sensemaking 0.24.3 → 0.24.5

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (131) hide show
  1. package/README.md +16 -12
  2. package/dist/cjs/cli/status.js +14 -24
  3. package/dist/cjs/cli/status.js.map +1 -1
  4. package/dist/cjs/commands/search.js +22 -0
  5. package/dist/cjs/commands/search.js.map +1 -1
  6. package/dist/cjs/config/types.js.map +1 -1
  7. package/dist/cjs/errors.d.cts +1 -1
  8. package/dist/cjs/errors.d.ts +1 -1
  9. package/dist/cjs/errors.js.map +1 -1
  10. package/dist/cjs/lib/worker-file.d.cts +1 -0
  11. package/dist/cjs/lib/worker-file.d.ts +1 -0
  12. package/dist/cjs/lib/worker-file.js +34 -0
  13. package/dist/cjs/lib/worker-file.js.map +1 -0
  14. package/dist/cjs/output/search-error.js +38 -0
  15. package/dist/cjs/output/search-error.js.map +1 -1
  16. package/dist/cjs/scan/pool.js +2 -23
  17. package/dist/cjs/scan/pool.js.map +1 -1
  18. package/dist/cjs/store/builder.d.cts +2 -1
  19. package/dist/cjs/store/builder.d.ts +2 -1
  20. package/dist/cjs/store/builder.js +4 -4
  21. package/dist/cjs/store/builder.js.map +1 -1
  22. package/dist/cjs/store/duckdb/lexical.d.cts +1 -0
  23. package/dist/cjs/store/duckdb/lexical.d.ts +1 -0
  24. package/dist/cjs/store/duckdb/lexical.js +28 -9
  25. package/dist/cjs/store/duckdb/lexical.js.map +1 -1
  26. package/dist/cjs/store/duckdb/native.d.cts +2 -0
  27. package/dist/cjs/store/duckdb/native.d.ts +2 -0
  28. package/dist/cjs/store/duckdb/native.js +11 -1
  29. package/dist/cjs/store/duckdb/native.js.map +1 -1
  30. package/dist/cjs/store/duckdb/ordered-bm25.d.cts +17 -0
  31. package/dist/cjs/store/duckdb/ordered-bm25.d.ts +17 -0
  32. package/dist/cjs/store/duckdb/ordered-bm25.js +238 -0
  33. package/dist/cjs/store/duckdb/ordered-bm25.js.map +1 -0
  34. package/dist/cjs/store/lock-wait.d.cts +2 -2
  35. package/dist/cjs/store/lock-wait.d.ts +2 -2
  36. package/dist/cjs/store/lock-wait.js +14 -11
  37. package/dist/cjs/store/lock-wait.js.map +1 -1
  38. package/dist/cjs/store/native.d.cts +4 -0
  39. package/dist/cjs/store/native.d.ts +4 -0
  40. package/dist/cjs/store/native.js +64 -9
  41. package/dist/cjs/store/native.js.map +1 -1
  42. package/dist/cjs/store/open.js +11 -7
  43. package/dist/cjs/store/open.js.map +1 -1
  44. package/dist/cjs/store/reconcile.d.cts +5 -1
  45. package/dist/cjs/store/reconcile.d.ts +5 -1
  46. package/dist/cjs/store/reconcile.js +5 -4
  47. package/dist/cjs/store/reconcile.js.map +1 -1
  48. package/dist/cjs/store/turso/connection.js +2 -2
  49. package/dist/cjs/store/turso/connection.js.map +1 -1
  50. package/dist/cjs/store/turso/lexical.js +5 -1
  51. package/dist/cjs/store/turso/lexical.js.map +1 -1
  52. package/dist/cjs/store/turso/reconcile.js +7 -6
  53. package/dist/cjs/store/turso/reconcile.js.map +1 -1
  54. package/dist/cjs/text/segment.js +168 -3
  55. package/dist/cjs/text/segment.js.map +1 -1
  56. package/dist/cjs/watch-claim.d.cts +39 -0
  57. package/dist/cjs/watch-claim.d.ts +39 -0
  58. package/dist/cjs/watch-claim.js +446 -0
  59. package/dist/cjs/watch-claim.js.map +1 -0
  60. package/dist/cjs/watch.js +240 -141
  61. package/dist/cjs/watch.js.map +1 -1
  62. package/dist/cjs/workers/watch-heartbeat.d.cts +1 -0
  63. package/dist/cjs/workers/watch-heartbeat.d.ts +1 -0
  64. package/dist/cjs/workers/watch-heartbeat.js +308 -0
  65. package/dist/cjs/workers/watch-heartbeat.js.map +1 -0
  66. package/dist/esm/cli/status.js +6 -5
  67. package/dist/esm/cli/status.js.map +1 -1
  68. package/dist/esm/commands/search.js +21 -0
  69. package/dist/esm/commands/search.js.map +1 -1
  70. package/dist/esm/config/types.js.map +1 -1
  71. package/dist/esm/errors.d.ts +1 -1
  72. package/dist/esm/errors.js.map +1 -1
  73. package/dist/esm/lib/worker-file.d.ts +1 -0
  74. package/dist/esm/lib/worker-file.js +24 -0
  75. package/dist/esm/lib/worker-file.js.map +1 -0
  76. package/dist/esm/output/search-error.js +21 -0
  77. package/dist/esm/output/search-error.js.map +1 -1
  78. package/dist/esm/scan/pool.js +2 -22
  79. package/dist/esm/scan/pool.js.map +1 -1
  80. package/dist/esm/store/builder.d.ts +2 -1
  81. package/dist/esm/store/builder.js +4 -4
  82. package/dist/esm/store/builder.js.map +1 -1
  83. package/dist/esm/store/duckdb/lexical.d.ts +1 -0
  84. package/dist/esm/store/duckdb/lexical.js +16 -7
  85. package/dist/esm/store/duckdb/lexical.js.map +1 -1
  86. package/dist/esm/store/duckdb/native.d.ts +2 -0
  87. package/dist/esm/store/duckdb/native.js +5 -1
  88. package/dist/esm/store/duckdb/native.js.map +1 -1
  89. package/dist/esm/store/duckdb/ordered-bm25.d.ts +17 -0
  90. package/dist/esm/store/duckdb/ordered-bm25.js +56 -0
  91. package/dist/esm/store/duckdb/ordered-bm25.js.map +1 -0
  92. package/dist/esm/store/lock-wait.d.ts +2 -2
  93. package/dist/esm/store/lock-wait.js +17 -14
  94. package/dist/esm/store/lock-wait.js.map +1 -1
  95. package/dist/esm/store/native.d.ts +4 -0
  96. package/dist/esm/store/native.js +60 -6
  97. package/dist/esm/store/native.js.map +1 -1
  98. package/dist/esm/store/open.js +9 -6
  99. package/dist/esm/store/open.js.map +1 -1
  100. package/dist/esm/store/reconcile.d.ts +5 -1
  101. package/dist/esm/store/reconcile.js +4 -3
  102. package/dist/esm/store/reconcile.js.map +1 -1
  103. package/dist/esm/store/turso/connection.js +1 -1
  104. package/dist/esm/store/turso/connection.js.map +1 -1
  105. package/dist/esm/store/turso/lexical.js +5 -1
  106. package/dist/esm/store/turso/lexical.js.map +1 -1
  107. package/dist/esm/store/turso/reconcile.js +8 -5
  108. package/dist/esm/store/turso/reconcile.js.map +1 -1
  109. package/dist/esm/text/segment.js +130 -4
  110. package/dist/esm/text/segment.js.map +1 -1
  111. package/dist/esm/watch-claim.d.ts +39 -0
  112. package/dist/esm/watch-claim.js +176 -0
  113. package/dist/esm/watch-claim.js.map +1 -0
  114. package/dist/esm/watch.js +168 -117
  115. package/dist/esm/watch.js.map +1 -1
  116. package/dist/esm/workers/watch-heartbeat.d.ts +1 -0
  117. package/dist/esm/workers/watch-heartbeat.js +106 -0
  118. package/dist/esm/workers/watch-heartbeat.js.map +1 -0
  119. package/package.json +2 -2
  120. package/skills/sense/SKILL.md +105 -104
  121. package/skills/sense/references/search.md +47 -0
  122. package/skills/sense/references/sql.md +120 -0
  123. package/skills/sense/references/stores/duckdb.md +35 -0
  124. package/skills/sense/references/stores/sqlite.md +66 -0
  125. package/skills/sense/references/stores/turso.md +36 -0
  126. package/skills/sense-setup/SKILL.md +93 -34
  127. package/skills/sense-setup/references/embeddings.md +51 -0
  128. package/skills/sense-setup/references/store-benchmarks.md +16 -0
  129. package/skills/sense-setup/references/stores/duckdb.md +28 -0
  130. package/skills/sense-setup/references/stores/sqlite.md +21 -0
  131. package/skills/sense-setup/references/stores/turso.md +25 -0
@@ -1 +1 @@
1
