Skip to content

CHANGELOG.md

Latest commit

 

History

History
103 lines (93 loc) · 5.66 KB

CHANGELOG.md

File metadata and controls

103 lines (93 loc) · 5.66 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

Fixed

  • Experiment resume no longer skips materially different runs. run_id was derived only from model alias, regime, condition and seed, so changing the case count (or the resolved model, the distributions config, or provider parameters) produced the same identifier and skip_completed=True treated the new experiment as already completed.

Added

  • mech_gov.experiment.identity — deterministic, versioned run identity. A readable run_label is now separate from an identity_sha256 digest computed over a canonical, allowlisted payload (case count, regime, condition, seed, hashed model ID, semantic distributions fingerprint, provider parameters, package version). Only the full digest decides resume.
  • LLMInterface.resume_identity_parameters() — optional hook for providers to declare adapter/deployment identity plus generation-affecting parameters. The default returns None, so existing custom providers remain API-compatible and are treated as non-resumable. Implemented for the bundled mock, OpenAI-compatible and Bedrock providers. SageMaker remains non-resumable because endpoint name and region do not establish an account-scoped deployment identity.
  • Results now include run_label, identity_schema, identity_sha256 and identity alongside the existing fields.
  • Stable identity warning codes MGI001–MGI008 on the mech_gov.experiment.identity logger. Diagnostics name the fields that differ, use a pseudonymous reference for correlation, distinguish unreadable completion state, and never print raw labels, paths, or parameter values.
  • README section documenting resume semantics, non-resumable runs and legacy result files.

Changed

  • run_id keeps the previous readable format for backward compatibility. identity_sha256, not run_id, decides whether a run may be skipped.

  • Runs whose identity cannot be established (no stable model ID, undeclared provider/deployment parameters, unreadable completion state, missing package version) always execute rather than risking a wrong skip.

  • Legacy records without identity metadata are read, never rewritten, and never treated as equivalent to a versioned identity.

  • The distributions configuration is loaded before the completion check, since its content participates in run identity.

  • OpenAI-compatible endpoint targets and Bedrock regions are represented only by SHA-256 fingerprints. Raw endpoint URLs, credentials and private paths are excluded. These fingerprints provide stable pseudonymous identity, not confidentiality.

  • Open-source readiness scaffolding:

    • Apache 2.0 NOTICE, expanded CONTRIBUTING.md (CLA + issue/PR flow), CODE_OF_CONDUCT.md, SECURITY.md, CODEOWNERS
    • CITATION.cff and a README citation block
    • Issue templates (bug, feature) and PR template
    • pyproject.toml tooling config (ruff, black, mypy, pytest, coverage) and real project URLs
    • SPDX headers on Python sources, scripts, examples and tests
    • GitHub Actions workflows (third-party actions pinned to SHA digests):
      • ci.yml — ruff + black + mypy + pytest matrix (3.10/3.11/3.12) with Codecov
      • codeql.yml — CodeQL SAST (push, PR, weekly cron)
      • dep-scan.yml — pip-audit (push, PR, daily cron)
      • license-check.yml — SPDX header verification + dependency-license allowlist (pip-licenses)
      • pattern-check.yml — internal-pattern scan with allowlist
      • scorecard.yml — OpenSSF Scorecard supply-chain analysis
      • cla.yml — CLA Assistant Lite
      • stale.yml — stale issues/PRs automation
      • release.yml — versioned source archive attached to GitHub Releases
    • .github/dependabot.yml — monthly Python and GitHub Actions updates
    • README badges, attribution and Citation sections

  • Privacy gate (R2): a pre-LLM governance primitive (mech_gov.governance.primitives.privacy_gate) that reversibly tokenizes direct identifiers (EMAIL, PHONE, SSN, PAN, IBAN, IP) before the model is consulted, and mechanically DEFERs a case when residual identifiers exceed a configurable budget or detection fails (fail-closed). Stdlib-only, vendor- neutral, configurable via PrivacyConfig; records privacy_entities_found and privacy_residual_pii counts in DecisionResult.metadata (the reversible token map is never persisted). Supports a pluggable PiiRecognizer. Ships an offline examples/privacy_demo.py.

0.1.0 - 2026-06-12

Added

  • mech_gov framework: model-agnostic governance for LLM decisions in high-stakes settings
  • Governance regimes: R1 (text-only), R2 (mechanical enforcement — hard gates, candidate freezing, argument-quality / I6Q checks, ambiguity gate, commit–reveal entropy step) and R3 (adaptive)
  • Vendor-neutral LLM interface (mech_gov.llm.base.LLMInterface) with a registry and mock, callable, openai_compatible providers plus optional bedrock/sagemaker backends behind an extra
  • Governance metrics (CDL, DIU, FVS, ESD, FSR, IPI) and task metrics (accuracy, macro-F1, MCC, deferral rate)
  • Synthetic banking decision dataset generator and an experiment runner with ablation, framing/FVS and seed tests
  • CLI scripts generate_dataset.py and run_governance.py, plus offline examples and a mock-backed regression test suite