Read Latest Documentation - Browse GitHub Code Repository
isort your imports, so you don't have to.
isort is a Python utility / library to sort imports alphabetically, and automatically separated into sections and by type. It provides a command line utility, Python library and plugins for various editors to quickly sort all your imports. It requires Python 3.6+ to run but supports formatting Python 2 code too.
Try isort now from your browser!
Before isort:
from my_lib import Object
import os
from my_lib import Object3
from my_lib import Object2
import sys
from third_party import lib15, lib1, lib2, lib3, lib4, lib5, lib6, lib7, lib8, lib9, lib10, lib11, lib12, lib13, lib14
import sys
from __future__ import absolute_import
from third_party import lib3
print("Hey")
print("yo")
After isort:
from __future__ import absolute_import
import os
import sys
from third_party import (lib1, lib2, lib3, lib4, lib5, lib6, lib7, lib8,
lib9, lib10, lib11, lib12, lib13, lib14, lib15)
from my_lib import Object, Object2, Object3
print("Hey")
print("yo")
Installing isort is as simple as:
pip install isort
Install isort with requirements.txt support:
pip install isort[requirements_deprecated_finder]
Install isort with Pipfile support:
pip install isort[pipfile_deprecated_finder]
Install isort with both formats support:
pip install isort[requirements_deprecated_finder,pipfile_deprecated_finder]
From the command line:
isort mypythonfile.py mypythonfile2.py
or recursively:
isort .
which is equivalent to:
isort **/*.py
or to see the proposed changes without applying them:
isort mypythonfile.py --diff
Finally, to atomically run isort against a project, only applying changes if they don't introduce syntax errors do:
isort --atomic .
(Note: this is disabled by default as it keeps isort from being able to run against code written using a different version of Python)
From within Python:
import isort
isort.file("pythonfile.py")
or:
import isort
sorted_code = isort.code("import b\nimport a\n")
Several plugins have been written that enable to use isort from within a variety of text-editors. You can find a full list of them on the isort wiki. Additionally, I will enthusiastically accept pull requests that include plugins for other text editors and add documentation for them as I am notified.
You will notice above the "multi_line_output" setting. This setting defines how from imports wrap when they extend past the line_length limit and has 12 possible settings:
0 - Grid
from third_party import (lib1, lib2, lib3,
lib4, lib5, ...)
1 - Vertical
from third_party import (lib1,
lib2,
lib3
lib4,
lib5,
...)
2 - Hanging Indent
from third_party import \
lib1, lib2, lib3, \
lib4, lib5, lib6
3 - Vertical Hanging Indent
from third_party import (
lib1,
lib2,
lib3,
lib4,
)
4 - Hanging Grid
from third_party import (
lib1, lib2, lib3, lib4,
lib5, ...)
5 - Hanging Grid Grouped
from third_party import (
lib1, lib2, lib3, lib4,
lib5, ...
)
6 - Hanging Grid Grouped, No Trailing Comma
In Mode 5 isort leaves a single extra space to maintain consistency of
output when a comma is added at the end. Mode 6 is the same - except
that no extra space is maintained leading to the possibility of lines
one character longer. You can enforce a trailing comma by using this in
conjunction with -tc
or include_trailing_comma: True
.
from third_party import (
lib1, lib2, lib3, lib4,
lib5
)
7 - NOQA
from third_party import lib1, lib2, lib3, ... # NOQA
Alternatively, you can set force_single_line
to True
(-sl
on the
command line) and every import will appear on its own line:
from third_party import lib1
from third_party import lib2
from third_party import lib3
...
8 - Vertical Hanging Indent Bracket
Same as Mode 3 - Vertical Hanging Indent but the closing parentheses on the last line is indented.
from third_party import (
lib1,
lib2,
lib3,
lib4,
)
9 - Vertical Prefix From Module Import
Starts a new line with the same from MODULE import
prefix when lines are longer than the line length limit.
from third_party import lib1, lib2, lib3
from third_party import lib4, lib5, lib6
10 - Hanging Indent With Parentheses
Same as Mode 2 - Hanging Indent but uses parentheses instead of backslash for wrapping long lines.
from third_party import (
lib1, lib2, lib3,
lib4, lib5, lib6)
11 - Backslash Grid
Same as Mode 0 - Grid but uses backslashes instead of parentheses to group imports.
from third_party import lib1, lib2, lib3, \
lib4, lib5
To change the how constant indents appear - simply change the indent property with the following accepted formats:
- Number of spaces you would like. For example: 4 would cause standard 4 space indentation.
