busya/ss-tools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity
1,039 Commits 2 Branches 19 Tags
0c895cf416ca2c42a57eca0c90ec268920f7b308
Commit Graph

96 Commits

Author SHA1 Message Date
busya
0c895cf416 feat(logging): unify canonical task CoT events 2026-08-24 17:00:17 +03:00
busya
ffa4d6a85b feat(scenarios): implement execution engine contracts 2026-08-21 16:15:40 +03:00
busya
82a519a347 feat(scenarios): complete editor execution and analytics 2026-08-20 11:32:26 +03:00
busya
488a8f349b test(backend): raise coverage to 95%+ statements and branches (97.8%/95.0%) 
- ~60 new/extended test files across api, core, plugins, services, schemas:
  routes, superset clients, task_manager, lineage, git, translate,
  dashboard-testing, load-testing, migration, llm_analysis, scheduler, ssl
- .coveragerc: enable branch coverage; exclude src/__tests__ (test files)
  and src/scripts (CLI/ops tools) from the denominator
- bug fixes found while testing:
  * settings: PUT /settings/reports registered under duplicated prefix
  * schemas/lineage: FleetReportDTO missing run_status (route always 500)
  * dashboard_testing/baseline_inheritance: visual entry read wrong field
  * superset_client/_databases: logger extra name shadowed LogRecord attr
  * routes/datasets: _yaml_string_paths recursion without yield from
  * translate/sql_generator: restore explicit-type timestamp contract
  * baseline_catalog: remove unreachable dashboard_id fallback
- conftest fixes: pytest_plugins to rootdir conftest (pytest 9), test
  filename collision, TMPDIR-safe integration fixtures
 2026-08-19 17:14:32 +03:00
busya
81ba94684f feat: sed-мутации датасетов, правки миграции и фиксы 
- Dataset viewer: sed-замена (find→replace) по всем/выбранным/текущим датасетам
  с обязательным визуальным предпросмотром и целями (sql/metrics/yaml); сохранение
  и выбор именованных правил (sedRules store + SedRuleEditor).
- Migration: literal-замена в dataset YAML при переносе; рескан ID чартов/датасетов
  перед миграцией + метрика средней длительности синка (GET /migration/sync-stats);
  логическая группировка опций (маппинги БД, сервер изменений, rescan) под галочками.
- Fix: дашборды хаба скрывались дефолтным профиль-фильтром show_only_slug_dashboards=true
  (теперь false); мастер миграции не грузил дашборды предзаполненного источника;
  потеря терминального task_status из-за утечки pending-корутин в /ws/logs (панель
  результата не появлялась); горизонтальный скролл в MappingTable; чистая per-env
  ошибка в mapping coverage.
- Backend: record_sync_duration + alembic-миграция sync-duration; literal_replace модуль.
 2026-08-19 10:29:28 +03:00
busya
c32b7ef509 feat(translate): extend run metrics with observed flow stats; test hardening 
- backend: aggregate cache_hits/observed_runs/source_records_read/eligible/
  translated/same_language_skipped/insert_rows_* preserving NULL for
  historical runs; drop legacy translate plugin module
- frontend: history page metric cards + RunOutcomeCompact per-run summary,
  totals with observed-scope notice; tabular numbers
- tests: fix banner date-format expectations, api-key env-scope fixture,
  rate-limiter cache pollution pinning, chart/candidates guards
- specs: sync dashboard-testing openapi contract
 2026-08-13 08:23:43 +03:00
busya
512e9223e3 feat(maintenance): configurable date format for banner timestamps 2026-08-10 11:22:48 +03:00
busya
3421484347 chore: synchronize remaining workspace updates 2026-08-10 07:29:08 +03:00
busya
1c577e8561 docs(semantics): normalize code contracts 2026-08-10 07:28:05 +03:00
busya
c9664dfabc fix(040/038): address code-review criticals (C1-C4, H1, M1, M3) 
QA review of the 036-041 closure range returned FAIL with 3 criticals, all
confirmed. Fixes:

