Locoti

Locoti is a fork of Locot. A simple utility for tracking lines of code over time in a Git repository. Outputs CSV for visualization and analysis.

Improvements over Locot

• Reads commit trees directly from the local Git object database instead of cloning and checking out every revision.
• Caches subtree stats by Git tree ID, so identical trees across history are counted once
• Caches blob-level parsing and ignore decisions to reduce repeated work during history scans
• Uses faster hash maps (rustc_hash::FxHashMap) in the hot path
• Improves ignore filtering by respecting .gitignore and .ignore rules
• Improves language detection for extensionless files via shebang detection

Performance

$ hyperfine 'locoti ./tokei' 'locot --path ./tokei'
Benchmark 1: locoti ./tokei
  Time (mean ± σ):     249.2 ms ±   8.6 ms
 
Benchmark 2: locot --path ./tokei
  Time (mean ± σ):      5.996 s ±  0.088 s
 
Summary
  locoti ./tokei ran 24.06 times faster than locot --path ./tokei

Install

From crates.io:

cargo install locoti

From this repository:

cargo install --path .

Usage

Analyze the current repository and save the results:

locoti > results.csv

Run interactively in a terminal to see a graph instead of raw CSV rows:

locoti

Analyze a specific repository path:

locoti /path/to/repo > results.csv

CLI options

Output format

The CSV contains:

• Commit: commit hash
• Time: commit timestamp (UTC, RFC 3339 format)
• Total: total lines counted at that commit, after language exclusions
• Language columns: one column per language with non-zero lines in at least one commit

When stdout is attached to a terminal, locoti renders a terminal graph of total LOC over commit history instead of printing CSV rows. The graph labels the x-axis endpoints with the first and last commit dates. Redirect stdout to a file to keep the CSV output unchanged.

Example output

License

GNU GPL v3