides a large range of templates for special class methods (__init__, __call__, __add__, …). To call them, simply type _ followed by the method name (for example _init) and hit TAB.

Other useful snippets

  • enc: Encoding statement
  • env: Hashbang statement
  • from: Import statement
  • pdb: Snippet to run pdb at the current script position
  • super: Super statement to call parent methods
  • def: Function definition
  • class: Class definition
  • defs: Class method definition

Folding

Elpy offers code folding by enhancing the builtin folding minor mode Hideshow.

When opening a python buffer, Elpy will indicate foldable things with an arrow in the left fringe. Clicking on an arrow will fold the corresponding code blocks. Folded code blocks can be unfolded by clicking on the button at the end of the line.

If you don’t want to use your mouse, you can achieve the same thing with the function

C-c @ C-c (elpy-folding-toggle-at-point)

Toggle folding for the thing at point, it can be a docstring, a comment or a code block.

The display of arrows in the fringe can be disable with the option

elpy-folding-fringe-indicators (Customize Option)

If elpy should display folding fringe indicators or not.

Elpy also provides some other useful features:

C-c @ C-b (elpy-folding-toggle-docstrings)

Toggle folding of all python docstrings.

C-c @ C-m (elpy-folding-toggle-comments)

Toggle folding of all comments.

C-c @ C-f (elpy-folding-hide-leafs)

Hide all code leafs, i.e. code blocks that do not contains any other blocks.

The classical keybindings for hideshow are also available, e.g.:

C-c @ C-a (hs-show-all)

Unfold everything.

(see Hideshow documentation for more information)

Debugging

Elpy provides an interface to pdb, the builtin Python debugger. Note that this interface is only available for Emacs 25 and above.

C-c C-u d (elpy-pdb-debug-buffer)

Run pdb on the current buffer. If no breakpoints has been set using elpy-pdb-toggle-breakpoint-at-point, the debugger will pause at the beginning of the buffer. Else, the debugger will pause at the first breakpoint. Once pdb is started, the pdb commands can be used to step through and look into the code evaluation.

With a prefix argument C-u, ignore the breakpoints and always pause at the beginning of the buffer.

C-c C-u b (elpy-pdb-toggle-breakpoint-at-point)

Add (or remove) a breakpoint on the current line. Elpy adds a red circle to the fringe to indicate the presence of a breakpoint. You can then use elpy-pdb-debug-buffer to start pdb and pause at each of the breakpoints.

With a prefix argument C-u, remove all the breakpoints.

C-c C-u p (elpy-pdb-break-at-point)

Run pdb on the current buffer and pause at the cursor position.

C-c C-u e (elpy-pdb-debug-last-exception)

Run post-mortem pdb on the last exception.

Testing

Testing is an important part of programming. Elpy provides a central interface to testing, which allows for a good workflow for tests.

Elpy’s test interface is built around Emacs’ compilation framework. Elpy will run test commands as a compilation job, with all the advantages this brings.

C-c C-t (elpy-test)

Start a test run. This uses the currently configured test runner to discover and run tests. If point is inside a test case, the test runner will run exactly that test case. Otherwise, or if a prefix argument is given, it will run all tests.

M-x elpy-set-test-runner

This changes the current test runner. Elpy supports the standard unittest discovery runner, the Django discovery runner, nose, green, py.test and Twisted trial. You can also write your own, as described in Writing Test Runners.

Note on Django runners: Elpy tries to find manage.py within your project structure. If it’s unable to find it, it falls back to django-admin.py. You must set the environment variable DJANGO_SETTINGS_MODULE accordingly.

This enables a good workflow. You write a test and use C-c C-t to watch it fail. You then go to your implementation file, for example using C-u C-c C-f, and make the test pass. You can use a key bound to recompile (I use <f5> for this) to just re-run that one test. Once that passes, you can use C-c C-t again to run all tests to make sure they all pass as well. Repeat.

For an even more automated way, you can use tdd.el, which will run your last compile command whenever you save a file.

Refactoring

C-c C-e (elpy-multiedit-python-symbol-at-point)

Edit all occurrences of the symbol at point at once. This will highlight all such occurrences, and editing one of them will edit all. This is an easy way to rename identifiers.

If the backend does not support finding occurrences (currently only Jedi does), or if a prefix argument is given, this will edit syntactic occurrences instead of semantic ones. This can match more occurrences than it should, so be careful. You can narrow the current buffer to the current function using C-x n d to restrict where this matches.

