codevira 2.0.0rc1__tar.gz → 2.1.2__tar.gz

{codevira-2.0.0rc1 → codevira-2.1.2}/CHANGELOG.md

@@ -11,7 +11,378 @@ This project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.htm
 
 ## [Unreleased]
 
-### Planned for v1.9
+## [2.1.2] — 2026-05-19 — Trust recovery + QoL
+
+Trust-recovery release based on **four independent field-test reports** that
+converged on "trust" as the gap (not capability). Full plan:
+[docs/plans/v2.1.2.md](docs/plans/v2.1.2.md).
+
+### Added — Smart similarity threshold (Item 1)
+
+- `search_decisions` now applies a per-project, self-calibrating similarity
+  threshold before RRF fusion. Gibberish queries (`"zzzzzz xqzv9"`,
+  `"how to make a cake"`, `""`) return zero results with
+  `retrieval: "semantic-no-results-above-threshold"` instead of the
+  v2.1.1 regression where they surfaced the "least bad" matches.
+- New `codevira calibrate` CLI command for manual threshold re-fit.
+  Auto-recalibration runs in a daemon thread every ~10 decisions added.
+- Per-project `<data_dir>/calibration.json` (search threshold + hook
+  threshold + positive-sample count + ISO timestamp).
+- Cross-session hook injection (`CrossSessionConsistency`) applies the
+  stricter `hook` threshold (search − 0.10). Commit-message-shaped
+  prompts (`feat(api):` / `fix:` / etc.) skipped entirely.
+
+### Added — Honest cleanup (Item 3)
+
+- `codevira reset --vectors / --graph / --all` — destructive operations
+  split out of `codevira heal` (whose name implied fix-in-place).
+- Auto-export of decisions + outcomes + preferences + learned_rules to
+  `<data_dir>/exports/<ts>-pre-<target>.json` BEFORE any wipe of `graph/`.
+  Pass `--no-backup` to skip.
+- Typed confirmation: user must type `reset` / `graph` / `vectors` /
+  `all` (not just `y`) to proceed. `--yes` skips for scripts.
+- `codevira heal --vectors / --graph / --all` deprecation cycle:
+  forwards to `cmd_reset` with a one-time warning. Removal planned v2.2.
+- New `codevira export decisions [--format json|sql] [--out PATH]`
+  standalone backup command. Closes Report 1 §7 gap.
+- New `confirm_typed(...)` helper in `_prompts.py`.
+
+### Added — Proactive correctness
+
+- **Item 20**: `check_conflict(decision_text, file_path?)` MCP tool detects
+  duplicates and conflicts vs `do_not_revert=True` decisions. Uses Item 1's
+  calibrated threshold. `record_decision` runs it automatically pre-write
+  and surfaces `_conflict_warning` in the response (suppressible with
+  `force=True`).
+- **Item 26**: `supersede_decision(old_id, new_decision, reason)` retires
+  a prior decision with auditable history. Schema auto-migrates with
+  `is_superseded INTEGER + superseded_by INTEGER`. `list_decisions` filters
+  superseded rows by default; `include_superseded=True` opts back in.
+
+### Added — Enumeration + filtering
+
+- **Item 11**: `list_decisions(limit, since_date, file_pattern,
+  protected_only, session_id, tags, include_superseded, full)` MCP tool.
+  Closes Report 3 "remembers but can't list" gap.
+- **Item 27 (partial)**: `tags=[...]` on `record_decision`; `list_tags()`
+  MCP tool; tag filter on `list_decisions`. `decision_tags` table
+  auto-migrated.
+- **Item 25**: `since="YYYY-MM-DD"` (or ISO 8601) filter on
+  `search_decisions`, `get_history`, `get_session_context`. SQL-layer
+  for BM25, post-filter for semantic results.
+
+### Added — Batch APIs (Items 23 + 24)
+
+- `record_decisions([...])` and `write_session_logs([...])` cut
+  memory-dump sessions from ~26 separate round trips to 1. Returns
+  `{count, recorded:[ids], errors:[{idx, error}]}` with per-item
+  partial-failure surfacing.
+
+### Added — Trust + correctness fixes
+
+- **Item 2**: `get_node` / `get_impact` / `query_graph` / `update_node`
+  return `not_indexed: True` + null counts instead of misleading 0 for
+  un-indexed paths.