C1 - breaker dead: on_result=persist_batch is now wired into RunnerPool
  (breaker.record() fed per result); added test_breaker_abort_persists_partials
  proving CIRCUIT_BREAKER_ABORT reachability + partial persistence.
C2 - index-based result mapping corrupted data under concurrency: results now
  map by execution_id to their source item; test uses two distinct payloads
  and asserts chart->digest pairing (previously masked by identical fixtures).
C3 - double-acquire of the shared client semaphore (deadlock invariant):
  RunnerPool no longer manually acquires the client semaphore; capacity is
  enforced by worker count, the client bounds total concurrency.
C4 - duplicated ScenarioGraph.Vlm.Analyze region: outer region renamed
  ScenarioGraph.Vlm [TYPE Module].
H1 - _default_submit stub removed: analyze_screenshot requires submit=; no
  silent empty-findings fallback.
M1 - test_capture_dispatch.py region closed.
M3 - capture.py raw_sha256 bypass removed: digest always derived from real
  capture_bytes (no caller-supplied hash).

Verification: load_testing (77) + scenario (103) = 180 passed; ruff clean;
all region pairs balanced.
 2026-08-07 16:18:29 +07:00
busya
d1e15904e1 feat(038): wire real VLM submit + real capture bytes (T057-T059) 
Close the 038 MVP runtime gaps found in the audit: VLM analysis previously
returned empty findings with no provider call, and capture registered a
synthetic sha256 derived from run/step ids instead of real image bytes.

T057 - vlm.py: replace _default_submit stub with real submit_screenshot that
  resolves a multimodal provider via LLMProviderService (decrypted key,
  multimodal-required gate) and calls Plugin.Service.LLMClient.get_json_completion
  with the masked screenshot; analyze_screenshot is now async.
T058 - capture.py: dispatch_capture now REQUIRES real capture_bytes/masked_bytes
  and computes sha256 from the actual image bytes (synthetic hashes forbidden);
  scenario API accepts base64 capture/masked bytes.
T059 - test_scenario_vlm_e2e.py: capture -> VLM -> disposition end-to-end with
  real bytes (masked bytes reach the provider; digest matches sha256 of bytes).

Verification: tests/services/dashboard_testing/scenario/ = 103 passed, ruff clean
(existing B008 on pre-existing draft-pack route lines untouched).
 2026-08-07 14:08:04 +07:00
busya
a1b20bf2cf feat(040): wire RunnerPool into run_load_run — real load execution (T075-T079) 
Close the 040 MVP runtime gap: run_load_run previously only slept through
RAMP->STEADY->DRAIN->COMPLETED with zero Superset requests. Now it actually
executes load-test chart queries through the 037 client.

- executor.py: 037-native adapter (execute_superset_chart via
  execute_dashboard_query_envelope; build_execution_items with stable ids)
- persistence.py: write_load_executions batched LoadExecution persistence
- plugin load_testing.py: run_load_run -> _resolve_execution_scope ->
  _run_bounded_pool (RunnerPool + shared client_registry.get_semaphore +
  CircuitBreaker -> circuit_breaker_abort)
- test_executor_runtime.py: 5 tests proving executor delegation, item
  determinism, real-execution persistence (2 items -> 2 LoadExecution rows,
  run -> COMPLETED), and lifecycle-only degradation on missing env

