weight-progress-dashboard

Run the generate.py script via the talent-run command, which reads morning weights from a SQLite database and renders a matplotlib PNG with trend lines, goal bands, and stat cards. The script prints the output image path on success.

Compute a linear regression over the last 7 days of weights, then extend the slope forward until it intersects the goal weight to get an ETA date. The script draws this as a dashed projection line alongside the raw time series.

The script requires at least two morning weight rows in the database and exits with code 2 otherwise. It filters on the period column so evening entries are excluded, which can leave too few qualifying rows early on.

No, matplotlib is installed only in the Talent's locked uv environment, not system python. Invoke it through talent-run or uv run with the project path; if the environment is missing, the script degrades gracefully and exits with code 2.

When the 7-day slope turns positive, the ETA cards display a dash or recovery week message instead of a projected date. The chart still renders the raw data and full-period statistics.