- {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/open.ts"],"sourcesContent":["import { channel } from 'node:diagnostics_channel';\nimport { mkdirSync } from 'node:fs';\nimport { join } from 'node:path';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { featureSignature, STATE_DIR } from '../config/index.ts';\nimport { rekeyChunkText } from '../embed/handoff.ts';\nimport { SenseError } from '../errors.ts';\nimport { FEATURES } from '../features/index.ts';\nimport { createBuilder } from './builder.ts';\nimport { clearCache } from './cache.ts';\nimport type { EmbedChangeKind } from './embed-scope.ts';\nimport { classifyEmbedChange } from './embed-scope.ts';\nimport type { FeatureToggle } from './feature-scope.ts';\nimport { classifyFeatureToggles, isFeatureOnlyChange } from './feature-scope.ts';\nimport { lockWaitBudgetMs } from './lock-wait.ts';\nimport { forcedPresetPaths, isPresetOnlyChange } from './preset-scope.ts';\nimport { getMeta, setMeta } from './shared.ts';\nimport { changedSignatureKeys, embedIdentityAdopted, signatureDiff } from './signature.ts';\nimport type { Stages } from './stages.ts';\nimport type { Connection, OpenDialect, Store } from './types.ts';\n\n// One open algorithm shared by every store, parameterised by a per-engine OpenDialect (types.ts).\n// Ordering is universal, not a dialect concern: connect, meta table, schema-version and\n// feature-signature comparison (rebuild, adopt, or partial rebuild), any extra rebuild trigger,\n// ensureSchema, the derived busy_timeout PRAGMA, reconcile, then the store itself.\n\nexport interface OpenResult {\n store: Store;\n cfg: ResolvedConfig;\n dbPath: string;\n parsed: number;\n warnings: string[];\n // Per-stage wall time for this open's build pass (stages.ts), not for any narrow embed or\n // feature invalidation that ran beside it -- on a cold build there is none, so it is the whole cost.\n stages: Stages;\n}\n\ninterface ConnectResult<Handle> {\n handle: Handle;\n conn: Connection;\n cfg: ResolvedConfig;\n dbPath: string;\n parsed: number;\n warnings: string[];\n stages: Stages;\n}\n\n// duckdb/turso hold the cache file for their connection's whole life, so a second command waits.\n// Bounded by this tree's recorded reconcile time (lock-wait.ts), not a fixed guess.\nconst LOCK_POLL_MS = 50;\nconst lockWait = channel('sensemaking.store.lock-wait');\n\nasync function connectUnlocked<Handle>(dbPath: string, cfg: ResolvedConfig, dialect: OpenDialect<Handle>): Promise<{ handle: Handle; conn: Connection }> {\n const budgetMs = lockWaitBudgetMs(cfg.configDir ?? cfg.baseDir);\n const deadline = Date.now() + budgetMs;\n for (;;) {\n try {\n return await dialect.connect(dbPath, cfg);\n } catch (err) {\n if (!dialect.isLocked?.(err as Error)) throw err;\n // Observe a real rejected native connection, never merely the start of an async open.\n if (lockWait.hasSubscribers) lockWait.publish({ store: cfg.store, dbPath });\n if (Date.now() >= deadline) {\n throw new SenseError('STORE_BUSY', `another sense process is using this tree's ${cfg.store} cache (${dbPath}) and did not release it within ${Math.round(budgetMs / 1000)}s; wait for that command to finish, or set \"store\" to \"sqlite\" in sense.config.json, which serves concurrent commands`);\n }\n await new Promise((resolve) => setTimeout(resolve, LOCK_POLL_MS));\n }\n }\n}\n\nasync function connectWithDialect<Handle>(cfg: ResolvedConfig, dialect: OpenDialect<Handle>): Promise<ConnectResult<Handle>> {\n const stateDir = join(cfg.configDir ?? cfg.baseDir, STATE_DIR);\n mkdirSync(stateDir, { recursive: true });\n const dbPath = join(stateDir, dialect.filename);\n\n const { handle, conn } = await connectUnlocked(dbPath, cfg, dialect);\n // A throw below must release this handle, or the leaked WAL makes the cache file undeletable on\n // Windows. `closed` keeps the catch from double-closing a handle a rebuild branch already closed.\n let closed = false;\n // Created here, not inside reconcile(): invalidate()/invalidateFeatures() below share its pool\n // with the build() call that follows them.\n const builder = createBuilder(conn, cfg, cfg.rootDir ?? cfg.baseDir, dialect.reconcileDialect);\n try {\n await conn.exec('CREATE TABLE IF NOT EXISTS meta (key TEXT PRIMARY KEY, value TEXT)');\n\n // Schema-version or feature-set mismatch: reconcile only reparses changed files, so an\n // old cache can't be patched incrementally -- rebuild instead (cheap: nothing expensive lives here).\n const version = await getMeta(conn, 'schema_version');\n const features = await getMeta(conn, 'features');\n const wantFeatures = featureSignature(cfg, FEATURES);\n // Set only by the preset-membership branch below; reconcile treats these paths as touched\n // despite an unchanged stamp, for this one build() call.\n let forcedPaths: Set<string> | undefined;\n // Set only by the embed-narrow branch below; applied via builder.invalidate() after ensureSchema.\n let embedInvalidate: EmbedChangeKind | undefined;\n // Set only by the feature-toggle branch below; applied via builder.invalidateFeatures() after ensureSchema.\n let featureToggles: FeatureToggle[] | undefined;\n if ((version !== null && version !== dialect.schemaVersion) || (features !== null && features !== wantFeatures)) {\n // Indexing derives from presets, so a config edit rebuilding the cache must say so and\n // name what changed -- silent rebuilds make derived indexing look like a hang or a bug.\n if (version !== null && version !== dialect.schemaVersion) {\n console.error('sense: cache format changed (new sensemaking version); rebuilding the index');\n // Close before clearCache deletes the files out from under the still-open handle.\n closed = true;\n await dialect.close(handle);\n clearCache(cfg);\n return connectWithDialect(cfg, dialect);\n }\n const changedKeys = changedSignatureKeys(features ?? '', wantFeatures);\n // Null (rather than the empty set) when changedKeys isn't preset-only, or a changed\n // preset's stored segment can't be parsed back into old include/exclude/vectors -- either\n // way the branch below is skipped and the full rebuild stays the fallback.\n const presetForced = isPresetOnlyChange(changedKeys) ? forcedPresetPaths(cfg, cfg.rootDir ?? cfg.baseDir, features ?? '', changedKeys) : null;\n // Null unless exactly one segment changed and it decomposes into a recognised embed-only\n // case (embed-scope.ts): anything else, including embed toggled on/off, stays null.\n const embedKind = changedKeys.size === 1 && changedKeys.has('embed') ? classifyEmbedChange(features ?? '', wantFeatures) : null;\n // Null (rather than the empty set) when changedKeys isn't feature-toggle-only, or a changed\n // feature's stored segment can't be parsed back into old/new on-off -- either way the branch\n // below is skipped and the full rebuild stays the fallback.\n const toggles = isFeatureOnlyChange(changedKeys) ? classifyFeatureToggles(features ?? '', wantFeatures, changedKeys) : null;\n if (changedKeys.size === 1 && changedKeys.has('embed') && embedIdentityAdopted(features ?? '', wantFeatures)) {\n // First sight of a resolved weight identity: the model itself hasn't changed, so\n // adopt it into meta with no rebuild and no re-embed.\n console.error(\"sense: recorded the embedding model's resolved identity; vectors are unaffected\");\n await setMeta(conn, 'features', wantFeatures);\n } else if (embedKind !== null) {\n // Chunk boundaries are file-derived, not model-derived (embed.ts's chunksOf), so neither\n // case needs any file reparsed for any feature but embed itself.\n embedInvalidate = embedKind;\n console.error(embedKind === 'model' ? 'sense: config change (embed settings) invalidates only the stale vectors' : 'sense: config change (embed settings) rebuilds only the embeddings it affects');\n await setMeta(conn, 'features', wantFeatures);\n } else if (presetForced !== null) {\n // Membership only: which files a preset covers moved, not a file's own content, a\n // global feature, or the embed model. reconcile() already knows how to\n // add, update, and remove a file correctly (every cross-feature cascade included); it\n // just needs telling which unchanged files to treat as touched (forcedPaths, above).