loc - Advanced Lines of Code Counter

A fast, feature-rich LOC tool written in Rust.
Author: kelexine

Features

• Fast project scanning: Counts text files across a target directory with optional Rayon-powered parallel processing.
• Code/comment/blank split: Classifies source lines using language-aware single-line and block-comment rules.
• Tree view: Renders a recursive project tree when --tree is enabled, with optional binary-file display.
• Function extraction: Uses Tree-sitter-backed extractors for Rust, Python, JavaScript/TypeScript, Go, C/C++, Java/Kotlin/C#/Scala, PHP, Swift, and Ruby.
• Complexity analysis: Reports function length and a branch-count cyclomatic complexity estimate.
• Git-aware discovery: Uses git ls-files in repositories and can attach last-modified dates from git log.
• Agent mode auto-detection: Switches to token-efficient TSV output when run inside AI coding agents (Claude Code, Gemini CLI, etc.).
• Lockfile awareness: Automatically detects dependency lockfiles (Cargo.lock, package-lock.json, etc.), excluding them from line metrics to prevent skewed stats.
• Machine-readable outputs: Supports direct-to-stdout JSON (--json), TSV (--format agent), and pipe-friendly raw path lists (-q).
• Multi-format export: Writes JSON, JSONL, CSV, TSV, and HTML reports.
• Global configuration: Reads defaults from ~/.config/loc-rs/config.toml through the platform config directory.
• Size warnings: Flags files above a configured line threshold.
• BOM-aware binary detection: Treats UTF-16/UTF-32 files as text instead of false-positive binary files.

Installation

From crates.io

cargo install loc-rs

From source

Requires Rust 1.92.0 or newer.

git clone https://github.com/kelexine/loc-rs
cd loc-rs
cargo build --release
cargo install --path .

The release binary is built at ./target/release/loc.

Quick Start

loc
loc src/
loc -d
loc --tree
loc -f --func-analysis
loc -t rust python typescript
loc -e report.html -f
loc --json
loc -q | head -20
loc --format agent -e out.tsv

Documentation

• CONTRIBUTING: Local setup, test commands, and release workflow.
• Architecture Guide: Module layout and runtime flow.
• Release Guide: Step-by-step cargo release sequence.
• Security Notes: Threat model and CI safety guidance.
• FAQ: Quick answers for common usage questions.
• Troubleshooting Guide: Common failure modes and fixes.

Generate Rust API docs locally:

cargo doc --no-deps

Open docs in your browser:

cargo doc --no-deps --open

Usage

loc [OPTIONS] [DIRECTORY]

DIRECTORY defaults to the current directory.

Common workflows

Export examples

loc -e results.json
loc -e results.jsonl
loc -e stats.csv -f
loc -e report.html -f

All Flags

Configuration

You can persist defaults in ~/.config/loc-rs/config.toml on Linux, or the equivalent platform config directory returned by the OS.

warn_size = 500
default_types = ["rust", "python"]
always_extract_functions = true

.locignore (project root) is also supported for custom file and directory ignores:

# Ignore generated snapshots
snapshots
*.min.js

GitHub Action Integration

The repository includes a comprehensive composite GitHub Action for CI line-count and complexity checks. The action features sub-second installs via pre-built binaries, GitHub Step Summary integration, and automated artifact uploads.

steps:
  - uses: actions/checkout@v4
  - uses: kelexine/loc-rs/.github/actions/loc-rs@main
    with:
      target_dir: .
      warn_size: 500
      functions: true
      fail_on_warn: true     # Fails the build if any files exceed warn_size
      export_html: true      # Automatically uploads HTML report to artifacts
      export_json: true      # Automatically uploads JSON report to artifacts
      version: latest        # Uses pre-built binaries for instant execution

Use static workflow values for target_dir, warn_size, and args; do not pass untrusted pull request, issue, or comment text into shell-backed action inputs.

Supported Languages

Language aliases are supported for common names such as py, js, ts, tsx, rs, rb, sh, bash, zsh, md, yml, kt, hs, c++, cxx, cc, and cs.

Function Extraction Support

Function extraction is available when -f or --func-analysis is enabled. The extractors return function names, line ranges, parameters where supported, async markers where supported, class/struct markers, docstrings where supported, decorators where supported, and a complexity estimate.

Export Formats

Troubleshooting

License

MIT © kelexine