Rust69.4%
Python21.5%
JavaScript7%
CSS1.2%
HTML0.4%
Shell0.2%
Other0.4%

Commit activity

Last 12 weeks · 59 commits

Full graph →

Community health

3 of 6 standards met

Community profile →
71
✓README✓License✓Contributing○Code of Conduct○Issue Template○PR Template

Recent PRs & issues

Active · 1 in progress · Last activity 1d ago
See all on GitHub →
clintecker

Infer accessory sizing for jewelry and watches categories and are now recognized category roots: in maps them to the existing , the same way and already work. Purchases in these categories no longer need a manually declared in , and no longer fail with a confusing item-level when the author forgets to declare one. Behavior change worth noting now fires for / categories that declare a non-accessory (for example on a ring). Before this change the check was skipped for these roots (inference returned nothing to compare against), so such catalogs compiled. The one-line fix is . Catalogs with no declared are unaffected — inference covers them now. What's in the box Inference arms (): / roots and dot-prefixed children infer accessory sizing. No new // variants; ring sizes (numeric with halves, e.g. ) and watch inventory () fit the existing accessory forms and bounds. Compiler tests (): one test proves both roots infer accessory sizing with no declared (ring renders , watch renders ); one proves a wrongly declared kind now fails (asserted as the single recorded error). These are also the first tests anywhere covering accessory-kind numeric sizing. Fictional examples (): a "Band ring" (, metal composition, size on a bounded accessory profile) and a "Field watch" (, ), following the existing tee entry's shape, with URLs. The example categories omit , so the existing example-catalog CI tests exercise the inference end to end. Docs (): one sentence documenting that recognized roots infer their sizing kind while unrecognized roots declare explicitly, plus a stale example-description fix. One over-broad assertion in the numeric-pair roundtrip test was scoped to the tee purchases it actually tests (the new heterogeneous example purchases made the blanket assertion untestable). Out of scope, deliberately Today outfit generation () is untouched — jewelry and watches stay out of outfit suggestions, consistent with the existing belts/hats exclusion and the project's "Today is an optional experiment" stance. No web UI, HA integration, OpenAPI, or migration changes were needed: category and size travel through those layers as opaque strings.

clintecker · 1d ago
jessfraz

Use Kit v0.1.8 for quiet hover feedback that respects touch, focus, selected states, and disabled controls. Checked the app views and dialogs in Playwright on phone, tablet, and desktop in both orientations, light/dark themes, Chromium, and WebKit.

jessfraz · 1d ago
jessfraz

Measure the purchase source and tracking status in one browser frame. Separate calls could straddle a layout update and compare positions from different scroll offsets; the ordering assertion stays the same.

jessfraz · 2d ago
jessfraz

Keep the changed-position test focused on invalidating and reselecting the reading position. The fixture shuts down the backend and simulated reader; the extra UI cleanup command was racing a reader revision.

jessfraz · 2d ago
Structured data for AI agents

Repository: jessfraz/ha-closet. Description: A JSON-backed wardrobe inventory for Home Assistant Stars: 2, Forks: 1. Primary language: Rust. Languages: Rust (69.4%), Python (21.5%), JavaScript (7%), CSS (1.2%), HTML (0.4%). License: MIT. Topics: cli, home-assistant, home-assistant-add-on, home-assistant-custom-component, inventory, rust, wardrobe. Latest release: v0.4.13 (1d ago). Open PRs: 1, open issues: 0. Last activity: 1d ago. Community health: 71%. Top contributors: jessfraz, dependabot[bot].