\n forcedPaths = presetForced;\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) reparses only the files it affects`);\n await setMeta(conn, 'features', wantFeatures);\n } else if (toggles !== null) {\n // A feature turned off or on: reconcile.ts's activeFeatures skips a disabled feature's\n // hooks entirely, so its rows rot while off -- invalidateFeatures drops them and, for a\n // feature turning back on, fully re-derives across every indexed file rather than\n // trusting rows written before or during that window.\n featureToggles = toggles;\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) reparses only the feature it affects`);\n await setMeta(conn, 'features', wantFeatures);\n } else {\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) rebuilds the index`);\n closed = true;\n await dialect.close(handle);\n clearCache(cfg);\n return connectWithDialect(cfg, dialect);\n }\n }\n\n await dialect.ensureSchema(handle, conn, cfg);\n\n // 3x the largest reconcile this cache has recorded, floored at 30s and capped at 10min.\n // Installed before build() -- that call is the one that races a watcher's transaction.\n if (dialect.setDerivedBusyTimeout) {\n const recordedMaxMs = Number((await getMeta(conn, 'reconcile_max_ms')) ?? '0');\n await dialect.setDerivedBusyTimeout(handle, conn, Math.min(Math.max(30000, 3 * recordedMaxMs), 600_000));\n }\n\n // Runs before build(): forcedPaths, embedInvalidate and featureToggles never target the same\n // call (each comes from its own mutually exclusive branch above: a preset-only, embed-only, or\n // feature-toggle-only changed-key set).\n const embedParsed = embedInvalidate ? (await builder.invalidate(embedInvalidate)).parsed : 0;\n const featureParsed = featureToggles ? (await builder.invalidateFeatures(featureToggles)).parsed : 0;\n\n const { parsed, warnings, stages } = await builder.build(forcedPaths);\n // Every open reconciles once and is done with it -- a watcher opens fresh per event instead of holding this connection (src/watch.ts).\n await builder.close();\n\n return { handle, conn, cfg, dbPath, parsed: parsed + embedParsed + featureParsed, warnings, stages };\n } catch (err) {\n await builder.close();\n if (!closed) await dialect.close(handle);\n throw err;\n }\n}\n\n// A store's open(): connects (see connectWithDialect above), then wraps the resulting connection\n// in the Store interface. The builder's pool closes right after the initial build.\nexport async function openWithDialect<Handle>(cfg: ResolvedConfig, dialect: OpenDialect<Handle>): Promise<OpenResult> {\n // `baseDir` was the original public open() input. Keep direct API callers working while\n // making the two ownership paths explicit for every internal operation.\n if (!cfg.rootDir) {\n if (!cfg.baseDir) throw new SenseError('CONFIG_INVALID', 'open requires rootDir (or legacy baseDir)');\n cfg.rootDir = cfg.baseDir;\n }\n if (!cfg.configDir) cfg.configDir = cfg.rootDir;\n const { handle, conn, cfg: resolvedCfg, dbPath, parsed, warnings, stages } = await connectWithDialect(cfg, dialect);\n const store = dialect.createStore(handle, conn, resolvedCfg);\n // reconcile ran before this object existed, so its chunk text is keyed by the connection.\n rekeyChunkText(conn, store);\n return { store, cfg: resolvedCfg, dbPath, parsed, warnings, stages };\n}\n\nexport async function docCount(store: Store): Promise<number> {\n const stmt = await store.prepare('SELECT COUNT(*) AS n FROM frontmatter');\n return ((await stmt.get()) as { n: number }).n;\n}\n"],"names":["channel","mkdirSync","join","featureSignature","STATE_DIR","rekeyChunkText","SenseError","FEATURES","createBuilder","clearCache","classifyEmbedChange","classifyFeatureToggles","isFeatureOnlyChange","lockWaitBudgetMs","forcedPresetPaths","isPresetOnlyChange","getMeta","setMeta","changedSignatureKeys","embedIdentityAdopted","signatureDiff","LOCK_POLL_MS","lockWait","connectUnlocked","dbPath","cfg","dialect","budgetMs","configDir","baseDir","deadline","Date","now","connect","err","isLocked","hasSubscribers","publish","store","Math","round","Promise","resolve","setTimeout","connectWithDialect","stateDir","recursive","filename","handle","conn","closed","builder","rootDir","reconcileDialect","exec","version","features","wantFeatures","forcedPaths","embedInvalidate","featureToggles","schemaVersion","console","error","close","changedKeys","presetForced","embedKind","size","has","toggles","changed","ensureSchema","setDerivedBusyTimeout","recordedMaxMs","Number","min","max","embedParsed","invalidate","parsed","featureParsed","invalidateFeatures","warnings","stages","build","openWithDialect","resolvedCfg","createStore","docCount","stmt","prepare","get","n"],"mappings":"AAAA,SAASA,OAAO,QAAQ,2BAA2B;AACnD,SAASC,SAAS,QAAQ,UAAU;AACpC,SAASC,IAAI,QAAQ,YAAY;AAEjC,SAASC,gBAAgB,EAAEC,SAAS,QAAQ,qBAAqB;AACjE,SAASC,cAAc,QAAQ,sBAAsB;AACrD,SAASC,UAAU,QAAQ,eAAe;AAC1C,SAASC,QAAQ,QAAQ,uBAAuB;AAChD,SAASC,aAAa,QAAQ,eAAe;AAC7C,SAASC,UAAU,QAAQ,aAAa;AAExC,SAASC,mBAAmB,QAAQ,mBAAmB;AAEvD,SAASC,sBAAsB,EAAEC,mBAAmB,QAAQ,qBAAqB;AACjF,SAASC,gBAAgB,QAAQ,iBAAiB;AAClD,SAASC,iBAAiB,EAAEC,kBAAkB,QAAQ,oBAAoB;AAC1E,SAASC,OAAO,EAAEC,OAAO,QAAQ,cAAc;AAC/C,SAASC,oBAAoB,EAAEC,oBAAoB,EAAEC,aAAa,QAAQ,iBAAiB;AA8B3F,iGAAiG;AACjG,oFAAoF;AACpF,MAAMC,eAAe;AACrB,MAAMC,WAAWtB,QAAQ;AAEzB,eAAeuB,gBAAwBC,MAAc,EAAEC,GAAmB,EAAEC,OAA4B;QACpED;IAAlC,MAAME,WAAWd,kBAAiBY,iBAAAA,IAAIG,SAAS,cAAbH,4BAAAA,iBAAiBA,IAAII,OAAO;IAC9D,MAAMC,WAAWC,KAAKC,GAAG,KAAKL;IAC9B,OAAS;QACP,IAAI;YACF,OAAO,MAAMD,QAAQO,OAAO,CAACT,QAAQC;QACvC,EAAE,OAAOS,KAAK;gBACPR;YAAL,IAAI,GAACA,oBAAAA,QAAQS,QAAQ,cAAhBT,wCAAAA,uBAAAA,SAAmBQ,OAAe,MAAMA;YAC7C,sFAAsF;YACtF,IAAIZ,SAASc,cAAc,EAAEd,SAASe,OAAO,CAAC;gBAAEC,OAAOb,IAAIa,KAAK;gBAAEd;YAAO;YACzE,IAAIO,KAAKC,GAAG,MAAMF,UAAU;gBAC1B,MAAM,IAAIxB,WAAW,cAAc,CAAC,2CAA2C,EAAEmB,IAAIa,KAAK,CAAC,QAAQ,EAAEd,OAAO,gCAAgC,EAAEe,KAAKC,KAAK,CAACb,WAAW,MAAM,qHAAqH,CAAC;YAClS;YACA,MAAM,IAAIc,QAAQ,CAACC,UAAYC,WAAWD,SAASrB;QACrD;IACF;AACF;AAEA,eAAeuB,mBAA2BnB,GAAmB,EAAEC,OAA4B;QACnED,gBAUmBA;IAVzC,MAAMoB,WAAW3C,MAAKuB,iBAAAA,IAAIG,SAAS,cAAbH,4BAAAA,iBAAiBA,IAAII,OAAO,EAAEzB;IACpDH,UAAU4C,UAAU;QAAEC,WAAW;IAAK;IACtC,MAAMtB,SAAStB,KAAK2C,UAAUnB,QAAQqB,QAAQ;IAE9C,MAAM,EAAEC,MAAM,EAAEC,IAAI,EAAE,GAAG,MAAM1B,gBAAgBC,QAAQC,KAAKC;IAC5D,gGAAgG;IAChG,kGAAkG;IAClG,IAAIwB,SAAS;IACb,+FAA+F;IAC/F,2CAA2C;IAC3C,MAAMC,UAAU3C,cAAcyC,MAAMxB,MAAKA,eAAAA,IAAI2B,OAAO,cAAX3B,0BAAAA,eAAeA,IAAII,OAAO,EAAEH,QAAQ2B,gBAAgB;IAC7F,IAAI;QACF,MAAMJ,KAAKK,IAAI,CAAC;QAEhB,uFAAuF;QACvF,qGAAqG;QACrG,MAAMC,UAAU,MAAMvC,QAAQiC,MAAM;QACpC,MAAMO,WAAW,MAAMxC,QAAQiC,MAAM;QACrC,MAAMQ,eAAetD,iBAAiBsB,KAAKlB;QAC3C,0FAA0F;QAC1F,yDAAyD;QACzD,IAAImD;QACJ,kGAAkG;QAClG,IAAIC;QACJ,4GAA4G;QAC5G,IAAIC;QACJ,IAAI,AAACL,YAAY,QAAQA,YAAY7B,QAAQmC,aAAa,IAAML,aAAa,QAAQA,aAAaC,cAAe;gBAejChC;YAd9E,uFAAuF;YACvF,wFAAwF;YACxF,IAAI8B,YAAY,QAAQA,YAAY7B,QAAQmC,aAAa,EAAE;gBACzDC,QAAQC,KAAK,CAAC;gBACd,kFAAkF;gBAClFb,SAAS;gBACT,MAAMxB,QAAQsC,KAAK,CAAChB;gBACpBvC,WAAWgB;gBACX,OAAOmB,mBAAmBnB,KAAKC;YACjC;YACA,MAAMuC,cAAc/C,qBAAqBsC,qBAAAA,sBAAAA,WAAY,IAAIC;YACzD,oFAAoF;YACpF,0FAA0F;YAC1F,2EAA2E;YAC3E,MAAMS,eAAenD,mBAAmBkD,eAAenD,kBAAkBW,MAAKA,gBAAAA,IAAI2B,OAAO,cAAX3B,2BAAAA,gBAAeA,IAAII,OAAO,EAAE2B,qBAAAA,sBAAAA,WAAY,IAAIS,eAAe;YACzI,yFAAyF;YACzF,oFAAoF;YACpF,MAAME,YAAYF,YAAYG,IAAI,KAAK,KAAKH,YAAYI,GAAG,CAAC,WAAW3D,oBAAoB8C,qBAAAA,sBAAAA,WAAY,IAAIC,gBAAgB;YAC3H,4FAA4F;YAC5F,6FAA6F;YAC7F,4DAA4D;YAC5D,MAAMa,UAAU1D,oBAAoBqD,eAAetD,uBAAuB6C,qBAAAA,sBAAAA,WAAY,IAAIC,cAAcQ,eAAe;YACvH,IAAIA,YAAYG,IAAI,KAAK,KAAKH,YAAYI,GAAG,CAAC,YAAYlD,qBAAqBqC,qBAAAA,sBAAAA,WAAY,IAAIC,eAAe;gBAC5G,iFAAiF;gBACjF,sDAAsD;gBACtDK,QAAQC,KAAK,CAAC;gBACd,MAAM9C,