+- **Item 4**: New `PostEditGraphRefresh` policy refreshes graph nodes in
+  a daemon thread after Edit/Write/MultiEdit so subsequent
+  `get_node` / `get_impact` calls see fresh data.
+- **Item 9**: `global_db.get_rules()` strict-language match by default
+  (was `language = ? OR language IS NULL`). Prevents Go-project rules
+  with NULL language from leaking into Python projects. Pass
+  `strict_language=False` for legacy behavior.
+- **Item 17**: Rule extractor noise filter — stopword filter + minimum
+  content-density gate + substring suppression in `_find_common_phrases`.
+  Pre-code projects (0 indexed source files) skip
+  `_infer_decision_pattern_rules` entirely.
+- **Item 18**: `add_phase()` silently replaces the bootstrap
+  "Getting Started" placeholder when called with the SAME number (and
+  the placeholder is pristine — status=pending, no changesets, default
+  description).
+- **Item 19**: Regression test for `file_path` serialization round-trip
+  through `get_session_context`'s `recent_decisions`.
+- **Item 22**: `write_session_log` / `log_session` auto-suffix
+  `session_id` on content collision (was: silent `INSERT OR REPLACE`).
+  Same id + same summary remains idempotent; different summary returns
+  the new suffixed id with `collision_resolved: True`.
+- **Item 33**: Hook commit-message pre-filter suppresses injection on
+  prompts matching `^(feat|fix|chore|docs|refactor|test|style|perf|build|ci|revert)(\(.*\))?:`.
+
+### Added — Roadmap workflow
+
+- **Item 10**: `complete_phase(backfill=True, completed_at='YYYY-MM-DD')`
+  for retroactive phase completion (current / upcoming / synthetic
+  cases).
+- **Item 12**: `complete_phase(git_ref="...")` links a commit sha or PR
+  reference to the completion entry.
+- **Item 29**: `bulk_import_phases([...])` for adopting codevira on a
+  project that already shipped N phases in git. Idempotent.
+
+### Added — QoL
+
+- **Item 5**: `do_not_revert` int→bool coercion at SQLite read boundary
+  (`search_decisions`, missing-rows fetch path). API contract now
+  matches schema.
+- **Item 6**: Smart truncation in `top_signals.rules` (word-boundary +
+  path-aware, 160-char limit).
+- **Item 7**: `summary` derived from first 80 chars of decision text
+  instead of `"ad-hoc record_decision"` placeholder.
+- **Item 8**: `get_session_context` returns `confidence_note` instead of
+  `confidence=null` on fresh projects.
+- **Item 28**: `summary_only=True` mode on `search_decisions` returns
+  id + summary + score + do_not_revert only — ~70% smaller payload for
+  AI triage queries.
+- **Item 30**: `record_decision` input-coerced echo — when
+  `do_not_revert` is passed as a non-bool (int 1, string "true"),
+  response carries `_input_coerced_warning`.
+- **Item 31**: Bundled non-Python playbooks (TypeScript / Go / generic)
+  in `mcp_server/data/rules/coding-standards-<lang>.md`. Auto-selected
+  by detected project language. Closes Report 1 §3.5.
+- **Items 13 + 14**: `clean --orphans` catches bare global.db rows (no
+  data dir + path missing on disk). `clean --ghosts` catches truly-
+  empty data dirs (<10 KB, status='stale').
+
+### Added — Plan + governance (Item 16)
+
+- `docs/plans/v2.1.2.md` mirrors the master plan (33 items + 4 deferred
+  v2.2-class items). Establishes release-planning discipline: every
+  vX.Y.Z release with 3+ items gets its own `docs/plans/` doc.
+- `ROADMAP.md` v2.1.2 section.
+- `CONTRIBUTING.md` Release planning + Documentation discipline.
+
+### Fixed
+
+- **Item 21**: Multi-language `get_signature` / `get_code` confirmed
+  working in v2.1.1 (15+ languages via tree-sitter-language-pack). No
+  new code needed; doc fix only.
+- **Item 32**: All 42 pre-existing mypy errors cleared via real fixes
+  (type narrowing, missing imports, `Counter` / `dict[str, Any]`
+  annotations, AST isinstance gating) and targeted
+  `# type: ignore[code]` for invariant pre-existing patterns. mypy is
+  now a hard pre-commit gate.