Finally, if there is a region active, Elpy will edit all occurrences of the text in the region.

C-c C-r f (elpy-format-code)

Format code using the available formatter.

If a region is selected, only that region is formatted. Otherwise current buffer is formatted.

elpy-formatter (Customize Option)

Allow to select the formatter you want to use amongst yapf , autopep8 and black.

yapf and autopep8 can be configured with style files placed in the project root directory (determined by elpy-project-root). black can be configured in the pyproject.toml file of your project.

Profiling

Elpy allows one to profile python scripts asynchronously using cProfile.

M-x elpy-profile-buffer-or-region

Send the current buffer or region to the profiler and display the result with elpy-profile-visualizer. The default visualizer is snakeviz, a browser-based graphical profile viewer that can be installed with pip install snakeviz. RunSnakeRun (Python-2 only GTK GUI) and pyprof2calltree (uses QCacheGrind or KCacheGrind for display) are supported alternatives and can be similarly installed with pip.

Django

Elpy has basic Django support such as parsing either manage.py or django-admin.py (If it does not find manage.py it falls back to django-admin.py) for command completion assistance. Can also start runserver automatically and you can give an ip address and port.

C-c C-x c (elpy-django-command)

Choose what command you’d like to run via django-admin.py or manage.py. Please note that for compatibility reasons, the shell command will use python (instead of ipython by default).

C-c C-x r (elpy-django-runserver)

Start the development server command, runserver. Default arguments are 127.0.0.1 for ip address and 8000 for port. These can be changed via elpy-django-server-ipaddr and elpy-django-server-port.

Concepts

The RPC Process

Elpy works by starting a Python process in the background and communicating with it through a basic Remote Procedure Call (RPC) interface. Ideally, you should never see this process and not worry about it, but when things don’t work as expected, it’s good to know what’s going on in the background.

Every project and virtual env combination gets their own RPC process. You can see them both in the process list (M-x list-process) as well as in the buffer list (C-x C-b) as buffers named *elpy-rpc[...]*.

RPC processes are used to provide code completion, documentation and other features. To do so, they use python packages (jedi, yapf, rope, …) that are installed in a dedicated virtualenv (.emacs.d/elpy/rpc-venv by default). Those packages can be updated through the configuration panel (accessible with M-x elpy-config).

By default, Elpy will also find the library root of the current file and pass that to the RPC functions. The library root is the directory from which the current file can be imported.

There are a few options and commands related to the RPC process.

M-x elpy-rpc-restart

Close all running RPC processes. Elpy will re-start them on demand with current settings.

elpy-rpc-python-command (Customize Option)

The Python interpreter Elpy should use to run the RPC process. This defaults to "python", which should be correct for most cases.

Please do note that this is not an interactive interpreter, so do not set this to "ipython", "jupyter" or similar.

As the RPC should be independent of any virtual environment, Elpy will try to use the system interpreter if it exists. If you wish to use a specific python interpreter (from a virtual environment for example), set this to the full interpreter path.

elpy-rpc-virtualenv-path (Customize Option)

Path to the virtualenv used by the RPC.

Can be ‘default (create a dedicated virtualenv .emacs.d/elpy/rpc-venv), ‘system (use the system environment), ‘current (use the currently active environment), a virtualenv path or a function returning a virtualenv path.

If the default virtual environment does not exist, it will be created using elpy-rpc-python-command and populated with the needed packages from elpy-rpc–get-package-list.

elpy-rpc-large-buffer-size (Customize Option)

The size in character starting from which Elpy will transfer buffer contents via temporary files instead of via the normal RPC mechanism.

When Elpy communicates with the RPC process, it often needs to tell Python about the contents of the current buffer. As the RPC protocol encodes all data in JSON, this can be a bit slow for large buffers. To speed things up, Elpy can transfer file contents in temporary files, which is a lot faster for large files, but slightly slower for small ones.

elpy-rpc-pythonpath (Customize Option)

A directory to add to the PYTHONPATH for the RPC process. This should point to the directory where the elpy module is installed. Usually, there is no need to change this.

Backends

For introspection and analysis of Python sources, Elpy mainly relies on Jedi, a python package for static code analysis.

Due to the dynamic nature of python and its minimalist structure syntax, python code can be difficult to understand in certain situations. Jedi documentation provides some tips to make jedi job easier.