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIU,cAAc,MAAM;gBAC7B,yFAAyF;gBACzF,iEAAiE;gBACjER,kBAAkBQ;gBAClBL,QAAQC,KAAK,CAACI,cAAc,UAAU,6EAA6E;gBACnH,MAAMlD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIS,iBAAiB,MAAM;gBAChC,kFAAkF;gBAClF,uEAAuE;gBACvE,sFAAsF;gBACtF,qFAAqF;gBACrFR,cAAcQ;gBACd,MAAMK,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,oCAAoC,CAAC;gBACpF,MAAMtD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIa,YAAY,MAAM;gBAC3B,uFAAuF;gBACvF,wFAAwF;gBACxF,kFAAkF;gBAClF,sDAAsD;gBACtDV,iBAAiBU;gBACjB,MAAMC,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,sCAAsC,CAAC;gBACtF,MAAMtD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO;gBACL,MAAMc,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,oBAAoB,CAAC;gBACpErB,SAAS;gBACT,MAAMxB,QAAQsC,KAAK,CAAChB;gBACpBvC,WAAWgB;gBACX,OAAOmB,mBAAmBnB,KAAKC;YACjC;QACF;QAEA,MAAMA,QAAQ8C,YAAY,CAACxB,QAAQC,MAAMxB;QAEzC,wFAAwF;QACxF,uFAAuF;QACvF,IAAIC,QAAQ+C,qBAAqB,EAAE;gBACH;YAA9B,MAAMC,gBAAgBC,QAAQ,OAAA,MAAM3D,QAAQiC,MAAM,iCAApB,kBAAA,OAA4C;YAC1E,MAAMvB,QAAQ+C,qBAAqB,CAACzB,QAAQC,MAAMV,KAAKqC,GAAG,CAACrC,KAAKsC,GAAG,CAAC,OAAO,IAAIH,gBAAgB;QACjG;QAEA,6FAA6F;QAC7F,+FAA+F;QAC/F,wCAAwC;QACxC,MAAMI,cAAcnB,kBAAkB,AAAC,CAAA,MAAMR,QAAQ4B,UAAU,CAACpB,gBAAe,EAAGqB,MAAM,GAAG;QAC3F,MAAMC,gBAAgBrB,iBAAiB,AAAC,CAAA,MAAMT,QAAQ+B,kBAAkB,CAACtB,eAAc,EAAGoB,MAAM,GAAG;QAEnG,MAAM,EAAEA,MAAM,EAAEG,QAAQ,EAAEC,MAAM,EAAE,GAAG,MAAMjC,QAAQkC,KAAK,CAAC3B;QACzD,uIAAuI;QACvI,MAAMP,QAAQa,KAAK;QAEnB,OAAO;YAAEhB;YAAQC;YAAMxB;YAAKD;YAAQwD,QAAQA,SAASF,cAAcG;YAAeE;YAAUC;QAAO;IACrG,EAAE,OAAOlD,KAAK;QACZ,MAAMiB,QAAQa,KAAK;QACnB,IAAI,CAACd,QAAQ,MAAMxB,QAAQsC,KAAK,CAAChB;QACjC,MAAMd;IACR;AACF;AAEA,iGAAiG;AACjG,mFAAmF;AACnF,OAAO,eAAeoD,gBAAwB7D,GAAmB,EAAEC,OAA4B;IAC7F,wFAAwF;IACxF,wEAAwE;IACxE,IAAI,CAACD,IAAI2B,OAAO,EAAE;QAChB,IAAI,CAAC3B,IAAII,OAAO,EAAE,MAAM,IAAIvB,WAAW,kBAAkB;QACzDmB,IAAI2B,OAAO,GAAG3B,IAAII,OAAO;IAC3B;IACA,IAAI,CAACJ,IAAIG,SAAS,EAAEH,IAAIG,SAAS,GAAGH,IAAI2B,OAAO;IAC/C,MAAM,EAAEJ,MAAM,EAAEC,IAAI,EAAExB,KAAK8D,WAAW,EAAE/D,MAAM,EAAEwD,MAAM,EAAEG,QAAQ,EAAEC,MAAM,EAAE,GAAG,MAAMxC,mBAAmBnB,KAAKC;IAC3G,MAAMY,QAAQZ,QAAQ8D,WAAW,CAACxC,QAAQC,MAAMsC;IAChD,0FAA0F;IAC1FlF,eAAe4C,MAAMX;IACrB,OAAO;QAAEA;QAAOb,KAAK8D;QAAa/D;QAAQwD;QAAQG;QAAUC;IAAO;AACrE;AAEA,OAAO,eAAeK,SAASnD,KAAY;IACzC,MAAMoD,OAAO,MAAMpD,MAAMqD,OAAO,CAAC;IACjC,OAAO,AAAE,CAAA,MAAMD,KAAKE,GAAG,EAAC,EAAqBC,CAAC;AAChD"}
1
+ {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/open.ts"],"sourcesContent":["import { channel } from 'node:diagnostics_channel';\nimport { mkdirSync } from 'node:fs';\nimport { join } from 'node:path';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { featureSignature, STATE_DIR } from '../config/index.ts';\nimport { rekeyChunkText } from '../embed/handoff.ts';\nimport { SenseError } from '../errors.ts';\nimport { FEATURES } from '../features/index.ts';\nimport { createBuilder } from './builder.ts';\nimport { clearCache } from './cache.ts';\nimport type { EmbedChangeKind } from './embed-scope.ts';\nimport { classifyEmbedChange } from './embed-scope.ts';\nimport type { FeatureToggle } from './feature-scope.ts';\nimport { classifyFeatureToggles, isFeatureOnlyChange } from './feature-scope.ts';\nimport { lockWaitBudgetMs } from './lock-wait.ts';\nimport { forcedPresetPaths, isPresetOnlyChange } from './preset-scope.ts';\nimport { getMeta, setMeta } from './shared.ts';\nimport { changedSignatureKeys, embedIdentityAdopted, signatureDiff } from './signature.ts';\nimport type { Stages } from './stages.ts';\nimport type { Connection, OpenDialect, Store } from './types.ts';\n\n// One open algorithm shared by every store, parameterised by a per-engine OpenDialect (types.ts).\n// Ordering is universal, not a dialect concern: connect, meta table, schema-version and\n// feature-signature comparison (rebuild, adopt, or partial rebuild), any extra rebuild trigger,\n// ensureSchema, the derived busy_timeout PRAGMA, reconcile, then the store itself.\n\nexport interface OpenResult {\n store: Store;\n cfg: ResolvedConfig;\n dbPath: string;\n parsed: number;\n warnings: string[];\n // Per-stage wall time for this open's build pass (stages.ts), not for any narrow embed or\n // feature invalidation that ran beside it -- on a cold build there is none, so it is the whole cost.\n stages: Stages;\n}\n\ninterface ConnectResult<Handle> {\n handle: Handle;\n conn: Connection;\n cfg: ResolvedConfig;\n dbPath: string;\n parsed: number;\n warnings: string[];\n stages: Stages;\n}\n\n// duckdb/turso hold the cache file for their connection's whole life, so a second command waits.\n// Bounded by this tree's recorded reconcile time (lock-wait.ts), not a fixed guess.\nconst LOCK_POLL_MS = 50;\nconst lockWait = channel('sensemaking.store.lock-wait');\n\nasync function connectUnlocked<Handle>(dbPath: string, cfg: ResolvedConfig, configDir: string, dialect: OpenDialect<Handle>): Promise<{ handle: Handle; conn: Connection }> {\n const budgetMs = lockWaitBudgetMs(configDir);\n const deadline = Date.now() + budgetMs;\n for (;;) {\n try {\n return await dialect.connect(dbPath, cfg);\n } catch (err) {\n if (!dialect.isLocked?.(err as Error)) throw err;\n // Observe a real rejected native connection, never merely the start of an async open.\n if (lockWait.hasSubscribers) lockWait.publish({ store: cfg.store, dbPath });\n if (Date.now() >= deadline) {\n throw new SenseError('STORE_BUSY', `another sense process is using this tree's ${cfg.store} cache (${dbPath}) and did not release it within ${Math.round(budgetMs / 1000)}s; wait for that command to finish, or set \"store\" to \"sqlite\" in sense.config.json, which serves concurrent commands`);\n }\n await new Promise((resolve) => setTimeout(resolve, LOCK_POLL_MS));\n }\n }\n}\n\nasync function connectWithDialect<Handle>(cfg: ResolvedConfig, dialect: OpenDialect<Handle>): Promise<ConnectResult<Handle>> {\n const configDir = cfg.configDir ?? cfg.baseDir;\n const stateDir = join(configDir, STATE_DIR);\n mkdirSync(stateDir, { recursive: true });\n const dbPath = join(stateDir, dialect.filename);\n\n const { handle, conn } = await connectUnlocked(dbPath, cfg, configDir, dialect);\n // A throw below must release this handle, or the leaked WAL makes the cache file undeletable on\n // Windows. `closed` keeps the catch from double-closing a handle a rebuild branch already closed.\n let closed = false;\n // Created here, not inside reconcile(): invalidate()/invalidateFeatures() below share its pool\n // with the build() call that follows them.\n const builder = createBuilder(conn, cfg, { rootDir: cfg.rootDir ?? cfg.baseDir, configDir }, dialect.reconcileDialect);\n try {\n await conn.exec('CREATE TABLE IF NOT EXISTS meta (key TEXT PRIMARY KEY, value TEXT)');\n\n // Schema-version or feature-set mismatch: reconcile only reparses changed files, so an\n // old cache can't be patched incrementally -- rebuild instead (cheap: nothing expensive lives here).\n const version = await getMeta(conn, 'schema_version');\n const features = await getMeta(conn, 'features');\n const wantFeatures = featureSignature(cfg, FEATURES);\n // Set only by the preset-membership branch below; reconcile treats these paths as touched\n // despite an unchanged stamp, for this one build() call.\n let forcedPaths: Set<string> | undefined;\n // Set only by the embed-narrow branch below; applied via builder.invalidate() after ensureSchema.\n let embedInvalidate: EmbedChangeKind | undefined;\n // Set only by the feature-toggle branch below; applied via builder.invalidateFeatures() after ensureSchema.\n let featureToggles: FeatureToggle[] | undefined;\n if ((version !