Fixit: Advanced Python Linting Framework with Autofixes

2.2.1 · active · verified Thu Apr 16

Fixit is an advanced Python linting framework developed by Meta, providing robust auto-fixing capabilities and hierarchical configuration. Built on LibCST, it allows for precise code modifications and makes it straightforward to write custom, in-repo lint rules. Fixit includes a wide array of built-in lint rules, optimizing for efficiency and ease of customization. The current version is 2.2.1, and it undergoes active development with a consistent release cadence.

Common errors

```
KeyError: 'rules' or similar when running fixit with custom rules after upgrade.
```
cause The configuration format for Fixit 2.x changed from YAML to TOML, and the way rules are enabled/disabled also changed. Older configuration files are no longer valid.

fix Migrate your `.fixit.config.yaml` to `pyproject.toml` or `fixit.toml`. Rules are now enabled using `enable = ["mypackage.rules:CustomLintRule", "another_package.rules"]` in the `[tool.fixit]` section.
```
AttributeError: 'module' object has no attribute 'LegacyRule' or 'BaseLintRule'
```
cause The base class for custom lint rules and other core API types were renamed or refactored in Fixit 2.x for conciseness.

fix Update your custom lint rule definitions. The primary base class for new rules is now `fixit.LintRule`. Review the 'Upgrading' guide and API reference for specific class name changes.
```
SyntaxError: invalid syntax (on newer Python features) or similar parsing errors.
```
cause Fixit relies on LibCST for parsing. If your Python environment's LibCST version is outdated, it might not support the latest Python syntax features.

fix Ensure both `fixit` and `libcst` are updated to their latest versions compatible with your Python interpreter (`pip install --upgrade fixit libcst`). Fixit requires Python >=3.9.

Warnings

breaking Fixit 2.0 (and newer) introduced a foundational rewrite from Fixit 1.x. Configuration files changed from YAML to TOML (e.g., `pyproject.toml` or `fixit.toml`), rule referencing syntax was updated, and the internal API for custom rules changed significantly.
Fix: Refer to the 'Upgrading' guide in the official documentation. Migrate `.fixit.config.yaml` to `pyproject.toml` or `fixit.toml` using the new TOML format. Update custom lint rules to align with the Fixit 2.x API.
breaking Flake8-style suppression comments (`# noqa`) are no longer supported in Fixit 2.x. They have been replaced by `# lint-ignore` or `# lint-fixme` directives.
Fix: Replace all instances of `# noqa` with `# lint-ignore` or `# lint-fixme` in your codebase. You can optionally specify rule names after the directive for granular suppression.
gotcha Fixit's autofix functionality might not work as expected when integrated with external formatters like 'Black' if Black makes no changes to the file. This can lead to a state where `fixit fix` reports no changes, even if lint issues are present that Fixit *could* fix.
Fix: This is a known issue (GitHub issue #427). A workaround might involve running `fixit fix` separately from formatter runs, or inspecting `fixit`'s output carefully when an external formatter is also in use.
gotcha When using `fixit_file()` or similar functions with `parallel=True` (for multiprocessing), Fixit currently does not support applying individual fixes interactively. Automated autofix (`autofix=True`) will still apply changes, but interactive selection is disabled.
Fix: If interactive fixes are desired, set `parallel=False` when calling relevant API functions. For fully automated fixes, set `autofix=True`.
gotcha While Fixit offers per-file multiprocessing for performance, it may still be slow on very large codebases without caching or further optimizations.
Fix: Monitor Fixit's performance on your codebase. Consider implementing caching mechanisms if not already integrated. The Fixit team has plans for future performance improvements, including caching.

Install

pip install fixit Install stable version

Imports

LintRule
```
from fixit import LintRule
```
InvalidTestCase
```
from fixit import InvalidTestCase
```
Used for defining invalid code examples in custom lint rule tests. Also 'Invalid' is an alias.
ValidTestCase
```
from fixit import ValidTestCase
```
Used for defining valid code examples in custom lint rule tests. Also 'Valid' is an alias.
Configuration
wrong:
```
from fixit.configuration import Configuration
```
correct:
```
from fixit.ftypes import Configuration
```
Configuration types and utilities have moved to `fixit.ftypes`.

Quickstart

This quickstart demonstrates how to use Fixit from the command line to lint a Python file and then apply available autofixes. It creates a temporary Python file with common linting issues, runs `fixit lint` to identify them, and then `fixit fix` to resolve them automatically.

import os
import subprocess

# Create a dummy Python file
with open('my_module.py', 'w') as f:
    f.write('x =  1\n')
    f.write('def foo():\n')
    f.write('    return 1 +  2 # Bad spacing\n')
    f.write('class MyClass(object): # Inheriting from object is redundant in Python 3\n')
    f.write('    pass\n')

# Run fixit lint to see issues
print('--- Running fixit lint ---')
subprocess.run(['fixit', 'lint', 'my_module.py'])

# Run fixit fix to apply autofixes
print('\n--- Running fixit fix ---')
subprocess.run(['fixit', 'fix', 'my_module.py'])

# Print the fixed file content
print('\n--- Fixed file content ---')
with open('my_module.py', 'r') as f:
    print(f.read())

# Clean up
os.remove('my_module.py')

view raw JSON →