Virtual Envs

Elpy has full support for Python’s virtual envs. Every RPC process is associated with a specific virtual env and completions are done based on that environment.

Outside of RPC processes, though, it is not easy to have more than one virtual env active at the same time. Elpy allows you to set a single global virtual env and change it whenever you like, though.

M-x pyvenv-workon
M-x pyvenv-activate
M-x pyvenv-deactivate

These commands are the main interaction point with virtual envs, mirroring the normal activate and deactivate commands of virtual envs and the workon command of virtualenvwrapper.sh.

The pyvenv-workon command will allow auto-completion of existing virtual envs and also supports virtualenvwrapper’s setup hooks to set environment variables.

Elpy won’t pollute your Emacs command namespaces, but it might be an idea to create an alias for the workon command:

(defalias 'workon 'pyvenv-workon)

Modules

As the last concept, Elpy has a number of optional features you can enable or disable as per your preferences.

elpy-modules (Customize Option)

The list of modules to activate by default. See the section on Writing Modules for details on how to write your own modules.

Extending Elpy

Writing Modules

Modules are a way of easily extending Elpy with modular extensions. In essence, a module is a function which is called once to initialize itself globally, then once every time elpy-mode is enabled or disabled, and also once if elpy is disabled globally.

To achieve this, a module function receives one or more arguments, the first of which is the command specifier symbol, which can be one of the following:

global-init
Called once, when Elpy is enabled using elpy-enable.
global-stop
Called once, when Elpy is disabled using elpy-disable.
buffer-init
Called in a buffer when elpy-mode is enabled.
buffer-stop
Called in a buffer when elpy-mode is disabled.

To activate a module, the user has to add the function to elpy-modules.

Writing Test Runners

A test runner is a function that receives four arguments, described in the docstring of elpy-test-at-point. If only the first argument is given, the test runner should find tests under this directory and run them. If the others are given, the test runner should run the specified test only, or as few as it can.

Test runners should use an interactive spec of (interactive (elpy-test-at-point)) so they can be called directly by the user. For their main work, they can use the helper function elpy-test-run. See the elpy-test-discover-runner for an example.

To make it possible to set the test runner as a file-, directory- or project-local variable, the function symbol should get the elpy-test-runner property with a value of t.

Running Tests:

You can set up a working environment for Elpy using pip and cask. After installing Cask, create a new virtual environment and run the setup script in it:

virtualenv ~/.virtualenvs/elpy
source ~/.virtualenvs/elpy/bin/activate
./scripts/setup

You can now run ./scripts/test to run Elpy’s test suite.

If you cannot (or do not want to) use virtual environments on your system, you can set the environment variable ELPY_TEST_DONT_USE_VIRTUALENV to skip tests involving virtual environments: ELPY_TEST_DONT_USE_VIRTUALENV=t ./scripts/test.

Customization tips

Use flycheck instead of flymake

(when (load "flycheck" t t)
  (setq elpy-modules (delq 'elpy-module-flymake elpy-modules))
  (add-hook 'elpy-mode-hook 'flycheck-mode))

Enable emacs 26 flymake indicators in an otherwise light modeline

(setq elpy-remove-modeline-lighter t)

(advice-add 'elpy-modules-remove-modeline-lighter
            :around (lambda (fun &rest args)
                      (unless (eq (car args) 'flymake-mode)
                        (apply fun args))))

See also the associated issue

Auto-format code on save

Elpy allows you to auto-format your code with C-c C-r f (elpy-format-code). Many people find it useful to auto-format their code automatically on save. This can be achieved with the following snippet:

(add-hook 'elpy-mode-hook (lambda ()
                            (add-hook 'before-save-hook
                                      'elpy-format-code nil t)))

The formatting function (elpy-format-code here) can be replaced with your prefered one. For example to use black:

(add-hook 'elpy-mode-hook (lambda ()
                            (add-hook 'before-save-hook
                                      'elpy-black-fix-code nil t)))

Alternatives to elpy-goto-definition

Fallback to rgrep

You may sometimes find when you try to navigate to a function/class definition with elpy-goto-definition (M-.), that instead of jumping to the definition, you get the message “No definition found”. If you see this error often (because of the nature of the code you work on), you can use the following function instead of/in addition to elpy-goto-definition:

