Public CI only runs synthetic fixtures, but every regression we have actually
shipped showed up on real projects, which are confidential and cannot reach a
GitHub runner. corpus_check.py runs a local corpus through the full pipeline
(parse, checks, HA YAML, entity suggestions) and diffs the counts against a
recorded baseline.
Only numbers are committed: counts, severity and code histograms, and a short
digest per source file. The project map with real paths lives in
tools/corpus_map.json, which is gitignored; corpus_map.example.json shows the
shape. Verified both directions: a deliberate change to the diagnostics filter
made the guard exit 1 and name the two metrics that moved, a clean tree exits 0,
and a machine without a map exits 2 without doing anything.
Wired into CONTRIBUTING and into the release checklist before the build step.
A synthetic 239-GA / 47-Function demo .knxproj (ETS6 schema, parseable by
xknxproject) with 5 deliberate mistakes, the tool's generated artifacts
(project report, Home Assistant entities, ETS XML/CSV), and a ha-brain/ smart
layer: circadian lighting, an 8-factor climate setpoint, presence/season/time
state machine and statistics. Whitelist the demo .knxproj past the *.knxproj
gitignore (synthetic, no real topology/secrets).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Design-time MCP server that reads .knxproj (read-only), validates naming/DPT/status,
and generates Home Assistant KNX YAML + ETS-importable group addresses (XML/CSV).
No live bus access — confined-workspace writes only.
Includes: 12 MCP tools, end-to-end smoke test, MIT license, English-first README
(+ Russian), CONTRIBUTING with a real-project test call, SECURITY policy, CHANGELOG,
GitHub Actions CI (Python 3.10–3.12), and issue/PR templates.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Nikolay Miroshnichenko