+
+### Late additions (caught by post-tag smoke testing)
+
+These three patches landed AFTER the initial v2.1.2 release commit but
+BEFORE shipping the wheel. They're all part of the v2.1.2 line:
+
+- **bulk_import_phases placeholder fix**: importing `phase=1` on a
+  fresh project was silently SKIPPING phase 1 because the bootstrap
+  "Getting Started" placeholder occupies that number. Adopters
+  migrating multi-phase git history (Report 3 #5 — the exact use case
+  Item 29 exists for) would hit this. Fixed by applying Item 18's
+  placeholder-recognition logic to bulk_import too.
+- **calibrate doc range fix**: `codevira calibrate --help` said
+  "Clamped to [0.20, 0.55]" but actual code clamps to [0.35, 0.80]
+  (the empirically-tuned values from Item 1 after measuring
+  all-MiniLM-L6-v2's distance distribution on real query/decision
+  pairs). Doc string corrected.
+- **Issue #10 — Antigravity sandbox + torch dlopen**: graceful
+  degradation across 3 tiers. (1) Removed `prewarm_embedding_model()`
+  from MCP server startup — torch loads lazily on first
+  `search_codebase` / `search_decisions` call. MCP `initialize` and
+  `tools/list` complete instantly without touching torch. All
+  non-search tools work in Antigravity. (2) `_decisions_collection_or_none()`
+  traps `OSError` (macOS dlopen errors arrive as OSError, not
+  ImportError) and surfaces `_semantic_warning` in `search_decisions`
+  responses with a clear explanation + issue link.
+  (3) `docs/troubleshooting/antigravity.md` documents the root cause
+  and four user-side workarounds. Closes
+  [#10](https://github.com/sachinshelke/codevira/issues/10).
+
+### Tests
+
+- 2401/2401 unit tests pass + 4/4 e2e cross-tool universality.
+- Replaced: `test_log_session_replaces_on_duplicate` → idempotent +
+  auto-suffix variants (Item 22).
+- Renamed: `test_8_evaluation_under_5ms_p95` → `_50ms_p95` (semantic
+  gate is inherently slower than BM25-only).
+- Updated: all `test_default_heroes_*` / `test_*_default_policies_*`
+  acceptance tests to expect `post_edit_graph_refresh` in the default
+  set; `test_dispatch_complete_phase` (×2) + `test_dispatch_get_history`
+  for new kwarg defaults.
+- New positive tests: `test_get_node_not_indexed`,
+  `test_get_impact_not_indexed`,
+  `test_gibberish_query_returns_zero_above_threshold`,
+  `test_session_context_recent_decisions_preserve_file_path`,
+  `test_add_phase_replaces_pristine_placeholder`,
+  `test_log_session_idempotent_on_same_content`,
+  `test_log_session_auto_suffixes_on_different_content`,
+  `test_language_filter_strict_excludes_null_language`,
+  `test_language_filter_loose_includes_null_language`.
+
+### Deferred to v2.1.3 / v2.2
+
+- Full README rewrite with animated 60-second demo GIF (Item 15) — pair
+  with v2.2 launch / benchmark publishing once benchmark suite ships.
+- Bundled `coding-standards-<lang>.md` for Rust / Java / etc. — extend
+  the per-language playbook system as adopters request.
+
+## [2.1.1] — 2026-05-17 — Hybrid decision search
+
+### Added
+- **Hybrid search for `search_decisions`** (BM25 + ChromaDB semantic + RRF fusion).
+  Closes the UDAP-benchmark gap: queries like `"DDD architecture layer"` or
+  `"codevira backfill"` that returned 0 hits in v2.0/v2.1.0 now surface the
+  right decisions via semantic recall. Response includes a `retrieval` field
+  indicating which path contributed (hybrid / keyword / semantic).
+- **`codevira heal --decisions`** — non-destructive backfill embedding all
+  existing decisions into the semantic index. Run once after upgrading from
+  v2.0/v2.1.0 to pick up hybrid recall on pre-existing decision history.
+  Idempotent (upsert pattern).
+- **`mcp_server/tools/_decision_embeddings.py`** — new helper module:
+  embed_decision, semantic_search_decisions, rrf_merge, backfill_all_decisions.
