health-sleep-analysis

Export sleep and blood oxygen samples using the apple-healthkit CLI into JSON files, then run the provided Python scripts to aggregate nightly sleep stages and render visual reports. The scripts handle night attribution, stage mapping, and chart generation automatically.

Fetch sleep and SpO2 samples with apple-healthkit into /tmp files, run sleep_report_data.py to build a summary JSON, then render it with sleep_report_librsvg.py using rsvg-convert. Pass --lang zh, en, or ja to localize all chart text.

SpO2 at or above 95% is normal, 90-94% is low and worth monitoring, and below 90% is dangerous. Frequent drops below 95% or any reading under 90% may indicate obstructive sleep apnea and warrants medical evaluation.

The workspace directory has an iSH to iOS sync delay that causes stale reads. Always write raw HealthKit JSON and script outputs to /tmp first, then copy the final PNG files to the workspace only for display.

The rendering environment is missing CJK fonts. Install them with apk add font-noto-cjk before running the rsvg-convert or matplotlib rendering scripts so localized chart text displays correctly.

Sleep stage and SpO2 data require an Apple Watch worn during sleep. Nights with less than one hour of recorded data are automatically filtered out, so missing nights usually mean the watch was not worn.