Verification: tests/services/load_testing/ = 76 passed, ruff clean.
 2026-08-07 13:37:46 +07:00
busya
c1c35e5855 fix(scenario): route draft storage through StorageService with legacy fallback, persist run marker for post-restart recovery 2026-08-07 10:24:32 +07:00
busya
a5e69de5ea fix(scenario): complete save flow with auto-created repo, robust approval gate, idempotent auto-start 2026-08-07 01:56:36 +07:00
busya
9cb5717a78 fix(agent): auto-start scenario chat, robust HITL resume, strict service auth 
- frontend: fix auto-start on dashboards->/agent navigation (undefined params
  ReferenceError), route initial connect through ConnectionManager with auto-retry,
  reset runModel on objectId change and failed recovery
- agent: fix closure-over-loop-variable bug in _inject_env_id_into_tools (env now
  resolved from request-local ContextVar; idempotent wrapping), make
  execute_dashboard_result.result_key optional, resilient checkpoint resume with
  ToolMessage repair + direct-tool fallback, remove dead fast-path, consolidate
  tool_call parsing in _tool_resolver, context-safe ContextVar resets
- backend: llm-config gated by strict service-only auth (no user-JWT fallback),
  tighten idempotent run reuse (dashboard/env/intent match + 6h staleness),
  terminal event transitions run.status to COMPLETED/FAILED/CANCELLED,
  null-safe metric parsing in dashboard query model
- run.sh/docker-compose: require SERVICE_JWT (random per-run secret) instead of
  public default
 2026-08-06 18:26:50 +07:00
busya
6d19ecf81a fix: harden feature security and e2e integrations 2026-08-06 13:47:28 +07:00
busya
0d9f3e09f3 chore: close 041/040 tails — contract test, lifecycle test, fixtures, integration test 
- 041 T038: 040 blast-radius consumer contract test (R6 pinned snapshot shape).
- 040 T003: materialize load-testing fixtures into frontend __fixtures__.
- 040 T029: run lifecycle test (ramp/steady/terminal immutability) + T038 aggregate-progress.
- 040 T051/T059: LoadResults + LoadComparison L2 UX tests.
- 040 T064: LoadModels integration test (profile -> gate -> run -> reconnect -> results).
- Check off verified tasks; remaining 040 open: e2e (T065), integration (T066/T067), a11y (T068).
 2026-08-05 21:28:54 +07:00
busya
9de74baa08 feat: dashboard testing suite — scenario UI, load testing, dataset lineage 
- 039-dashboard-scenario-ui: agent workspace (WorkspaceModel, scenario
  views, parameters/baselines, artifact preview, HITL save/approval,
  evidence/VLM review, pipeline verification views) + contextVersion=2
  scenario intent + prototype.
- 040-dashboard-load-testing: capacity/matrix/profile, runner pool,
  timing, circuit breaker, PROD gate, cache/bounded/consistency,
  comparison/schedule, API + frontend + prototype.
- 041-dataset-lineage-blast-radius: usage index, severity/schema-diff,
  propagation, deprecation, fanout, API + frontend + prototype (R9
  labels/metrics/recreate).
- 036/037 amendments: dataset_updated trigger + fanout_plan_id, lineage
  deprecation gate, 040 read-model traceability.
- Includes pre-existing 042-rls-management-workspace and
  043-idm-account-integration work.
 2026-08-05 18:15:30 +07:00
busya
f9a15a0a7b feat(maintenance): opt-in auto-end at end_time via scheduler scan 
Add auto_end flag to maintenance_events: when set with an end_time, a
60s APScheduler scan dispatches the end task automatically. The scan
survives restarts and is de-duped by task_id; end_time alone stays
informational. Includes alembic migration, route/schema wiring, Svelte
checkbox with validation, examples, and backend + frontend tests.
 2026-08-04 16:38:20 +07:00
busya
d54f903660 feat(maintenance): settings (timezone, height), snapshot restore, UX polish 
Settings:
- display_timezone defaults to Europe/Moscow everywhere (model, service
  fallbacks, settings form); migration flips untouched 'UTC' default row.
- New banner_height setting (1-200 grid units, NULL = auto): threaded from
  settings through start/rebuild flows into MARKDOWN insert/content updates;
  settings panel gains Auto/Manual radio + number input.

Banner removal:
- update_dashboard_layout returns the pre-mutation position_json; stored on
  maintenance_dashboard_banners.original_position_json at creation.
- Removal restores the snapshot verbatim when the current layout matches the
  deterministic replay of the insert (normalize + y-shift + banner keys);
  diverged layouts (user edits during maintenance) and legacy banners fall
  back to surgical removal. Fixes layout drift from y-shift and unreverted
  ROOT->TABS -> ROOT->GRID normalization.

UX:
- StartMaintenanceForm: recent-tables chips from event history, Enter-to-
  submit, schema.table format hint, inline task progress panel (status +
  progress bar + task-log link), templates section removed, Button atoms.
- Maintenance page: environment selector + env context init.
- Events table: Active/Completed tabs with count badges.
- Backend: plugins/services report monotonic progress via
  context.logger.progress (per-dashboard) for start/end/end-all.