== null && version !== dialect.schemaVersion) || (features !== null && features !== wantFeatures)) {\n // Indexing derives from presets, so a config edit rebuilding the cache must say so and\n // name what changed -- silent rebuilds make derived indexing look like a hang or a bug.\n if (version !== null && version !== dialect.schemaVersion) {\n console.error('sense: cache format changed (new sensemaking version); rebuilding the index');\n // Close before clearCache deletes the files out from under the still-open handle.\n closed = true;\n await dialect.close(handle);\n clearCache(cfg);\n return connectWithDialect(cfg, dialect);\n }\n const changedKeys = changedSignatureKeys(features ?? '', wantFeatures);\n // Null (rather than the empty set) when changedKeys isn't preset-only, or a changed\n // preset's stored segment can't be parsed back into old include/exclude/vectors -- either\n // way the branch below is skipped and the full rebuild stays the fallback.\n const presetForced = isPresetOnlyChange(changedKeys) ? forcedPresetPaths(cfg, cfg.rootDir ?? cfg.baseDir, features ?? '', changedKeys) : null;\n // Null unless exactly one segment changed and it decomposes into a recognised embed-only\n // case (embed-scope.ts): anything else, including embed toggled on/off, stays null.\n const embedKind = changedKeys.size === 1 && changedKeys.has('embed') ? classifyEmbedChange(features ?? '', wantFeatures) : null;\n // Null (rather than the empty set) when changedKeys isn't feature-toggle-only, or a changed\n // feature's stored segment can't be parsed back into old/new on-off -- either way the branch\n // below is skipped and the full rebuild stays the fallback.\n const toggles = isFeatureOnlyChange(changedKeys) ? classifyFeatureToggles(features ?? '', wantFeatures, changedKeys) : null;\n if (changedKeys.size === 1 && changedKeys.has('embed') && embedIdentityAdopted(features ?? '', wantFeatures)) {\n // First sight of a resolved weight identity: the model itself hasn't changed, so\n // adopt it into meta with no rebuild and no re-embed.\n console.error(\"sense: recorded the embedding model's resolved identity; vectors are unaffected\");\n await setMeta(conn, 'features', wantFeatures);\n } else if (embedKind !== null) {\n // Chunk boundaries are file-derived, not model-derived (embed.ts's chunksOf), so neither\n // case needs any file reparsed for any feature but embed itself.\n embedInvalidate = embedKind;\n console.error(embedKind === 'model' ? 'sense: config change (embed settings) invalidates only the stale vectors' : 'sense: config change (embed settings) rebuilds only the embeddings it affects');\n await setMeta(conn, 'features', wantFeatures);\n } else if (presetForced !== null) {\n // Membership only: which files a preset covers moved, not a file's own content, a\n // global feature, or the embed model. reconcile() already knows how to\n // add, update, and remove a file correctly (every cross-feature cascade included); it\n // just needs telling which unchanged files to treat as touched (forcedPaths, above).\n forcedPaths = presetForced;\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) reparses only the files it affects`);\n await setMeta(conn, 'features', wantFeatures);\n } else if (toggles !== null) {\n // A feature turned off or on: reconcile.ts's activeFeatures skips a disabled feature's\n // hooks entirely, so its rows rot while off -- invalidateFeatures drops them and, for a\n // feature turning back on, fully re-derives across every indexed file rather than\n // trusting rows written before or during that window.\n featureToggles = toggles;\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) reparses only the feature it affects`);\n await setMeta(conn, 'features', wantFeatures);\n } else {\n const changed = signatureDiff(features ?? '', wantFeatures);\n console.error(`sense: config change (${changed}) rebuilds the index`);\n closed = true;\n await dialect.close(handle);\n clearCache(cfg);\n return connectWithDialect(cfg, dialect);\n }\n }\n\n await dialect.ensureSchema(handle, conn, cfg);\n\n // 3x the largest reconcile this cache has recorded, floored at 30s and capped at 10min.\n // Installed before build() -- that call is the one that races a watcher's transaction.\n if (dialect.setDerivedBusyTimeout) {\n const recordedMaxMs = Number((await getMeta(conn, 'reconcile_max_ms')) ?? '0');\n await dialect.setDerivedBusyTimeout(handle, conn, Math.min(Math.max(30000, 3 * recordedMaxMs), 600_000));\n }\n\n // Runs before build(): forcedPaths, embedInvalidate and featureToggles never target the same\n // call (each comes from its own mutually exclusive branch above: a preset-only, embed-only, or\n // feature-toggle-only changed-key set).\n const embedParsed = embedInvalidate ? (await builder.invalidate(embedInvalidate)).parsed : 0;\n const featureParsed = featureToggles ? (await builder.invalidateFeatures(featureToggles)).parsed : 0;\n\n const { parsed, warnings, stages } = await builder.build(forcedPaths);\n // Every open reconciles once and is done with it -- a watcher opens fresh per event instead of holding this connection (src/watch.ts).\n await builder.close();\n\n return { handle, conn, cfg, dbPath, parsed: parsed + embedParsed + featureParsed, warnings, stages };\n } catch (err) {\n await builder.close();\n if (!closed) await dialect.close(handle);\n throw err;\n }\n}\n\n// A store's open(): connects (see connectWithDialect above), then wraps the resulting connection\n// in the Store interface. The builder's pool closes right after the initial build.\nexport async function openWithDialect<Handle>(cfg: ResolvedConfig, dialect: OpenDialect<Handle>): Promise<OpenResult> {\n // `baseDir` was the original public open() input. Keep direct API callers working while\n // making the two ownership paths explicit for every internal operation.\n if (!cfg.rootDir) {\n if (!cfg.baseDir) throw new SenseError('CONFIG_INVALID', 'open requires rootDir (or legacy baseDir)');\n cfg.rootDir = cfg.baseDir;\n }\n if (!cfg.configDir) cfg.configDir = cfg.rootDir;\n const { handle, conn, cfg: resolvedCfg, dbPath, parsed, warnings, stages } = await connectWithDialect(cfg, dialect);\n const store = dialect.createStore(handle, conn, resolvedCfg);\n // reconcile ran before this object existed, so its chunk text is keyed by the connection.\n rekeyChunkText(conn, store);\n return { store, cfg: resolvedCfg, dbPath, parsed, warnings, stages };\n}\n\nexport async function docCount(store: Store): Promise<number> {\n const stmt = await store.prepare('SELECT COUNT(*) AS n FROM frontmatter');\n return ((await stmt.get()) as { n: number }).n;\n}\n"],"names":["channel","mkdirSync","join","featureSignature","STATE_DIR","rekeyChunkText","SenseError","FEATURES","createBuilder","clearCache","classifyEmbedChange","classifyFeatureToggles","isFeatureOnlyChange","lockWaitBudgetMs","forcedPresetPaths","isPresetOnlyChange","getMeta","setMeta","changedSignatureKeys","embedIdentityAdopted","signatureDiff","LOCK_POLL_MS","lockWait","connectUnlocked","dbPath","cfg","configDir","dialect","budgetMs","deadline","Date","now","connect","err","isLocked","hasSubscribers","publish","store","Math","round","Promise","resolve","setTimeout","connectWithDialect","baseDir","stateDir","recursive","filename","handle","conn","closed","builder","rootDir","reconcileDialect","exec","version","features","wantFeatures","forcedPaths","embedInvalidate","featureToggles","schemaVersion","console","error","close","changedKeys","presetForced","embedKind","size","has","toggles","changed","ensureSchema","setDerivedBusyTimeout","recordedMaxMs","Number","min","max","embedParsed","invalidate","parsed","featureParsed","invalidateFeatures","warnings","stages","build","openWithDialect","resolvedCfg","createStore","docCount","stmt","prepare","get","n"],"mappings":"AAAA,SAASA,OAAO,QAAQ,2BAA2B;AACnD,SAASC,SAAS,QAAQ,UAAU;AACpC,SAASC,IAAI,QAAQ,YAAY;AAEjC,SAASC,gBAAgB,EAAEC,SAAS,QAAQ,qBAAqB;AACjE,SAASC,cAAc,QAAQ,sBAAsB;AACrD,SAASC,UAAU,QAAQ,eAAe;AAC1C,SAASC,QAAQ,QAAQ,uBAAuB;AAChD,SAASC,aAAa,QAAQ,eAAe;AAC7C,SAASC,UAAU,QAAQ,aAAa;AAExC,SAASC,mBAAmB,QAAQ,mBAAmB;AAEvD,SAASC,sBAAsB,EAAEC,mBAAmB,QAAQ,qBAAqB;AACjF,SAASC,gBAAgB,QAAQ,iBAAiB;AAClD,SAASC,iBAAiB,EAAEC,kBAAkB,QAAQ,oBAAoB;AAC1E,SAASC,OAAO,EAAEC,OAAO,QAAQ,cAAc;AAC/C,SAASC,oBAAoB,EAAEC,oBAAoB,EAAEC,aAAa,QAAQ,iBAAiB;AA8B3F,iGAAiG;AACjG,oFAAoF;AACpF,MAAMC,eAAe;AACrB,MAAMC,WAAWtB,QAAQ;AAEzB,eAAeuB,gBAAwBC,MAAc,EAAEC,GAAmB,EAAEC,SAAiB,EAAEC,OAA4B;IACzH,MAAMC,WAAWf,iBAAiBa;IAClC,MAAMG,WAAWC,KAAKC,GAAG,KAAKH;IAC9B,OAAS;QACP,IAAI;YACF,OAAO,MAAMD,QAAQK,OAAO,CAACR,QAAQC;QACvC,EAAE,OAAOQ,KAAK;gBACPN;YAAL,IAAI,GAACA,oBAAAA,QAAQO,QAAQ,cAAhBP,wCAAAA,uBAAAA,SAAmBM,OAAe,MAAMA;YAC7C,sFAAsF;YACtF,IAAIX,SAASa,cAAc,EAAEb,SAASc,OAAO,CAAC;gBAAEC,OAAOZ,IAAIY,KAAK;gBAAEb;YAAO;YACzE,IAAIM,KAAKC,GAAG,MAAMF,UAAU;gBAC1B,MAAM,IAAIvB,WAAW,cAAc,CAAC,2CAA2C,EAAEmB,IAAIY,KAAK,CAAC,QAAQ,EAAEb,OAAO,gCAAgC,EAAEc,KAAKC,KAAK,CAACX,WAAW,MAAM,qHAAqH,CAAC;YAClS;YACA,MAAM,IAAIY,QAAQ,CAACC,UAAYC,WAAWD,SAASpB;QACrD;IACF;AACF;AAEA,eAAesB,mBAA2BlB,GAAmB,EAAEE,OAA4B;QACvEF,gBAWkCA;IAXpD,MAAMC,aAAYD,iBAAAA,IAAIC,SAAS,cAAbD,4BAAAA,iBAAiBA,IAAImB,OAAO;IAC9C,MAAMC,WAAW3C,KAAKwB,WAAWtB;IACjCH,UAAU4C,UAAU;QAAEC,WAAW;IAAK;IACtC,MAAMtB,SAAStB,KAAK2C,UAAUlB,QAAQoB,QAAQ;IAE9C,MAAM,EAAEC,MAAM,EAAEC,IAAI,EAAE,GAAG,MAAM1B,gBAAgBC,QAAQC,KAAKC,WAAWC;IACvE,gGAAgG;IAChG,kGAAkG;IAClG,IAAIuB,SAAS;IACb,+FAA+F;IAC/F,2CAA2C;IAC3C,MAAMC,UAAU3C,cAAcyC,MAAMxB,KAAK;QAAE2B,OAAO,GAAE3B,eAAAA,IAAI2B,OAAO,cAAX3B,0BAAAA,eAAeA,IAAImB,OAAO;QAAElB;IAAU,GAAGC,QAAQ0B,gBAAgB;IACrH,IAAI;QACF,MAAMJ,KAAKK,IAAI,CAAC;QAEhB,uFAAuF;QACvF,qGAAqG;QACrG,MAAMC,UAAU,MAAMvC,QAAQiC,MAAM;QACpC,MAAMO,WAAW,MAAMxC,QAAQiC,MAAM;QACrC,MAAMQ,eAAetD,iBAAiBsB,KAAKlB;QAC3C,0FAA0F;QAC1F,yDAAyD;QACzD,IAAImD;QACJ,kGAAkG;QAClG,IAAIC;QACJ,4GAA4G;QAC5G,IAAIC;QACJ,IAAI,AAACL,YAAY,QAAQA,YAAY5B,QAAQkC,aAAa,IAAML,aAAa,QAAQA,aAAaC,cAAe;gBAejChC;YAd9E,uFAAuF;YACvF,wFAAwF;YACxF,IAAI8B,YAAY,QAAQA,YAAY5B,QAAQkC,aAAa,EAAE;gBACzDC,QAAQC,KAAK,CAAC;gBACd,kFAAkF;gBAClFb,SAAS;gBACT,MAAMvB,QAAQqC,KAAK,CAAChB;gBACpBvC,WAAWgB;gBACX,OAAOkB,mBAAmBlB,KAAKE;YACjC;YACA,MAAMsC,cAAc/C,qBAAqBsC,qBAAAA,sBAAAA,WAAY,IAAIC;YACzD,oFAAoF;YACpF,0FAA0F;YAC1F,2EAA2E;YAC3E,MAAMS,eAAenD,mBAAmBkD,eAAenD,kBAAkBW,MAAKA,gBAAAA,IAAI2B,OAAO,cAAX3B,2BAAAA,gBAAeA,IAAImB,OAAO,EAAEY,qBAAAA,sBAAAA,WAAY,IAAIS,eAAe;YACzI,yFAAyF;YACzF,oFAAoF;YACpF,MAAME,YAAYF,YAAYG,IAAI,KAAK,KAAKH,YAAYI,GAAG,CAAC,WAAW3D,oBAAoB8C,qBAAAA,sBAAAA,WAAY,IAAIC,gBAAgB;YAC3H,4FAA4F;YAC5F,6FAA6F;YAC7F,4DAA4D;YAC5D,MAAMa,UAAU1D,oBAAoBqD,eAAetD,uBAAuB6C,qBAAAA,sBAAAA,WAAY,IAAIC,cAAcQ,eAAe;YACvH,IAAIA,YAAYG,IAAI,KAAK,KAAKH,YAAYI,GAAG,CAAC,YAAYlD,qBAAqBqC,qBAAAA,sBAAAA,WAAY,IAAIC,eAAe;gBAC5G,iFAAiF;gBACjF,sDAAsD;gBACtDK,QAAQC,KAAK,CAAC;gBACd,MAAM9C,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIU,cAAc,MAAM;gBAC7B,yFAAyF;gBACzF,iEAAiE;gBACjER,kBAAkBQ;gBAClBL,QAAQC,KAAK,CAACI,cAAc,UAAU,6EAA6E;gBACnH,MAAMlD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIS,iBAAiB,MAAM;gBAChC,kFAAkF;gBAClF,uEAAuE;gBACvE,sFAAsF;gBACtF,qFAAqF;gBACrFR,cAAcQ;gBACd,MAAMK,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,oCAAoC,CAAC;gBACpF,MAAMtD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO,IAAIa,YAAY,MAAM;gBAC3B,uFAAuF;gBACvF,wFAAwF;gBACxF,kFAAkF;gBAClF,sDAAsD;gBACtDV,iBAAiBU;gBACjB,MAAMC,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,sCAAsC,CAAC;gBACtF,MAAMtD,QAAQgC,MAAM,YAAYQ;YAClC,OAAO;gBACL,MAAMc,UAAUnD,cAAcoC,qBAAAA,sBAAAA,WAAY,IAAIC;gBAC9CK,QAAQC,KAAK,CAAC,CAAC,sBAAsB,EAAEQ,QAAQ,oBAAoB,CAAC;gBACpErB,SAAS;gBACT,MAAMvB,QAAQqC,KAAK,CAAChB;gBACpBvC,WAAWgB;gBACX,OAAOkB,mBAAmBlB,KAAKE;YACjC;QACF;QAEA,MAAMA,QAAQ6C,YAAY,CAACxB,QAAQC,MAAMxB;QAEzC,wFAAwF;QACxF,uFAAuF;QACvF,IAAIE,QAAQ8C,qBAAqB,EAAE;gBACH;YAA9B,MAAMC,gBAAgBC,QAAQ,OAAA,MAAM3D,QAAQiC,MAAM,iCAApB,kBAAA,OAA4C;YAC1E,MAAMtB,QAAQ8C,qBAAqB,CAACzB,QAAQC,MAAMX,KAAKsC,GAAG,CAACtC,KAAKuC,GAAG,CAAC,OAAO,IAAIH,gBAAgB;QACjG;QAEA,6FAA6F;QAC7F,+FAA+F;QAC/F,wCAAwC;QACxC,MAAMI,cAAcnB,kBAAkB,AAAC,CAAA,MAAMR,QAAQ4B,UAAU,CAACpB,gBAAe,EAAGqB,MAAM,GAAG;QAC3F,MAAMC,gBAAgBrB,iBAAiB,AAAC,CAAA,MAAMT,QAAQ+B,kBAAkB,CAACtB,eAAc,EAAGoB,MAAM,GAAG;QAEnG,MAAM,EAAEA,MAAM,EAAEG,QAAQ,EAAEC,MAAM,EAAE,GAAG,MAAMjC,QAAQkC,KAAK,CAAC3B;QACzD,uIAAuI;QACvI,MAAMP,QAAQa,KAAK;QAEnB,OAAO;YAAEhB;YAAQC;YAAMxB;YAAKD;YAAQwD,QAAQA,SAASF,cAAcG;YAAeE;YAAUC;QAAO;IACrG,EAAE,OAAOnD,KAAK;QACZ,MAAMkB,QAAQa,KAAK;QACnB,IAAI,CAACd,QAAQ,MAAMvB,QAAQqC,KAAK,CAAChB;QACjC,MAAMf;IACR;AACF;AAEA,iGAAiG;AACjG,mFAAmF;AACnF,OAAO,eAAeqD,gBAAwB7D,GAAmB,EAAEE,OAA4B;IAC7F,wFAAwF;IACxF,wEAAwE;IACxE,IAAI,CAACF,IAAI2B,OAAO,EAAE;QAChB,IAAI,CAAC3B,IAAImB,OAAO,EAAE,MAAM,IAAItC,WAAW,kBAAkB;QACzDmB,IAAI2B,OAAO,GAAG3B,IAAImB,OAAO;IAC3B;IACA,IAAI,CAACnB,IAAIC,SAAS,EAAED,IAAIC,SAAS,GAAGD,IAAI2B,OAAO;IAC/C,MAAM,EAAEJ,MAAM,EAAEC,IAAI,EAAExB,KAAK8D,WAAW,EAAE/D,MAAM,EAAEwD,MAAM,EAAEG,QAAQ,EAAEC,MAAM,EAAE,GAAG,MAAMzC,mBAAmBlB,KAAKE;IAC3G,MAAMU,QAAQV,QAAQ6D,WAAW,CAACxC,QAAQC,MAAMsC;IAChD,0FAA0F;IAC1FlF,eAAe4C,MAAMZ;IACrB,OAAO;QAAEA;QAAOZ,KAAK8D;QAAa/D;QAAQwD;QAAQG;QAAUC;IAAO;AACrE;AAEA,OAAO,eAAeK,SAASpD,KAAY;IACzC,MAAMqD,OAAO,MAAMrD,MAAMsD,OAAO,CAAC;IACjC,OAAO,AAAE,CAAA,MAAMD,KAAKE,GAAG,EAAC,EAAqBC,CAAC;AAChD"}
@@ -3,7 +3,11 @@ import type { ParsePool } from '../scan/pool.js';
3
3
  import { type Stages } from './stages.js';
4
4
  import type { Connection, ReconcileDialect } from './types.js';
5
5
  export declare const CORE_FRONTMATTER_COLUMNS: Set<string>;
6
- export declare function reconcile(conn: Connection, cfg: Config, baseDir: string, dialect: ReconcileDialect, pool?: ParsePool, forcedPaths?: ReadonlySet<string>): Promise<{
6
+ export interface ReconcilePaths {
7
+ rootDir: string;
8
+ configDir: string;
9
+ }
10
+ export declare function reconcile(conn: Connection, cfg: Config, paths: ReconcilePaths, dialect: ReconcileDialect, pool?: ParsePool, forcedPaths?: ReadonlySet<string>): Promise<{
7
11
  parsed: number;
8
12
  warnings: string[];
9
13
  stages: Stages;
@@ -20,12 +20,13 @@ export const CORE_FRONTMATTER_COLUMNS = new Set([
20
20
  '_size',
21
21
  '_parse_error'
22
22
  ]);
23
- export async function reconcile(conn, cfg, baseDir, dialect, pool, forcedPaths) {
23
+ export async function reconcile(conn, cfg, paths, dialect, pool, forcedPaths) {
24
+ const { rootDir, configDir } = paths;
24
25
  const start = process.hrtime.bigint();
25
26
  const features = activeFeatures(cfg);
26
27
  const stages = stageRecorder(features.map((f)=>f.name));
27
28
  const elapsed = ()=>Number(process.hrtime.bigint() - start) / 1e6;
28
- const files = await stages.time('list', ()=>listFiles(cfg, baseDir));
29
+ const files = await stages.time('list', ()=>listFiles(cfg, rootDir));
29
30
  const currentSet = new Set(files.map((f)=>f.relPath));
30
31
  const existingRows = await stages.time('existing', async ()=>{
31
32
  const existingStmt = await conn.prepare('SELECT "path", "_mtime", "_size" FROM frontmatter');
@@ -180,7 +181,7 @@ export async function reconcile(conn, cfg, baseDir, dialect, pool, forcedPaths)
180
181
  }, dialect.beginMode());
181
182
  const durationMs = Date.now() - txStart;
182
183
  // Every store, not just the ones with a PRAGMA to derive: connectUnlocked's lock-wait budget needs it too.
