Stats

g stats renders a usage dashboard right in your terminal — a contribution heatmap, the commands you run most, your conventional-commit mix, busiest hours, top authors, and more. Everything is computed locally and stored in ~/.config/g/g.db (SQLite). Nothing is uploaded; nothing phones home.

g stats

First run feels a little empty? Some sections (commit search, duplicates, message length) read from an imported history. Run g stats --import once to backfill, then re-run g stats.

Quick start

g stats                       # full report, last 365 days
g stats --days 90             # narrow the time-based sections to 90 days
g stats --no-git              # skip sections that shell out to git
g stats --import              # backfill commit history into the local db
g stats --search "fix login"  # fuzzy-search commit messages
g stats --duplicates          # find commit messages you've reused
g stats --message-stats       # subject-length trends over time

What’s in the report

A full g stats run prints these sections, each as its own panel:

Section	What it shows	Source
Usage Overview	Totals: commands run, commits, active days	local db
Commit Heatmap — Last 52 Weeks	GitHub-style contribution grid	git
Lines Changed — Last 60 Commits	Additions vs deletions per commit	git
Top Commands	The g/git subcommands you run most	local db
Commit Types	Your feat / fix / docs … mix from g commit	local db
Repository Activity	Commits, churn, and recent momentum	git
Activity by Hour (UTC)	When in the day you commit	git
Top Authors	Who’s been committing in this repo	git

Pass --no-git to print only the panels that come from the local database — handy in a directory that isn’t a git repository, or for a fast summary.

Importing history

The command-tracking panels grow naturally as you use g. The commit-message features (search, duplicates, message length) need a one-time import of your existing history:

g stats --import              # import everything
g stats --import-limit 2000   # cap the import to the most recent N commits

Re-running --import is safe: existing commits are skipped, only new ones are added.

Search your commit messages

Fuzzy-search across everything you’ve imported:

g stats --search "auth"
g stats --search "revert migration"

Each hit shows the short hash, author, date, and subject. If you get no results, import first:

g stats --import && g stats --search "auth"

Find duplicate & repetitive messages

Spot copy-pasted or low-signal commit subjects:

g stats --duplicates          # messages used more than once

Pair it with message statistics to see whether your subjects are trending too long (the conventional-commit sweet spot is ≤ 72 characters):

g stats --message-stats

This prints overall length statistics plus a 12-month trend of average subject length, and a Top Authors / Top Duplicate Messages breakdown.

Where the data lives

Path	Contents
~/.config/g/g.db	SQLite database: command usage + imported commits
~/.config/g/config.toml	Your configuration (see Configuration)

Because it’s a plain SQLite file, it’s easy to inspect or wipe:

g developer db                # open the database in sqlite3
rm ~/.config/g/g.db           # start fresh (re-created on next run)

Tips

• Run g stats --days 30 on a Monday for a quick “what did I ship last month”.
• g stats --no-git is instant — great for a shell prompt widget or alias.
• Combine with Theming: a heavy-border theme makes the heatmap and bars pop on a projector.

Related docs

• Configuration — where the database and config live
• CLI Reference — every flag for g stats
• Log & diff — the everyday history commands