crucible / engine /policy.py

Commit History

  1. ML models: walk-forward-judged panel rank, plain-language ML disclosure in report and progress

    Verified signature
    da2ea0f
  2. Legacy cleanup: state mode is the only debate path; text gate, claim sweep, report exporter, metric categories and 29 unused modules removed; LLM stage override replaced by a deterministic GAAP-distortion test; unused dependencies dropped

    Verified signature
    33b81e9

  1. AnalysisState pipeline: reconciliation, staged validation, frozen state, claim-token prose, audited renderer; public/QA split

    Verified signature
    c9fa753
  2. P0 hardening: one percentage type, exhaustive rendered-claim audit, engine-owned terminology, reliability as evidence permissions, typed metric categories, FPI leadership, 10b5-1 semantics, deterministic ranges

    Verified signature
    d5edb72
  3. Semantic types: normalized metrics (currency/scale/ADS basis) with gated derivations; engine-owned labels and methodology; one ScenarioSet and immutable object ids; reliability controls weight, language and persona visibility; structured thesis triggers with causal polarity; separate intrinsic/pricing/momentum signals; claim-registry render audit; 8 publication gates; 95 tests

    Verified signature
    742f5b8
  4. P1: primary-source segment facts (annual-report XBRL instance), SEC bank capital/credit facts, sourced ADS ratio and 20-F officers; mixed-issuer equity engine + real sum-of-parts; engines own the two-stage arithmetic; no risk-free/ROE/growth/sector-P/E defaults; fallback register

    Verified signature
    cdc46b3
  5. Architecture: typed concepts + immutable AnalysisContext with rejected-field quarantine and lineage; applicability policy; valuation engines with label invariants; quant reliability states; semantic comparisons; engine-side synthesis/scenario EV; five-state checks; six publication gates incl. final render audit; property/golden/corruption tests

    Verified signature
    43c8298