Tests: 203 backend (matcher, restore/fallback, height, progress) + 3645
frontend pass; migration E2E-verified (upgrade/downgrade).
 2026-08-04 11:08:11 +07:00
busya
80dad15458 fix(maintenance): render banner as native MARKDOWN element, not a chart 
Two defects fixed:
- ensure_banner_chart created an orphan 'Maintenance Banner' markdown chart
  (polluted Charts menu and dashboard exports). The banner is now a native
  MARKDOWN element in position_json; chart_id is a synthetic layout key.
- insert_banner_markdown_at_top blindly targeted GRID_ID; on ROOT->TABS
  dashboards (FI-0085) GRID_ID is orphaned and the banner never rendered.
  The layout is now normalized to ROOT->GRID_ID->[ROW-banner, ...] with
  recursive parents update, matching the proven-working manual example.

Review-driven hardening:
- liveness check verifies reachability from ROOT (children graph), so
  dashboards corrupted by the old bug self-heal on the next start.
- insert removes stale ROW-banner-*/MARKDOWN-banner-* keys (single banner).
- decision memory (@RATIONALE/@REJECTED) added to both modules.
- new ops script scripts/cleanup_maintenance_banner_charts.py (dry-run by
  default) deletes already-created bogus banner charts in prod.
 2026-08-04 08:46:22 +07:00
busya
02a97bfc9c fix(maintenance): survive sqlparse token cap on huge virtual dataset SQL 
Discovery of virtual datasets now works, but a runtime blocker remained: any
virtual dataset whose SQL exceeds sqlparse's MAX_GROUPING_TOKENS (10000 tokens)
raised SQLParseError 'Maximum number of tokens exceeded (10000)' from
extract_tables_from_sql_span, which is called unguarded in the scan loop — one
oversized virtual dataset aborted the whole maintenance preview/start.

- extract_tables_from_sql_span now wraps sqlparse.parse + token walk in
  try/except and falls back to regex-only extraction (keeping all schema.table
  matches) instead of raising, so huge SQL no longer fails the scan.
- Tier-1 virtual filter uses value "" (not None) so the sql is_not_null filter
  passes Superset's rison schema instead of always falling back to a full scan.

Tests: huge-SQL fallback (extractor) and huge-virtual-dataset scan resilience
(scanner). ADR-0020 updated with Decision 3.
 2026-08-03 23:48:01 +07:00
busya
52e909a2da fix(maintenance): discover virtual (SQL) datasets in dashboard scanner 
Virtual (SQL) datasets were never matched, so maintenance discovery returned
0 affected dashboards. Two defects fixed:

- find_affected_dashboards filtered by is_sqllab_view, which is NOT a
  filterable column in Superset's dataset list API (absent from search_columns),
  so the query was rejected. Now discover virtual datasets via the filterable
  sql column: primary server-side 'sql is_not_null' filter with a client-side
  non-empty-sql scan as fallback (best-effort vs pagination cap), dedupe by id.

- AsyncAPIClient.request never called raise_for_status(), so rejected filters
  (HTTP 400) were returned as bodies without a 'result' key and surfaced as
  'Found 0 datasets', dead-coding the filtered->full-scan fallback. request()
  now raises on non-2xx via the existing error mapper.

Tests cover both virtual-scan tiers, all fallback paths, the raise behavior,
and an end-to-end match with the real sql_table_extractor on production SQL.
Documented in ADR-0020.
 2026-08-03 22:21:00 +07:00
busya
912583acb7 fix: RBAC admin flag self-heal; await WS maintenance broadcast; scalable dataset discovery 
- RBAC: ensure_admin_role() guarantees the Admin role carries is_admin=True
  (startup self-heal + create_admin promotion + role-is_admin UI checkbox in
  admin/roles); update_role refuses to strip is_admin from the last admin role.
- WS: broadcast_maintenance_event is now awaited (3 sites) so maintenance
  events actually reach clients (was an un-awaited coroutine RuntimeWarning).
- Pagination: MAX_PAGINATION_PAGES cap + clear error in fetch_paginated_data
  to stop runaway loops on huge environments.
- Discovery: find_affected_dashboards and translate datasource picker filter
  datasets/dashboards server-side (table_name/id filters, opr operator per
  Superset OpenAPI) instead of full scans that hit the pagination token cap;
  fallback to full scan when filters are rejected; virtual-dataset dedupe.
 2026-08-02 23:09:08 +07:00
busya
5719029a71 fix(038): QA gate — uicontext None guard in agent handler, ruff compliance, belief-scope wiring 
- agent_handler: guard scenario_mode against None uicontext (regression in
  test_handler_missing_auth_continues_gracefully)
- tools_038.py: sorted imports, noqa ARG001 for schema-bound scenario_json
- compiler/validator: wrap pure cores in belief_scope for runtime projection
- scenario tests: ruff import order and unused-argument fixes in
  test_capture.py, test_capture_dispatch.py, test_vlm.py