+  P9 graceful degradation throughout — chromadb failures never block SQL
+  writes or BM25 reads.
+
+### Tests
+- `tests/test_decision_embeddings.py` — 15 new regression tests covering
+  RRF math, graceful-degradation paths, and the explicit benchmark queries.
+
+## [2.1.0] — 2026-05-17 — Reliability hardening + Pillar 3 discipline scaffold
+
+### Added — Pillar 3: AI development discipline scaffold (2026-05-16)
+
+The codevira repo now ships its own discipline scaffold — the same
+pattern that will be exposed as `codevira discipline install` in
+v2.2. This is the reference implementation, eaten as dog food.
+
+- **`.claude/skills/`** — 4 SKILL.md files: `development-discipline`,
+  `open-source-quality`, `release-readiness`, `epistemic-honesty`.
+- **`.claude/hooks/pre-release-block.sh`** — PreToolUse hard wall.
+  Refuses `twine upload`, `gh release ... --draft=false`,
+  `pipx publish`, etc. unless `.release-evidence/<version>.json`
+  shows G1–G5 pass. Bypass via `CODEVIRA_RELEASE_OVERRIDE=1`.
+- **`Makefile`** — 12 targets including `release-verify-version`,
+  `release-gauntlet`, `release-build`, `release-dry-run`,
+  `release-publish`, `release-smoke`, `release-full`.
+- **`.pre-commit-config.yaml`** — ruff lint+format, mypy, hygiene.
+- **`tests/e2e/test_first_contact.py`** + 4 fixtures (docs_only,
+  code_only_python, polyglot, monorepo). G2 gate.
+- **`.github/workflows/{ci,release-gate}.yml`** — CI hard wall.
+- **`codevira.discipline.yaml`** — central scaffold config.
+- **`scripts/check_real_ide_smoke.sh`** — G3 stub.
+- **`docs/release-process.md`** — step-by-step foolproof release.
+
+### Planned for v2.1
+
+See [ROADMAP.md](ROADMAP.md#-v21--new-user-first-contact--reliability-hardening).
+
+- **Reliability hardening (23 bugs A–O)** — silent-failure elimination
+  surfaced by the discipline-scaffold e2e fixtures.
+- **Hybrid search (BM25 + semantic + rerank)** — natural-language
+  decision search.
+- **Decision deduplication** (ADD/UPDATE/NOOP) + audit trail.
+- **Conditional hook injection** — kill the always-on token tax.
+- **Multi-language `get_signature` / `get_code`** — wire tree-sitter.
+- **`record_decisions_batch` API** — compress protocol overhead.
+- **CLI naming clarity** — `init` / `setup` / `register` / `configure`
+  canonical hierarchy.
+
+---
+
+## [2.0.0] — 2026-05-14 — First public 2.0 release
+
+The 2.0 release moves codevira from "memory layer for one developer in one IDE" to
+"active guardian for every AI coding tool you use, on every project, on your local
+machine." Five internal iterations (rc1..rc5 in dev tags) plus a same-day public
+release-candidate cycle (`2.0.0rc1`) of dogfood + audit + product-credibility
+work consolidate into 2.0.0. **Full changelog: [RELEASE_NOTES.md](RELEASE_NOTES.md).**
+
+### Added
+
+- **All 10 hero policies** — active guardian engine intercepts every AI tool
+  call (Edit, Write, prompt submit, session start) and routes through
+  registered policies (Decision Lock, Anti-Regression, Scope Contract,
+  Blast-Radius Veto, Cross-Session Consistency, Token Budget, Live Style
+  Enforcement, Decision Replay, Proactive Intent Inference, AI Promotion
+  Score).
+- **`codevira setup`** — one-prompt installer that detects every AI tool
+  on the machine (Claude Code, Cursor, Windsurf, Antigravity, OpenAI Codex,
+  GitHub Copilot, Continue.dev, Aider) and configures all of them at once.
+- **`codevira projects`** — canonical inventory with `tracked / ghost /
+  orphan / stale` classification (`--json` for scripting; `--ghosts-only`
+  pairs with `clean --ghosts`).
+- **`codevira hooks list / uninstall`** — admin commands for Claude Code
+  lifecycle hooks; surgical install + clean removal.
+- **`codevira clean --ghosts`** — surgical removal of incomplete project
+  data dirs without touching tracked projects.
