PragyaLint 🔍

A static dead-code analyzer for Python. PragyaLint builds a module graph from your entry points, traces what's actually reachable, and reports unused files, exports, functions, imports, and circular imports — so you can keep your codebase lean.

Inspired by OptiPrune (a dead-code analyzer for TypeScript/JavaScript) — PragyaLint brings the same concept to Python.

Features

Installation

pip install pragyalint
# or
pipx install pragyalint
# or
pip install --user pragyalint

Requires Python 3.11+.

Quick start

Run an analysis from the project root:

pragyalint

Specify entry points and output formats:

pragyalint --entry src/main.py
pragyalint --json
pragyalint --sarif > pragyalint.sarif

Commands

Analyze flags

Rules / finding types

Confidence

Every finding carries a confidence level:

• high — structurally certain (module unreachable, import unused)
• medium — strongly implied (export/definition not referenced)
• low — heuristic (constant assignments, cycles)

Use --fail-on to make your CI fail on a given threshold, e.g.:

pragyalint --fail-on high

This exits with a non-zero code when any high-confidence finding exists — perfect for blocking PRs that introduce dead code.

Fixes

Like OptiPrune, PragyaLint's fixer is explicit, not implicit. Analysis only reports; use --fix to actually remove dead code. Always start with a dry run, inspect the output, then drop --dry-run.

# Preview what would change (does NOT modify files)
pragyalint --fix --dry-run

# Actually remove dead code
pragyalint --fix

# Remove dead code and functions with lower-risk edits too
pragyalint --fix exports --confidence medium+

Fixes are confidence-gated:

• --confidence high (default) — only high-confidence removals.
• --confidence medium+ — also remove unused exports/definitions.
• --force — allow an edit otherwise considered unsafe (e.g. removing a name listed in __all__).

Always commit before fixing. Removing code can have ripple effects. Run --dry-run first and re-run after fixing.

Configuration

PragyaLint loads configuration from the first file it finds (starting at the root directory):

1. pragyalint.toml
2. pragyalint.json
3. [tool.pragyalint] section of pyproject.toml

# pragyalint.toml
entry = ["src"]
ignore = ["migrations/", "generated/"]
extensions = [".py"]
detect_cycles = true
fail_on = "medium"

Or inside your pyproject.toml:

[tool.pragyalint]
ignore = ["tests/"]
report_unused_exports = false

Why Python for analyzing Python?

PragyaLint is written in pure Python and relies on the standard-library ast module. That gives you:

• A complete, battle-tested Python parser with zero external dependencies
• Correct handling of every modern Python syntax construct
• Simple distribution via PyPI / pipx
• No parser to build or maintain

Development

git clone https://github.com/example/pragyalint
cd pragyalint
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest

License

GPL-3.0-or-later