183
- recordLockWaitMs(baseDir, durationMs);
184
+ recordLockWaitMs(configDir, durationMs);
184
185
  if (dialect.recordDuration) await stages.time('meta', ()=>{
185
186
  var _dialect_recordDuration;
186
187
  return (_dialect_recordDuration = dialect.recordDuration) === null || _dialect_recordDuration === void 0 ? void 0 : _dialect_recordDuration.call(dialect, conn, durationMs);
@@ -1 +1 @@
1
- {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/reconcile.ts"],"sourcesContent":["import type { Config } from '../config/index.ts';\nimport { activeFeatures } from '../features/index.ts';\nimport type { ExtractedDoc, ReconcileDelta } from '../features/types.ts';\nimport { progress } from '../output/progress.ts';\nimport { listFiles, RESERVED_COLUMNS } from '../scan/index.ts';\nimport type { ParsePool } from '../scan/pool.ts';\nimport { reparseFiles } from '../scan/reparse.ts';\nimport { recordLockWaitMs } from './lock-wait.ts';\nimport { appendRows, getColumns, quoteIdent } from './shared.ts';\nimport { featureStage, type Stages, stageRecorder } from './stages.ts';\nimport { withTransaction } from './transaction.ts';\nimport type { Connection, ReconcileDialect } from './types.ts';\n\n// One reconcile algorithm shared by every store, parameterised by a per-engine ReconcileDialect\n// (types.ts). Ordering is universal, not a dialect concern: ALTER, then the frontmatter upsert,\n// then reconcileContent, then preset_files, then the vanished-frontmatter delete, then feature hooks\n// -- a vanished path's content delete (inside reconcileContent) must precede its frontmatter\n// delete, since sqlite's delete SQL resolves the row via its frontmatter rowid.\n\n// Feature-owned columns (`_rank`) must stay out of the upsert: a reparse would null the last\n// computed value on every touch, not just the reconciles that recompute it.\nexport const CORE_FRONTMATTER_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_parse_error']);\n\nexport async function reconcile(conn: Connection, cfg: Config, baseDir: string, dialect: ReconcileDialect, pool?: ParsePool, forcedPaths?: ReadonlySet<string>): Promise<{ parsed: number; warnings: string[]; stages: Stages }> {\n const start = process.hrtime.bigint();\n const features = activeFeatures(cfg);\n const stages = stageRecorder(features.map((f) => f.name));\n const elapsed = () => Number(process.hrtime.bigint() - start) / 1e6;\n const files = await stages.time('list', () => listFiles(cfg, baseDir));\n const currentSet = new Set(files.map((f) => f.relPath));\n\n const existingRows = await stages.time('existing', async () => {\n const existingStmt = await conn.prepare('SELECT \"path\", \"_mtime\", \"_size\" FROM frontmatter');\n return (await existingStmt.all()) as Array<{ path: string; _mtime: number; _size: number }>;\n });\n const existing = new Map(existingRows.map((r) => [r.path, r]));\n // A path whose coverage moved between presets (forcedPaths) but is no longer covered at all is\n // already caught below by !currentSet.has, since it can only be forced by having existed under\n // an old preset's match, which means it was reconciled into `existing` already.\n const vanished = existingRows.filter((r) => !currentSet.has(r.path)).map((r) => r.path);\n\n // forcedPaths treats an unchanged file as touched because its preset coverage moved, not its\n // stamp -- reconcile still owns add/update/remove and every cross-feature cascade for it.\n const toReparse = files.filter((f) => {\n const row = existing.get(f.relPath);\n return !row || row._mtime !== f.mtimeMs || row._size !== f.size || (forcedPaths?.has(f.relPath) ?? false);\n });\n\n if (vanished.length === 0 && toReparse.length === 0) return { parsed: 0, warnings: [], stages: stages.take(elapsed(), 0) };\n\n const seenColumns = await getColumns(conn);\n\n // Bulk reparses (a sync, a cold build) are the long silences a query can hit; short\n // reconciles stay silent (progress() has a threshold).\n const report = progress('reparsing files', toReparse.length);\n // Pool wall time, dispatch to drain, so this stage shares a clock with every other one.\n const { docs: parsedDocs, warnings, newColumns, workerParseMs } = await stages.time('parse', () => reparseFiles(toReparse, features, cfg, seenColumns, report.tick, { pool }));\n report.finish();\n for (const col of newColumns) seenColumns.add(col);\n\n const allColumns = [...seenColumns];\n // Fence before ALTERing: a store's own failure past this point is a raw, engine-specific\n // error with no indication of the boundary or the levers -- dialect.checkColumnLimit names both.\n dialect.checkColumnLimit(allColumns.length);\n // Columns the frontmatter upsert actually writes: core + parsed frontmatter keys, never a\n // feature-owned reserved column (see CORE_FRONTMATTER_COLUMNS above).\n const writableColumns = allColumns.filter((c) => CORE_FRONTMATTER_COLUMNS.has(c) || !RESERVED_COLUMNS.has(c));\n // ON CONFLICT DO UPDATE (not OR REPLACE) keeps the row's rowid stable across reparses --\n // sqlite's content rows are coupled to that rowid.\n const insertSql = `INSERT INTO frontmatter (${writableColumns.map(quoteIdent).join(', ')}) VALUES (${writableColumns.map(() => '?').join(', ')}) ON CONFLICT(\"path\") DO UPDATE SET ${writableColumns\n .filter((c) => c !== 'path')\n .map((c) => `${quoteIdent(c)} = excluded.${quoteIdent(c)}`)\n .join(', ')}`;\n\n const added = toReparse.filter((f) => !existing.has(f.relPath)).map((f) => f.relPath);\n const delta: ReconcileDelta = { files, reparsed: parsedDocs.map((d) => d.relPath), added, vanished };\n const addedSet = new Set(added);\n const reparsedExisting = parsedDocs.map((d) => d.relPath).filter((p) => !addedSet.has(p));\n // Paths whose content (and, per feature, other rows) need clearing: gone entirely, or about to\n // be reinserted fresh. Disjoint from `added`, which has nothing to clear.\n const touched = [...vanished, ...reparsedExisting];\n\n const txStart = Date.now();\n await withTransaction(\n conn,\n async () => {\n // Re-read inside the write transaction: newColumns came from a read taken before it opened, so a\n // concurrent reconcile may have added some of them since. ALTER has no IF NOT EXISTS.\n const present = await getColumns(conn);\n const missingColumns = newColumns.filter((col) => !present.has(col));\n if (missingColumns.length > 0) await stages.time('alter', () => dialect.addColumns(conn, missingColumns));\n\n // Revalidated once the lock is held, before this process's own frontmatter upsert below:\n // `added` came from a path read taken before this transaction's lock, so a path still\n // called \"added\" here may already have a content row from a concurrent reconcile that\n // committed while this one waited. One SELECT, not a DELETE per added file -- with no\n // contention it returns the same set `existing` already ruled out, so nothing extra clears.\n let contentTouched = touched;\n if (added.length > 0)\n contentTouched = await stages.time('added-recheck', async () => {\n const currentPathsStmt = await conn.prepare('SELECT \"path\" FROM frontmatter');\n const currentPaths = new Set(((await currentPathsStmt.all()) as Array<{ path: string }>).map((r) => r.path));\n const staleAdded = added.filter((p) => currentPaths.has(p));\n return staleAdded.length > 0 ? [...touched, ...staleAdded] : touched;\n });\n\n if (parsedDocs.length > 0) {\n const toRow = (doc: (typeof parsedDocs)[number]) =>\n writableColumns.map((col) => {\n if (col === 'path') return doc.relPath;\n if (col === '_mtime') return doc.mtimeMs;\n if (col === '_ctime') return doc.ctimeMs;\n if (col === '_size') return doc.size;\n // Written per parse, unlike _rank, which a feature pass owns and the upsert skips.