Backend scenario 80 passed; dashboard-testing 378 passed;
agent 352 passed, 12 skipped; ruff clean for 038 scope.
 2026-07-31 14:22:43 +03:00
busya
cb95d79707 feat(038): Phase 8 — capture, VLM analysis, human disposition 
- T037-T047: CaptureProfile validated from capture-profile.schema.json (default
  profile + dispatch via 036 Evidence bridge: original + masked artifacts),
  VLM typed findings (parse/validate, stale-prompt guard, prompt template v1
  with versioned hash), human disposition (confirm requires comment,
  double-disposition 409, graph immutability)
- Belief runtime: REASON/REFLECT/EXPLORE on all C4/C5; audit 0 errors
- 85 backend tests pass; ruff clean; regions balanced
 2026-07-31 13:17:45 +03:00
busya
4e93a31407 feat(038): Phase 6 — safe draft pack compiler 
- T026-T030: ScenarioGraph.PackCompiler.Generate — registered versioned templates
  (scenario.yaml, runner.plan.json, report_template.md, evidence_manifest.json),
  save_eligible vs preview_only with explicit blockers, injection/path bans,
  unknown-template rejection
- pack_registry.py: register_pack_drafts through 036 AgentRuns.Artifacts.Register
  (save_eligible only, safe intended_paths)
- Belief runtime: REASON/REFLECT/EXPLORE (preview_only fallback); audit 0 errors
- 63 scenario tests pass; ruff clean
 2026-07-31 13:03:18 +03:00
busya
31e8524a32 feat(038): Phase 5 — US4 immutable resolver 
- T022-T025: ScenarioGraph.Resolver.Resolve — typed parameter/selector/manual/
  remove-step resolutions emit immutable revisions linked via parent_revision_hash;
  stale base revision rejected (409 semantics); unrelated step ids/order unchanged
- Selector hints recorded in step description for auditability
- 58 scenario tests pass; ruff clean; regions balanced
 2026-07-31 13:00:11 +03:00
busya
9224d1a9ca feat(038): Phase 4 — US3 canonical serializer + golden tests 
- T018-T021: ScenarioGraph.Serializer.Canonical — canonical JSON (sorted keys,
  stable separators) + canonical YAML; revision hash excludes volatile identity
  fields; JSON/YAML represent equal domain data
- Golden tests: repeated/shuffled serialization byte-identical, YAML round-trip
  equals JSON domain, revision hash derived from canonical bytes
- Belief runtime: REASON/REFLECT/scope in CanonicalYaml + ValidateCore +
  CompileImpl; audit_belief_runtime 0 errors
- 52 scenario tests pass; ruff clean
 2026-07-31 12:56:48 +03:00
busya
67e4b9fc42 feat(038): Phase 3 — US2 validator safety matrix 
- T013-T017b: ScenarioGraph.Validator.Validate with deterministic findings —
  duplicate steps, missing deps, cycles (with path), duplicate/missing refs,
  tool/action registry, SQL/code/path-traversal bans, raw baseline literals,
  unresolved params/selectors/baselines, coverage classification
- Decomposed to 8 helpers (C901 fixed: _validate_core 36→5 complexity)
- Property tests: chain DAGs of any length valid, self-dep cycle, dup refs
- Belief runtime: REASON/REFLECT in validate + validate_core; audit 0 errors
