How can I check for unused import in many Python files?

Question

I remember when I was developing in C++ or Java, the compiler usually complains for unused methods, functions or imports. In my Django project, I have a bunch of Python files which have gone through a number of iterations. Some of those files have a few lines of import statement at the top of the page and some of those imports are not used anymore. Is there a way to locate those unused imports besides eyeballing each one of them in each file?

All my imports are explicit, I don't usually write from blah import *

Answer 1

PyFlakes (similar to Lint) will give you this information.

pyflakes python_archive.py

Example output:
python_archive.py:1: 'python_archive2.SomeClass' imported but unused

Answer 2

Use a tool like pylint which will signal these code defects (among a lot of others).

Doing these kinds of 'pre-runtime' checks is hard in a language with dynamic typing, but pylint does a terrific job at catching these typos / leftovers from refactoring etc ...

Answer 3

I use flake8 to check the style, and then isort+autoflake to auto remove the unused imports.

Check: See more at flake8 vs pyflake

pip install flake8 --user
flake8 .

Reformat: see more at why isort+autoflake

pip install isort autoflake --user
isort -sl .
autoflake --remove-all-unused-imports -i -r .
isort -m 3 .

Recently(since 2023), I use ruff to replace autoflake/flake8

pip install --upgrade --user ruff
ruff --fix /path/to/file_or_folder

Answer 4

You can also consider vulture as one of several options.

Installation

pip install vulture  # from PyPI

Usage

vulture myscript.py

For all python files under your project.

find . -name "*.py" | xargs vulture | grep "unused import"

Example

Applies to the code below.

import numpy as np
import pandas as pd

df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})

The results are as follows.

➜ vulture myscript.py
myscript.py:1: unused import 'np' (90% confidence)
myscript.py:4: unused variable 'df' (60% confidence)

Answer 5

Have a look at PyChecker. It is a debugging tool and able to find unused variables and modules.

Answer 6

I have been using pyflakes successfully and wished to auto-remove the unused imports.

I recently found autoflake:

• Uses pyflakes for checking.
• Defaults to removing unused standard library imports and redundant pass statements.
• Has options for removing other unused imports and unused variables.

Answer 7

If you use the eclipse IDE with pydev and mylyn, it provides automatic checking and highlighting for unused imports, among other things. It integrates with pylint as well.

Answer 8

autoflake is an improved version of pyflakes with additional options. It uses pyflakes under the hood, I would recommend to use it instead of using pyflakes directly.

Answer 9

I agree with using PyFlakes. It's like linting, but it excludes styling errors.

UPDATE

How to run: pyflakes <your python file> or pyflakes <your folder containing python files>

BE CAREFUL!!!

If you run it just with command pyflakes, it takes a really long time like it is never-ending. My hypothesis is it is trying to check every python file in your machine/folder when you call it that way.

Answer 10

Importchecker is a commandline utility to find unused imports in Python modules.

Answer 11

You can use the following user setting:

"python.linting.pylintEnabled": true,
"python.linting.pylintArgs": [
    "--enable=W0614"
]

But I think, you'll need to go through all files yourself and hit "save".

Answer 12

You can easily use pycln to do that, just do:

pip3 install pycln
pycln path_of_your_file.py -a

And then all the unused imports are going to be removed!