+- **`codevira init --single-language`** — opt-out flag for the new
+  index-everything default.
+- **`codevira engine` subcommand** — internal hook dispatcher; surfaces in
+  `--help` so the lifecycle hooks can call it.
+- **4 new doctor checks** — `claude_mcp_visibility`, `codeindex_freshness`,
+  `semantic_search_health`, `ghost_projects` (total 14 per run).
+- **Per-project config opt-out for cross-session injection** —
+  `.codevira/config.yaml: project: { cross_session_mode: off }` disables
+  the per-prompt context block without touching env vars.
+
+### Changed
+
+- **`codevira init` default** — indexes every common source/config/docs
+  extension (~75 total: `.py`, `.ts`, `.go`, `.yaml`, `.toml`, `.md`,
+  `.html`, `.sql`, `.proto`, …) instead of narrowing to one language.
+- **`codevira agents` default** — renders nudge files for **detected** IDEs
+  only; `--ide=all` opt-in for the legacy "render for every supported IDE"
+  behavior.
+- **`codevira doctor`** — now genuinely read-only; snapshots the projects
+  dir at entry and removes any new dirs at exit.
+- **`search_codebase`** — graceful structural fallback (filename + symbol
+  substring) when the semantic index is unavailable, with the correct
+  `fix_command` instead of a misleading "reinstall codevira" hint.
+- **`get_node` / `get_impact` / `query_graph`** — three-case error
+  differentiation: "no graph DB" / "graph empty" / "file not in populated
+  graph", each with its own `fix_command`.
+- **`get_decision_confidence`** — exposes `decisions_in_db_total` and
+  `decisions_eligible_for_outcomes` plus a four-state interpretation so
+  users understand WHY their `total_decisions` may be zero. Outcome
+  tracker also classifies file-less decisions via mention-extraction.
+- **Playbooks** — project-scoped first
+  (`<data_dir>/playbooks/` or `<project>/.codevira/playbooks/`); bundled
+  Python defaults are skipped with a clear warning when project language
+  ≠ Python.
+- **`register_project`** uses `ON CONFLICT … COALESCE(excluded.git_remote,
+  projects.git_remote)` — subsequent registrations can't silently clear
+  the `git_remote` column.
+- **Auto-init self-heal** runs SYNCHRONOUSLY in the calling thread of every
+  CLI invocation — daemon thread death no longer leaves ghost data dirs.
+- **Default install** includes ChromaDB + sentence-transformers (no
+  `[search]` extra needed for semantic search).
+- **README "92% reduction" claim** qualified with honest scope, per-prompt
+  cost, and amortization curve.
+- **`register` deprecation** now names the removal version (v2.1).
+
+### Fixed
+
+- macOS Apple Silicon **fork-safety segfault on first `codevira index`**
+  (auto-applied at indexer import).
+- **Setup interactive prompt** silent-fail on unexpected input — replaced
+  with a shared `_prompts.confirm` helper that retries, flushes stdout,
+  and handles `KeyboardInterrupt` cleanly.
+- Three **`status --global` UI typos** that always rendered 0/0/0
+  regardless of actual `global.db` state.
+- Four **FK race conditions** in the watcher pipeline.
+- **Python `None` leaked into argparse choices** for `agents --ide` and
+  `budget` positional.
+- Several **silent argument clamps** in `replay --since`, `insights --since`,
+  and `insights --top` now print visible warnings.
+
+### Tests
+
+- 2395 / 2395 passing (deterministic).
+- ~1091 net new tests since v1.8.0 (mostly from the v2.0 hero policies +
+  audit-driven regression coverage).
+
+### Note on internal v1.8.1 + 2.0.0rc1
+
+A v1.8.1 production hotfix existed in dev tags but was never published to
+PyPI; its fixes are folded into 2.0.0. A `2.0.0rc1` was briefly published
+on PyPI (2026-05-14) as a same-day public release candidate; the code is
+identical to 2.0.0. Anyone who installed `codevira==2.0.0rc1` can
+`pipx install --upgrade codevira` to move to 2.0.0 final.
+
+---
+
+## [Original v1.9 plan, deferred]
 
 - **Interactive checkbox UI for `codevira configure`**. The current
   prompt asks users to type comma-separated indices ("1,3,5") into a