b7a7e707319689ee0cbc2de93b6896e5986112ee
7 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
|
afb3f83f30 |
feat(iace): cross-domain precision overhaul + component review + schema reconcile
Engine precision (stop foreign-machine patterns leaking into a project):
- Wire project.MachineType into the engine machine-type gate (empty input no
longer fires every machine class — press/cnc/excavator/crane/medical...).
- Capability-domain gating extended by 7 domains (outdoor, ventilation,
machining, bulk, palletizer, playground, fitness) so domain-specific hazards
only fire when the narrative names that domain; emitted via keyword_dictionary.
- Relevance backstop moved into iace (single gating contract, testable), and its
dominant false-anchor class removed (a long pattern word no longer matches a
short common token; prepositions/leitung added to the generic stoplist).
- New guard tests: TestCrossDomainPrecision (full pipeline, 0 foreign per GT) and
TestPatternReachability now asserts 0 dead patterns. Both GTs keep coverage 1.0.
Reachability fix: the 51 dead patterns required electrical/pneumatic/hydraulic
tags nothing produced — renamed to the canonical electrical_energy/
pneumatic_pressure/hydraulic_pressure/hydraulic_part.
Component review (negation is best-effort + expert-correctable):
- Parser surfaces negated components (ComponentMatch.Negated) instead of dropping
them; negated contribute no tags/energy → no phantom hazards.
- presence_status (vorhanden|nicht_vorhanden|geloescht) + ce_marked on components;
only `vorhanden` feed matching. CE+safety-relevant flags the PL/SIL obligation.
- Force re-seed preserves the expert's component decisions instead of wiping them.
- Tag-based component→hazard assignment (was: all on the first component).
- Negation-aware narrative parsing ("keine Pneumatik" no longer extracts it).
Local-dev DB: ai-sdk sets search_path=compliance,core,public; reconcile migrations
152-156 bring the consolidated local iace tables to the current schema + add the
presence_status/ce_marked columns. Machine-type vocabulary endpoint for the form.
[migration-approved]
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
||
|
|
7b3a6f0dcd |
fix(iace): close domain-gate gaps — generic patterns with press/welding/glass text
CI / loc-budget (push) Successful in 15s
CI / go-lint (push) Has been skipped
CI / python-lint (push) Has been skipped
CI / detect-changes (push) Successful in 6s
CI / branch-name (push) Has been skipped
CI / guardrail-integrity (push) Has been skipped
CI / secret-scan (push) Has been skipped
CI / dep-audit (push) Has been skipped
CI / sbom-scan (push) Has been skipped
CI / build-sha-integrity (push) Failing after 4s
CI / validate-canonical-controls (push) Successful in 12s
CI / nodejs-build (push) Has been skipped
CI / nodejs-lint (push) Has been skipped
CI / test-go (push) Failing after 37s
CI / iace-gt-coverage (push) Successful in 23s
CI / test-python-backend (push) Has been skipped
CI / test-python-document-crawler (push) Has been skipped
CI / test-python-dsms-gateway (push) Has been skipped
Observed on a real Kistenhubgeraet (lift) project: generic mechanical patterns (e.g. HP1000 "Quetschen Arm zwischen Pressenteilen") carry NO machine type and only generic tags (crush_point, rotating_part), so they fired for a lift; the narrow domain-gate terms missed their press/welding/glass wording. Broadens domainGateTerms (pressenteil, pressraum, blechbearbeitung, punktschweiss, schweisselektrod, elektrodenspalt) and adds a dom_glass domain (glasschneid/glasbearbeitung/...) with its emit keywords. New test pins that the four observed leakers now require a dom_* tag. Ghost=0, Leakage=0, coverage held on both GTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> |
||
|
|
b1357915ae |
feat(iace): Capability-Domain-Gating — Ghost 120→0, Leakage 25→0, Coverage 100%
CI / detect-changes (push) Successful in 8s
CI / branch-name (push) Has been skipped
CI / guardrail-integrity (push) Has been skipped
CI / secret-scan (push) Has been skipped
CI / build-sha-integrity (push) Failing after 4s
CI / validate-canonical-controls (push) Successful in 10s
CI / loc-budget (push) Successful in 11s
CI / go-lint (push) Has been skipped
CI / python-lint (push) Has been skipped
CI / nodejs-lint (push) Has been skipped
CI / nodejs-build (push) Has been skipped
CI / test-go (push) Failing after 40s
CI / iace-gt-coverage (push) Successful in 24s
CI / test-python-backend (push) Has been skipped
CI / test-python-document-crawler (push) Has been skipped
CI / test-python-dsms-gateway (push) Has been skipped
CI / dep-audit (push) Has been skipped
CI / sbom-scan (push) Has been skipped