- 47 scenario tests pass; ruff clean
 2026-07-31 12:53:33 +03:00
busya
40bbdc97f6 feat(038): Phase 1-2 — catalog, models, capability mapper, deterministic compiler 
- T001-T005: catalog_v1.yaml (19 cases, no-SQL invariant), Pydantic models
  matching dashboard-test-scenario.schema.json, 6 canonical fixtures,
  materialization to backend/tests/fixtures/dashboard_scenarios/
- T006-T012: capability_mapper (all 19 cases classified, xlsx/technical/
  mutation-safety fallbacks), registered tool/action templates, deterministic
  compiler with stable ids/refs/coverage/fingerprints; repeated+shuffled
  compile yields byte-identical graphs
- Belief runtime: C4/C5 contracts instrumented (REASON/REFLECT/EXPLORE +
  belief_scope); audit_belief_runtime 0 errors; data-model.md RATIONALE/REJECTED
- INV_1: all functions/classes have balanced #region/#endregion contracts
- 30 scenario tests pass; ruff clean
 2026-07-31 12:47:56 +03:00
busya
a32ca0631b feat(037): capture, verification lifecycle, inheritance + close 036 stabilization 
- Authoritative candidate capture with server-issued artifacts and raw-byte
  immutability hashing (source_response_hash server-owned)
- Closed-period lifecycle: request-hash bound approvals, persisted closure
  immutability violations, byte-for-byte catalog stability on reclosure
- Verification runs: persisted VerificationRun model + FK migration,
  publish gate (block_publish), scheduled observability runs (02:00 UTC)
- FR-013 baseline inheritance: prior_release_id migration, plan_inheritance/
  execute_inheritance classification and re-extraction, API endpoints
- Visual executor bound to release-deployment environment; caller mismatch
  rejected; visual SSIM/reconciliation modules
- Query execution decomposed: envelope/model/executor split, no direct SQL
- AgentRun approvals extracted to submodule; evidence adapter; _utils
- Dashboard testing service decomposed into 30+ modules (all <400 LOC)
- Five Feature-037 agent tools with permission guards (tools_037.py)
- API readiness endpoint; Alembic env/migrations; test fixture repos
- Specs 036/037 contracts, openapi.yaml, schema.json, tasks/traceability
  updated; semantic index rebuilt with 0 parse warnings
- Fix ADR-0003 parser ambiguity: remove [DEF🆔ADR] prose example
- Add axiom-mcp-agent-feedback.md: agent findings for MCP rework plan
- Tests: 298 service + 1464 API + 45 agent passing; ruff clean
 2026-07-31 11:28:50 +03:00
busya
2136082d6d feat(037): Phase 7 — Visual Baseline Support (T039-T047) 
- T039-T040: Schema ready (visualEntry + visualPolicy already in JSON schema)
- T041: visual_baseline.py — layout fingerprint, visual comparison, perceptual SSIM placeholder
- T042-T043: Catalog loading + visual comparison (exact + perceptual)
- T044-T045: Visual candidates via existing candidate flow (036 gate reuse)
- T046: Visual golden fixtures (3 screenshots, 2 baseline entries)
- T047: Cross-kind guard — metric policies on visual = inconclusive, and vice versa