- Tab
- A verbatim string with quotes around it.
For example:
" "
is equivalent to 4.
For the import styles that use parentheses, you can control whether or
not to include a trailing comma after the last import with the
include_trailing_comma
option (defaults to False
).
As of isort 3.1.0 support for balanced multi-line imports has been added. With this enabled isort will dynamically change the import length to the one that produces the most balanced grid, while staying below the maximum import length defined.
Example:
from __future__ import (absolute_import, division,
print_function, unicode_literals)
Will be produced instead of:
from __future__ import (absolute_import, division, print_function,
unicode_literals)
To enable this set balanced_wrapping
to True
in your config or pass
the -e
option into the command line utility.
You can change the section order with sections
option from the default
of:
FUTURE,STDLIB,THIRDPARTY,FIRSTPARTY,LOCALFOLDER
to your preference (if defined, omitting a default section may cause errors):
sections=FUTURE,STDLIB,FIRSTPARTY,THIRDPARTY,LOCALFOLDER
You also can define your own sections and their order.
Example:
known_django=django
known_pandas=pandas,numpy
sections=FUTURE,STDLIB,DJANGO,THIRDPARTY,PANDAS,FIRSTPARTY,LOCALFOLDER
would create two new sections with the specified known modules.
The no_lines_before
option will prevent the listed sections from being
split from the previous section by an empty line.
Example:
sections=FUTURE,STDLIB,THIRDPARTY,FIRSTPARTY,LOCALFOLDER
no_lines_before=LOCALFOLDER
would produce a section with both FIRSTPARTY and LOCALFOLDER modules combined.
IMPORTANT NOTE: It is very important to know when setting known
sections that the naming
does not directly map for historical reasons. For custom settings, the only difference is
capitalization (known_custom=custom
VS sections=CUSTOM,...
) for all others reference the
following mapping:
known_standard_library
:STANDARD_LIBRARY
extra_standard_library
:STANDARD_LIBRARY
# Like known standard library but appends instead of replacingknown_future_library
:FUTURE
known_first_party
:FIRSTPARTY
known_third_party
:THIRDPARTY
known_local_folder
:LOCALFOLDER
This will likely be changed in isort 6.0.0+ in a backwards compatible way.
Some projects prefer to have import sections uniquely titled to aid in
identifying the sections quickly when visually scanning. isort can
automate this as well. To do this simply set the
import_heading_{section_name}
setting for each section you wish to
have auto commented - to the desired comment.
For Example:
import_heading_stdlib=Standard Library
import_heading_firstparty=My Stuff
Would lead to output looking like the following:
# Standard Library
import os
import sys
import django.settings
# My Stuff
import myproject.test
isort also makes it easy to sort your imports by length, simply by
setting the length_sort
option to True
. This will result in the
following output style:
from evn.util import (
Pool,
Dict,
Options,
Constant,
DecayDict,
UnexpectedCodePath,
)
It is also possible to opt-in to sorting imports by length for only
specific sections by using length_sort_
followed by the section name
as a configuration item, e.g.:
length_sort_stdlib=1
By default isort places straight (import y
) imports above from imports (from x import y
):
import b
from a import a # This will always appear below because it is a from import.
However, if you prefer to keep strict alphabetical sorting you can set force sort within sections to true. Resulting in:
from a import a # This will now appear at top because a appears in the alphabet before b
import b
You can even tell isort to always place from imports on top, instead of the default of placing them on bottom, using from first.
from b import b # If from first is set to True, all from imports will be placed before non-from imports.
import a
To make isort ignore a single import simply add a comment at the end of
the import line containing the text isort:skip
:
import module # isort:skip
or:
from xyz import (abc, # isort:skip
yo,
hey)
To make isort skip an entire file simply add isort:skip_file
to the
module's doc string:
""" my_module.py
Best module ever
isort:skip_file
"""
import b
import a
isort makes it easy to add an import statement across multiple files, while being assured it's correctly placed.