Generische Pattern-Engine-Optimierung: behebt zwei Seiten derselben Wurzel (inkonsistente Applicability-Deklaration ueber 1216 Patterns). - Ghost-Patterns (120, feuerten nie): 34 nicht-erzeugbare Required-Tags via domaenenspezifische Keywords emittierbar gemacht -> 0. - Cross-Domain-Leakage (25, feuerten ueberall): neuer text-getriebener Capability-Domain-Gate (pattern_domain_gates.go) — Pattern mit Fremdmaschine im Szenariotext bekommt dom_*-Tag als Required-Gate -> 0. - Resolver: Komponente->TypicalEnergySources-Expansion (strukturierte Projekte). - Benchmark: GT-Platzhalter-Filter; faithful Cross-GT-Narrative-Harness. - Harte Regression-Guards: Ghosts=0, Leakage=0, Coverage>=90% (beide GTs). - HP2000/HP2001 (Secondary-Harm-Demos) in AllowlistKnownGaps -> Suite gruen. Echte Pipeline beide GTs: Coverage 100%/100%, 0 Leaks, 0 Ghosts. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> |
||
|
|
f534b52817 |
feat(iace): pattern audit suite + library hygiene wave
Add cmd/iace-audit CLI with 5 deterministic methods that find engine gaps without ground truth: - A reachability: 1058 patterns vs achievable tag universe - B consistency: components vs their declared hazard categories - C vocabulary: limits-form tokens vs keyword dictionary - D echo: limits-form sentences vs generated hazards (jaccard) - E hierarchy: hazards vs ISO 12100 design/protection/info levels Library fixes triggered by A+B+C findings: - tag_resolver: synonym map for electrical/pneumatic/hydraulic aliases - component_library: crush_point + EN03 (gravitational) on C014/C128 (Hubwerk family) - fixes HP1014/1015/1017/1018 which were silently weakly_reachable. noise_source added on 7 components (C006/C011/ C017/C020/C031/C041/C096). electrical_part on 8 drive components (C031/C032/C033/C034/C035/C036/C037/C038/C077/C092). cyber tag on 10 sensors (C081-C090) + 3 IT components (C111/C112/C116) + KI module C119 (ai_model added). pneumatic_part+hydraulic_part on valves C091/C093, hydraulic_part+chemical_risk on pump C097, moving_part on motion controller C075 - keyword_dictionary: EN03 added to aufzug/lift/hubwerk/hubgeraet (was wrongly EN04-only). New keyword entries for hub-action verbs: absenken/senken/anheben/heben + hubhoehe/hubweg/hubgeschwindig Audit impact: - A: weakly_reachable 409 -> 358 (-51 patterns now fully reachable) - B: incomplete components 46 -> 30 (-16, -33%) - HP1018 (Person unter absenkendem Maschinenteil eingeklemmt): weakly_reachable -> reachable Why: methods A/B/C surfaced that the Kistenhubgeraet test project generated 0 crush-under-load hazards despite OSHA 1910.212(a)(3) + EN ISO 12100 6.3.5.5 explicitly requiring them. Three orthogonal bugs (missing crush_point tag, wrong energy source mapping, missing action verbs in dictionary) silently disabled the entire lift crush pattern family. |
||
|
|
0371eecc03 |
fix: Struktureller Fix — Maschinentyp-Filter fuer Keywords + Patterns
PROBLEM: Cobot-Projekt hatte 52 Pressen-Hazards weil Keywords wie "stempel" und "stoessel" ohne Maschinentyp-Kontext matchten. FIX an 3 Stellen: 1. KeywordEntry.MachineTypes — Pressen-Keywords nur fuer press/*_press 2. ParseNarrative(text, machineType) — Parser laedt Maschinentyp aus Projekt 3. HazardPattern.MachineTypes — Pressen-Patterns (HP045-HP058) nur fuer Pressen Verhindert zukuenftig falsche Zuordnungen bei neuen Kundenprojekten. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> |
||
|
|
8dd1581fae |
feat: IACE SIL/PL calculator + Cobot patterns + library extensions
SIL/PL Calculator: Deterministic S×E×P → PL (a-e) → SIL (1-3) mapping Cobot Patterns (HP059-HP065): Human-robot collision, afterrun, misprogramming Press Patterns split into separate file (500-line guardrail) 5 new components (C136-C140), 5 new tags, 18 keyword entries Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> |
||
|
|
d4b7943d54 |
feat: IACE deterministic narrative parser + library extensions
Library Extensions: - 15 new components (C121-C135): knee lever, hydraulic ram, lubrication system, extraction system, vibrating plate, die tooling, transfer system, hoist, chute, oil drip tray, pressure relief valve, die space, flywheel, bin changeover station, inspection scale - 8 new tags: person_under_load, two_hand_control_required, thermal_accumulation, mechanical_transmission, oil_mist_risk, rapid_energy_release, gravity_suspended_load, bypass_risk - 14 new patterns (HP045-HP058): ram drop, die space crushing, oil mist inhalation, hot workpiece burns, suspended load, transfer draw-in, ejection fall, accumulator pressure release, impact noise, flywheel residual energy, guard bypass, two-hand misoperation, oil leakage, ergonomic bin changeover Deterministic Parser (NO LLM): - keyword_dictionary.go: ~100 entries mapping DE/EN keywords to component IDs, energy source IDs, and tags - narrative_parser.go: ParseNarrative() extracts components, energy sources, lifecycle phases, roles, tech specs, and context tags from free-text machine descriptions via keyword matching + regex - Tech spec regex: extracts kN, V, °C, bar, kW, rpm values and derives energy sources + severity tags automatically - iace_handler_parser.go: POST /projects/:id/parse-narrative endpoint chains parser → pattern engine → hazard suggestions Test: Paste Kniehebelpresse description → should detect 10+ components, 15+ hazards, all deterministically without LLM. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> |