(defun elpy-goto-definition-or-rgrep ()
  "Go to the definition of the symbol at point, if found. Otherwise, run `elpy-rgrep-symbol'."
    (interactive)
    (ring-insert find-tag-marker-ring (point-marker))
    (condition-case nil (elpy-goto-definition)
        (error (elpy-rgrep-symbol
                   (concat "\\(def\\|class\\)\s" (thing-at-point 'symbol) "(")))))

This function will try to find the definition of the symbol at point using elpy-goto-definition, but will do elpy-rgrep-symbol instead, if the former function fails to return a result. You can bind this function to the key combination of your choice, or you can bind it to M-. to use it as a replacement for the the default goto-definition function:

(define-key elpy-mode-map (kbd "M-.") 'elpy-goto-definition-or-rgrep)

Jumping to assignment

As an alternative to elpy-goto-definition, Elpy also provides the function elpy-goto-assignment that jumps to the line where the symbol at point has been assigned. For functions and classes, it behaves roughly like elpy-goto-definition but has some advantages in certain situations (like if you want to jump to a decorated function). You can try this alternative with the following code:

(define-key map (kbd "M-.") 'elpy-goto-assignment
(define-key map (kbd "C-x 4 M-.") 'elpy-goto-assignment-other-window)

Enable full font locking of inputs in the python shell

(advice-add 'elpy-shell--insert-and-font-lock
            :around (lambda (f string face &optional no-font-lock)
                      (if (not (eq face 'comint-highlight-input))
                          (funcall f string face no-font-lock)
                        (funcall f string face t)
                        (python-shell-font-lock-post-command-hook))))

(advice-add 'comint-send-input
            :around (lambda (f &rest args)
                      (if (eq major-mode 'inferior-python-mode)
                          (cl-letf ((g (symbol-function 'add-text-properties))
                                    ((symbol-function 'add-text-properties)
                                     (lambda (start end properties &optional object)
                                       (unless (eq (nth 3 properties) 'comint-highlight-input)
                                         (funcall g start end properties object)))))
                            (apply f args))
                        (apply f args))))

See details in https://github.com/jorgenschaefer/elpy/issues/1428 and https://debbugs.gnu.org/cgi/bugreport.cgi?bug=32344.

FAQ

Q: Can I tell elpy not to load one of the minor modes it enables by default?

A: You can simply remove the appropriate symbol from elpy-modules.

Q: How do I make Elpy work well with org-mode?

A: You can try the following customization by @mankoff:

(setq python-shell-interpreter "ipython"
      python-shell-interpreter-args "--pylab=osx --pdb --nosep --classic"
      python-shell-prompt-regexp ">>> "
      python-shell-prompt-output-regexp ""
      python-shell-completion-setup-code "from IPython.core.completerlib import module_completion"
      python-shell-completion-module-string-code "';'.join(module_completion('''%s'''))\n"
      python-shell-completion-string-code "';'.join(get_ipython().Completer.all_completions('''%s'''))\n")

See #191 for a discussion and background.

Q: Can I get documentation as pop-ups next to the completion pop-up?

A: Not by default with company-mode, but there’s expez’ excellent company-quickhelp package you can install for this feature. It’s available from MELPA.

Q: How to solve company, yasnippet conflicts?

A: Add this snippet to your emacs configuration to avoid that.

(defun company-yasnippet-or-completion ()
  "Solve company yasnippet conflicts."
  (interactive)
  (let ((yas-fallback-behavior
         (apply 'company-complete-common nil)))
    (yas-expand)))

(add-hook 'company-mode-hook
          (lambda ()
            (substitute-key-definition
             'company-complete-common
             'company-yasnippet-or-completion
             company-active-map)))

Q: How do I install development version of elpy?

A: Remove existing elpy if you already installed it.

rm -rf ~/.emacs.d/elpa/elpy*

Install dependencies

(package-install 'company)
(package-install 'yasnippet)
(package-install 'pyvenv)
(package-install 'highlight-indentation)
(package-install 's)

Get latest code from github

mkdir ~/Projects
cd ~/Projects
git clone https://github.com/jorgenschaefer/elpy

Add the following to your ~/.emacs.d/init.el or ~/.emacs:

(add-to-list 'load-path "~/Projects/elpy")
(load "elpy" nil t)
(elpy-enable)

To update, run git pull --rebase, M-x load-library RET elpy RET and M-x elpy-rpc-reload.

Indices and tables