To add an import to all files:
isort -a "from __future__ import print_function" *.py
To add an import only to files that already have imports:
isort -a "from __future__ import print_function" --append-only *.py
isort also makes it easy to remove an import from multiple files, without having to be concerned with how it was originally formatted.
From the command line:
isort --rm "os.system" *.py
isort can also be used to verify that code is correctly formatted
by running it with -c
. Any files that contain incorrectly sorted
and/or formatted imports will be outputted to stderr
.
isort **/*.py -c -v
SUCCESS: /home/timothy/Projects/Open_Source/isort/isort_kate_plugin.py Everything Looks Good!
ERROR: /home/timothy/Projects/Open_Source/isort/isort/isort.py Imports are incorrectly sorted.
One great place this can be used is with a pre-commit git hook, such as this one by @acdha:
https://gist.github.com/acdha/8717683
This can help to ensure a certain level of code quality throughout a project.
isort provides a hook function that can be integrated into your Git pre-commit script to check Python code before committing.
To cause the commit to fail if there are isort errors (strict mode),
include the following in .git/hooks/pre-commit
:
#!/usr/bin/env python
import sys
from isort.hooks import git_hook
sys.exit(git_hook(strict=True, modify=True, lazy=True, settings_file=""))
If you just want to display warnings, but allow the commit to happen
anyway, call git_hook
without the strict parameter. If you want to
display warnings, but not also fix the code, call git_hook
without the
modify parameter.
The lazy
argument is to support users who are "lazy" to add files
individually to the index and tend to use git commit -a
instead.
Set it to True
to ensure all tracked files are properly isorted,
leave it out or set it to False
to check only files added to your
index.
If you want to use a specific configuration file for the hook, you can pass its
path to settings_file. If no path is specifically requested, git_hook
will
search for the configuration file starting at the directory containing the first
staged file, as per git diff-index
ordering, and going upward in the directory
structure until a valid configuration file is found or
MAX_CONFIG_SEARCH_DEPTH
directories are checked.
The settings_file parameter is used to support users who keep their configuration
file in a directory that might not be a parent of all the other files.
Upon installation, isort enables a setuptools
command that checks
Python files declared by your project.
Running python setup.py isort
on the command line will check the files
listed in your py_modules
and packages
. If any warning is found, the
command will exit with an error code:
$ python setup.py isort
Also, to allow users to be able to use the command without having to
install isort themselves, add isort to the setup_requires of your
setup()
like so:
setup(
name="project",
packages=["project"],
setup_requires=[
"isort"
]
)
Place this badge at the top of your repository to let others know your project uses isort.
For README.md:
[![Imports: isort](https://img.shields.io/badge/%20imports-isort-%231674b1?style=flat&labelColor=ef8336)](https://pycqa.github.io/isort/)
Or README.rst:
.. image:: https://img.shields.io/badge/%20imports-isort-%231674b1?style=flat&labelColor=ef8336
:target: https://pycqa.github.io/isort/
To report a security vulnerability, please use the Tidelift security contact. Tidelift will coordinate the fix and disclosure.
isort simply stands for import sort. It was originally called "sortImports" however I got tired of typing the extra characters and came to the realization camelCase is not pythonic.
I wrote isort because in an organization I used to work in the manager came in one day and decided all code must have alphabetically sorted imports. The code base was huge - and he meant for us to do it by hand. However, being a programmer - I'm too lazy to spend 8 hours mindlessly performing a function, but not too lazy to spend 16 hours automating it. I was given permission to open source sortImports and here we are :)
Get professionally supported isort with the Tidelift Subscription
Professional support for isort is available as part of the Tidelift Subscription. Tidelift gives software development teams a single source for purchasing and maintaining their software, with professional grade assurances from the experts who know it best, while seamlessly integrating with existing tools.
Thanks and I hope you find isort useful!
~Timothy Crosley