Ruff

CodeRabbit runs Ruff to find issues in Python code and Jupyter Notebooks.

Ruff is a linter for Python and Jupyter Notebooks.

CodeRabbit runs Ruff v0.16.9.

Files

Ruff will run on files with the following extensions:

• .py
• .ipynb (using nbQA)

Configuration

enabled · boolean · default: true

Enable Ruff

config_file · string

Optional path to a Ruff configuration file relative to the repository. When set, this configuration is used for every reviewed file instead of Ruff’s closest-config discovery.

.coderabbit.yaml

.coderabbit.yaml

reviews:
  tools:
    ruff:
      enabled: true
      config_file: "ruff.toml"

Web UI

In your organization or repository settings, go to Reviews → Tools and turn the tool on or off with its toggle. To change its other settings, click the settings icon next to the toggle.

Ruff supports the following config files:

• pyproject.toml
• ruff.toml
• .ruff.toml

config_file must be a repository-relative path. If it is a URL, an absolute path, contains .., or does not exist, CodeRabbit skips Ruff for the review.

When config_file is not set, Ruff finds the closest applicable configuration for each reviewed file by searching its directory hierarchy. If no repository configuration applies, CodeRabbit uses an isolated fallback configuration based on the selected review profile.

Ruff runs inside the CodeRabbit sandbox. Repository configuration is parsed there by Ruff; CodeRabbit does not load user-provided plugins or Python init hooks.

When we skip Ruff

CodeRabbit will skip running Ruff when:

• Ruff is already running in GitHub Actions, GitLab CI, CircleCI, or Azure Pipelines.

Ignored codes

The following Ruff codes are automatically ignored:

• I001 - unsorted imports
• F401 - unused imports
• F405 - undefined name from star import
• W291 - trailing whitespace
• W293 - blank line contains whitespace
• Q000 - bad quotes
• ANN001 - missing type annotation
• ANN201 - missing return type annotation
• UP006 - deprecated type alias
• UP045 - deprecated import
• E501 - line too long
• S101 - use of assert detected
• EXE001 - Shebang is present but file is not executable
• RUF100 - unused noqa directive
• PT009 - pytest unittest assertion
• LOG015 - root logger call
• PLR2004 - magic-value-comparison
• PLR0913 - too-many-arguments
• PLR0917 - too-many-positional-arguments
• F841 - unused-variable
• F811 - redefinition-of-unused-name
• PIE794 - duplicate class field definition

When CodeRabbit generates a fallback configuration, it also ignores E402 in __init__.py and conftest.py files.

Profile behavior

These rule selections apply to CodeRabbit’s generated fallback configuration. A repository configuration or an explicit config_file supplies its own rule selections; CodeRabbit still filters the ignored codes from review comments.

In Chill mode, Ruff uses a focused set of rule codes including Pyflakes (F), flake8-bugbear (B), flake8-bandit (S), flake8-blind-except (BLE), flake8-debugger (T10), flake8-executable (EXE), Ruff-specific bug rules (RUF), Pylint error checks (PLE), flake8-builtins (A), and selected pycodestyle checks.

In Assertive mode, Ruff additionally enables:

• ANN - flake8-annotations (type annotations)
• ASYNC - flake8-async (async/await issues)
• FBT - flake8-boolean-trap
• C4 - flake8-comprehensions
• DTZ - flake8-datetimez (timezone issues)
• DJ - flake8-django
• ISC - flake8-implicit-str-concat
• LOG - flake8-logging
• G - flake8-logging-format
• PIE - flake8-pie
• PT - flake8-pytest-style
• FLY - flynt (f-string opportunities)
• UP - pyupgrade (upgrade opportunities)
• PLR - Pylint refactor checks
• SIM - flake8-simplify
• PERF - perflint
• RET - flake8-return
• ARG - flake8-unused-arguments
• TRY - tryceratops
• PLW - Pylint warning checks

Links

• Ruff Configuration
• nbQA Documentation