Weksa/tools/show-working
Repository files (latest commit first)
Filename Latest commit message Latest commit date
Weksa d4c72c54aa show-working guard: reject Validation/Test items by split before writing
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-11-10 18:04:40 +00:00
..
build.py show-working guard: reject Validation/Test items by split before writing 2026-11-10 18:04:40 +00:00
README.md Show her working: static page of the lowering chain from the real lowerer and scorer 2026-11-08 00:50:39 +00:00

Show her working

A static page that shows the lowering chain for committed fixture packets: typed slots, inventory entries chosen, lowered English, scorer closeness to the corpus target, with hedge, certainty and restraint annotated. Refusals (admission rules, LowerError, addressee consent, negation awaiting polarity) are results, not omissions.

  • crates/weksa/examples/show_working.rs: runs the real admission door, lower and pronounce over proof_packet() and its variants and prints JSON. Each feature is removed in turn and the packet re-run, so "the lowerer read this" is measured.
  • build.py: runs that, scores each lowered text with the weksa score CLI (needs Ollama with qwen3-embedding:0.6b), writes out/data.json and one dependency-free out/index.html. Fails if anything cannot run; there is no fallback text.
  • scripts/publish-show-working.sh: build, then publish through gamecult-ops publish-mind-preview.sh.

Adding a fixture: add an entry in fixtures() in the example. A packet from another store needs a loader there; none exists yet.