65/65 tests pass. SPEC 037 COMPLETE: 47/47 tasks.
 2026-07-28 19:41:05 +03:00
busya
a74e7b084f feat(037): Phase 5 — US4 Baseline Candidate Lifecycle (T023-T032) 
- T023-T026: baseline_catalog.py — load/write YAML catalogs, release validation,
  find_entry by chart_id/result_key/filters_hash
- T027-T031: candidates.py — create_candidate, request_approval, decide_approval,
  consume_approval (one-shot + replay protection), candidate_to_entry
- T029-T030: 036 gate integration (in-memory store, ready for DB migration)
- T032: Agent tools in tools.py
- T023-T031 tests: 10 catalog + candidate tests

55/55 tests pass.
 2026-07-28 19:37:38 +03:00
busya
c2d5b67404 feat(037): Phase 4 — US3 Normalize and Compare (T016-T022) 
- T016-T017: normalization.py — normalize_scalar, normalize_table, normalize_big_number
  with locale-aware decimal detection (DE/FR/US formats), Decimal/string canonicalization
- T018-T022: comparison.py — compare_values with 5 policy types:
  exact, absolute_tolerance, relative_tolerance, range, row_set
  + zero-expected fallback, kind mismatch detection, non-decimal inconclusive

43/43 tests pass (14 normalization + 13 comparison + existing 16 from phases 1-3)
 2026-07-28 19:35:05 +03:00
busya
b8235ef2a1 feat(037): Phase 3 — US2 Superset-Native Execution (T011-T015) 
- T011: 4 NO-SQL tests (reject SQL, scalar execution, error taxonomy, temporal filter)
- T012: _chart_data.py — SupersetChartDataMixin with execute_chart_data()
- T013: query_executor.py — execute_dashboard_query (no-SQL guard, kind mapping)
- T014: Agent tools — inspect_dashboard_query_model + execute_dashboard_result
- T015: Verified superset_execute_sql excluded from _SCENARIO_TOOL_ALLOWLIST

4/4 executor tests pass. ChartDataMixin registered in SupersetClient.
Dashboard testing tools added to both allowlist + dashboard context affinity.
 2026-07-28 19:32:09 +03:00
busya
012f903a57 feat(037): Phase 2 US1 — Inspect Dashboard Query Model (T006-T010) 
- T006: 4 deterministic inspection tests (basic, deterministic, inaccessible, missing)
- T007: 6 filter normalization tests (scope, hash, order, locale)
- T008: query_model.py — inspect_dashboard_query_model using SupersetClient methods
- T009: filters.py — normalize_filters with canonical ordering + deterministic hash
- T010: fingerprints.py — SHA-256 helpers for query model and filter hashing

10/10 tests pass. Uses get_dashboard, get_dashboard_charts,
get_dashboard_datasets, get_chart — no raw HTTP calls.
 2026-07-28 19:27:47 +03:00
busya
9254299da0 feat(036): enforce mask_selectors in RegisterDraft + test 2026-07-28 19:05:55 +03:00
busya
f2d844cd91 test(036): evidence tests (4) + denial tests (5) — 80 backend, 14 frontend 2026-07-28 18:57:48 +03:00
busya
343d9e3917 test(036): artifact tests (12) + API tests (10) — 76/76 backend, fix RBAC Depends 2026-07-28 18:48:52 +03:00
busya
c9c6636ae8 test(036): event tests (12) + approval tests (11) — 54/54 backend 2026-07-28 11:42:54 +03:00
busya
6185e94d25 test(036): repository (11 tests) + agent context v2 (9 tests) + tool filter (6 tests) 2026-07-28 11:34:09 +03:00
busya
24fca2ec8a test(036): backend schemas (22 tests) + frontend AgentRunModel (13 tests) 2026-07-28 10:35:47 +03:00
busya
b9c0fa4c28 feat(git): BI-first UI/UX rework of /git — wizard modal, smart grid, undo, pre-flight 
Grid (Phase 1):
- Smart row actions by sync status (Connect Git / Save version (N) / Manage / Diagnose)
- "What changed" column with compact stacked category badges
- Status filter chips, sticky bulk panel with live progress, inline row errors + retry
- Dedupe repo-status batch requests (grid feeds the Repositories tab)
- DashboardDataGrid: new actionsCell snippet

GitManager modal (Phases 2-4):
- Simple/Full mode toggle (localStorage), Simple = 3-step wizard
  (Changes → Verify → Publish) via new GitWizardStepper
- Undo center + undo toast: soft-undo unpublished commit
  (backend POST /repositories/{ref}/undo-commit, reset --soft HEAD~1,
  409 guards for pushed/detached/empty HEAD)
- Commit draft autosave per dashboard slug
- Keyboard: Ctrl+Enter commit, 1/2/3 step/tab navigation
- Contextual "You are here: step N" help on the /git page

Excellence (Phase 5):
- First-run GuidedTour (4 spotlight steps, restartable from help panel)
- Pre-flight checklist before create/publish release (ConfirmDialog children
  + confirmDisabled; red checks block the action)
- Rollback confirmation with revert-preview diff; guided conflict progress bar
- Preview link to PREPROD before publishing; human-readable version labels
- prefers-reduced-motion guard; page <title>; aria-live wizard announcements
- docs/design/git-ux-glossary.md — canonical action verbs, ru/en normalized

UX fixes (user feedback):
- Compact change chips (vertical stack, 10px) — no table horizontal scroll
- "Insert into version description" button next to AI key-changes summary
- Guided recovery for "binding belongs to another Git server" (CTA to settings)
- Instant rollback button reveal (CSS visibility via :global, no opacity repaint)

QA fixes:
- Glossary compliance: 0 "commit/коммит" in user-facing strings
- Contract coverage for rollback functions in CommitHistory
- Rollback label aligned to glossary ("Откат к версии" / "Revert to version")

Tests: backend 446 git passed + 5 new undo-commit edge cases;
frontend 3204 passed (8 pre-existing failures unrelated: pipeline locale,
ConfirmationCard, PasswordPrompt, assistant_chat, test_tasks);
new GitReleasePanel pre-flight tests 3/3 green; vite build green.
 2026-07-22 18:06:26 +03:00
root
632b730fff chore: migrate GRACE-Poly anchors to hierarchical dotted naming 
Systematic rename of all semantic anchors (#region, [DEF], @RELATION)
across 1400+ files — backend Python, frontend Svelte/TS, specs, docs:
- Flat anchors become Namespace.Module.Entity
- @RELATION references updated to match new anchor paths
- Zero business logic changes
 2026-07-22 11:48:15 +03:00
busya
456d531a13 feat: harden migration flows and integration coverage 2026-07-21 18:59:26 +03:00
busya
49a566359a feat: translate module — runtime knobs, GRACE anchors, two-layer testing, QA fixes 
Implementation:
- Performance knobs: llm_batch_max_rows, llm_concurrency, insert_concurrency,
  multi_lang_mode, batch_aggressiveness, max_in_flight_batches
- Alembic migration f7a8b9c0d1e2 (idempotent, nullable, non-destructive)
- LLM provider capabilities: throughput_class, reasoning_control,
  supports_json_object, default/max_llm_concurrency
- TargetSchemaValidationRequest with conditional validator (sqllab/direct_db)
- Scheduler: background dispatch, local imports for lazy bootstrap

GRACE-Poly compliance:
- Semantic anchors on orchestrator_aggregator, orchestrator_sql, llm_provider
- Shared module _llm_http.py: _apply_reasoning_control extraction (INV_4)
- Alembic migration anchors per C3/C1 template
- Four renamed Svelte components: RunOutcomeSummary, DetectionQualityCard,
  LanguageStatList, SourceLanguageOverride

QA (this session):
- 11 backend test regressions fixed: spec'd MagicMock null fields for new
  columns, _check_translation_cache_bulk retarget, scheduler mock wiring,
  language_detection=auto assertions, token budget constant update
- 4 frontend eslint errors fixed: SvelteSet/SvelteDate imports,
  unused lang parameter, dead isTransientError
- Production bugfix: job_to_response() mapped 6 missing fields
- Pre-existing auth test failure documented (dependencies.py untouched)
- Axiom: 6440 contracts, 0 warnings
 2026-07-20 14:19:47 +03:00
busya
c3ad0afc17 refactor: remove rejected dataset review feature 2026-07-14 15:56:31 +03:00
busya
c2bd6cb441 feat(git): clarify dashboard release flow 2026-07-13 12:55:24 +03:00
busya
b39d9991b9 fix(logs): resolve old prod execution issues + full GRACE-Poly compliance 
- translate: pure-skip scheduled runs now COMPLETE (not FAILED); extract _compute_final_status
- scheduler: cron validation at update; pre-validate + skip legacy bad expr (/15*) in load
- git: robust get_branch_commits (existence check for 'prod' to avoid fatal)
- health: use explore for network/auth failures (reduces spam)
- Full semantic protocol updates across modules:
  - complete #region/#endregion + @BRIEF/@PRE/@POST/@INVARIANT/@RELATION/@RATIONALE
  - belief_scope + CoT markers (reason/explore/reflect)
  - semantics-testing compliance in tests
- Addresses exact symptoms from container_backend.log and backend/logs/app.*

Related to 039-dashboard-scenario-ui log analysis.
 2026-07-10 12:35:07 +03:00