\n if (col === '_parse_error') return doc.parseError;\n return doc.data[col] ?? null;\n });\n // A path in `added` has no existing frontmatter row, so it can never conflict; the rest\n // genuinely can, and keep the upsert.\n await stages.time('fm-upsert', async () => {\n const newDocs = parsedDocs.filter((d) => addedSet.has(d.relPath));\n const updateDocs = parsedDocs.filter((d) => !addedSet.has(d.relPath));\n await appendRows(conn, 'frontmatter', writableColumns, insertSql, newDocs.map(toRow));\n if (updateDocs.length > 0) await conn.runBatch(insertSql, updateDocs.map(toRow));\n });\n }\n\n // After the upsert, so every doc already has the frontmatter row sqlite's content rowid\n // couples to. ON CONFLICT DO UPDATE preserves that rowid, so a reparse keeps its identity.\n await stages.time('text-index', () => dialect.reconcileContent(conn, contentTouched, parsedDocs, delta, cfg));\n\n // A preset edit forces a full rebuild, so an unchanged doc's coverage is already correct;\n // new docs have nothing to clear, which keeps cold builds linear.\n await stages.time('presets', async () => {\n if (touched.length > 0)\n await conn.runBatch(\n 'DELETE FROM preset_files WHERE \"path\" = ?',\n touched.map((p) => [p])\n );\n const presetRows: unknown[][] = [];\n for (const doc of parsedDocs) for (const presetName of doc.presets) presetRows.push([doc.relPath, presetName]);\n // DO NOTHING, not a bare INSERT: the added/touched split above comes from a read taken\n // before this transaction's lock, so a path this process calls \"added\" can already have\n // its (path, preset) row committed by a concurrent reconcile -- the row would be identical either way.\n await appendRows(conn, 'preset_files', ['path', 'preset'], 'INSERT INTO preset_files (\"path\", preset) VALUES (?, ?) ON CONFLICT(\"path\", preset) DO NOTHING', presetRows);\n });\n\n // Before the feature hooks, never after: rank's afterReconcile reads frontmatter as PageRank's\n // node set, so a lingering vanished row would dilute rank mass across every surviving note.\n if (vanished.length > 0)\n await stages.time('vanished', () =>\n conn.runBatch(\n 'DELETE FROM frontmatter WHERE \"path\" = ?',\n vanished.map((p) => [p])\n )\n );\n\n // Timed per feature per hook, so link resolution and PageRank are named stages without\n // links.ts or rank.ts knowing anything about this, and a new feature is visible for free.\n if (touched.length > 0) for (const feature of features) await stages.time(featureStage(feature.name, 'remove'), () => feature.remove?.(conn, touched, delta));\n for (const feature of features) {\n const docsForFeature: ExtractedDoc[] = parsedDocs.map((doc) => ({ path: doc.relPath, extracted: doc.extracted[feature.name] }));\n await stages.time(featureStage(feature.name, 'store'), () => feature.store?.(conn, docsForFeature, delta));\n }\n for (const feature of features) await stages.time(featureStage(feature.name, 'after'), () => feature.afterReconcile?.(conn, delta));\n },\n dialect.beginMode()\n );\n\n const durationMs = Date.now() - txStart;\n // Every store, not just the ones with a PRAGMA to derive: connectUnlocked's lock-wait budget needs it too.\n recordLockWaitMs(baseDir, durationMs);\n if (dialect.recordDuration) await stages.time('meta', () => dialect.recordDuration?.(conn, durationMs));\n\n return { parsed: parsedDocs.length, warnings, stages: stages.take(elapsed(), durationMs, workerParseMs) };\n}\n"],"names":["activeFeatures","progress","listFiles","RESERVED_COLUMNS","reparseFiles","recordLockWaitMs","appendRows","getColumns","quoteIdent","featureStage","stageRecorder","withTransaction","CORE_FRONTMATTER_COLUMNS","Set","reconcile","conn","cfg","baseDir","dialect","pool","forcedPaths","start","process","hrtime","bigint","features","stages","map","f","name","elapsed","Number","files","time","currentSet","relPath","existingRows","existingStmt","prepare","all","existing","Map","r","path","vanished","filter","has","toReparse","row","get","_mtime","mtimeMs","_size","size","length","parsed","warnings","take","seenColumns","report","docs","parsedDocs","newColumns","workerParseMs","tick","finish","col","add","allColumns","checkColumnLimit","writableColumns","c","insertSql","join","added","delta","reparsed","d","addedSet","reparsedExisting","p","touched","txStart","Date","now","present","missingColumns","addColumns","contentTouched","currentPathsStmt","currentPaths","staleAdded","toRow","doc","ctimeMs","parseError","data","newDocs","updateDocs","runBatch","reconcileContent","presetRows","presetName","presets","push","feature","remove","docsForFeature","extracted","store","afterReconcile","beginMode","durationMs","recordDuration"],"mappings":"AACA,SAASA,cAAc,QAAQ,uBAAuB;AAEtD,SAASC,QAAQ,QAAQ,wBAAwB;AACjD,SAASC,SAAS,EAAEC,gBAAgB,QAAQ,mBAAmB;AAE/D,SAASC,YAAY,QAAQ,qBAAqB;AAClD,SAASC,gBAAgB,QAAQ,iBAAiB;AAClD,SAASC,UAAU,EAAEC,UAAU,EAAEC,UAAU,QAAQ,cAAc;AACjE,SAASC,YAAY,EAAeC,aAAa,QAAQ,cAAc;AACvE,SAASC,eAAe,QAAQ,mBAAmB;AAGnD,gGAAgG;AAChG,gGAAgG;AAChG,qGAAqG;AACrG,6FAA6F;AAC7F,gFAAgF;AAEhF,6FAA6F;AAC7F,4EAA4E;AAC5E,OAAO,MAAMC,2BAA2B,IAAIC,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;CAAe,EAAE;AAEvG,OAAO,eAAeC,UAAUC,IAAgB,EAAEC,GAAW,EAAEC,OAAe,EAAEC,OAAyB,EAAEC,IAAgB,EAAEC,WAAiC;IAC5J,MAAMC,QAAQC,QAAQC,MAAM,CAACC,MAAM;IACnC,MAAMC,WAAWzB,eAAegB;IAChC,MAAMU,SAAShB,cAAce,SAASE,GAAG,CAAC,CAACC,IAAMA,EAAEC,IAAI;IACvD,MAAMC,UAAU,IAAMC,OAAOT,QAAQC,MAAM,CAACC,MAAM,KAAKH,SAAS;IAChE,MAAMW,QAAQ,MAAMN,OAAOO,IAAI,CAAC,QAAQ,IAAM/B,UAAUc,KAAKC;IAC7D,MAAMiB,aAAa,IAAIrB,IAAImB,MAAML,GAAG,CAAC,CAACC,IAAMA,EAAEO,OAAO;IAErD,MAAMC,eAAe,MAAMV,OAAOO,IAAI,CAAC,YAAY;QACjD,MAAMI,eAAe,MAAMtB,KAAKuB,OAAO,CAAC;QACxC,OAAQ,MAAMD,aAAaE,GAAG;IAChC;IACA,MAAMC,WAAW,IAAIC,IAAIL,aAAaT,GAAG,CAAC,CAACe,IAAM;YAACA,EAAEC,IAAI;YAAED;SAAE;IAC5D,+FAA+F;IAC/F,+FAA+F;IAC/F,gFAAgF;IAChF,MAAME,WAAWR,aAAaS,MAAM,CAAC,CAACH,IAAM,CAACR,WAAWY,GAAG,CAACJ,EAAEC,IAAI,GAAGhB,GAAG,CAAC,CAACe,IAAMA,EAAEC,IAAI;IAEtF,6FAA6F;IAC7F,0FAA0F;IAC1F,MAAMI,YAAYf,MAAMa,MAAM,CAAC,CAACjB;;QAC9B,MAAMoB,MAAMR,SAASS,GAAG,CAACrB,EAAEO,OAAO;QAClC,OAAO,CAACa,OAAOA,IAAIE,MAAM,KAAKtB,EAAEuB,OAAO,IAAIH,IAAII,KAAK,KAAKxB,EAAEyB,IAAI,aAAKjC,wBAAAA,kCAAAA,YAAa0B,GAAG,CAAClB,EAAEO,OAAO,wCAAK;IACrG;IAEA,IAAIS,SAASU,MAAM,KAAK,KAAKP,UAAUO,MAAM,KAAK,GAAG,OAAO;QAAEC,QAAQ;QAAGC,UAAU,EAAE;QAAE9B,QAAQA,OAAO+B,IAAI,CAAC3B,WAAW;IAAG;IAEzH,MAAM4B,cAAc,MAAMnD,WAAWQ;IAErC,oFAAoF;IACpF,uDAAuD;IACvD,MAAM4C,SAAS1D,SAAS,mBAAmB8C,UAAUO,MAAM;IAC3D,wFAAwF;IACxF,MAAM,EAAEM,MAAMC,UAAU,EAAEL,QAAQ,EAAEM,UAAU,EAAEC,aAAa,EAAE,GAAG,MAAMrC,OAAOO,IAAI,CAAC,SAAS,IAAM7B,aAAa2C,WAAWtB,UAAUT,KAAK0C,aAAaC,OAAOK,IAAI,EAAE;YAAE7C;QAAK;IAC3KwC,OAAOM,MAAM;IACb,KAAK,MAAMC,OAAOJ,WAAYJ,YAAYS,GAAG,CAACD;IAE9C,MAAME,aAAa;WAAIV;KAAY;IACnC,yFAAyF;IACzF,iGAAiG;IACjGxC,QAAQmD,gBAAgB,CAACD,WAAWd,MAAM;IAC1C,0FAA0F;IAC1F,sEAAsE;IACtE,MAAMgB,kBAAkBF,WAAWvB,MAAM,CAAC,CAAC0B,IAAM3D,yBAAyBkC,GAAG,CAACyB,MAAM,CAACpE,iBAAiB2C,GAAG,CAACyB;IAC1G,yFAAyF;IACzF,mDAAmD;IACnD,MAAMC,YAAY,CAAC,yBAAyB,EAAEF,gBAAgB3C,GAAG,CAACnB,YAAYiE,IAAI,CAAC,MAAM,UAAU,EAAEH,gBAAgB3C,GAAG,CAAC,IAAM,KAAK8C,IAAI,CAAC,MAAM,oCAAoC,EAAEH,gBAClLzB,MAAM,CAAC,CAAC0B,IAAMA,MAAM,QACpB5C,GAAG,CAAC,CAAC4C,IAAM,GAAG/D,WAAW+D,GAAG,YAAY,EAAE/D,WAAW+D,IAAI,EACzDE,IAAI,CAAC,OAAO;IAEf,MAAMC,QAAQ3B,UAAUF,MAAM,CAAC,CAACjB,IAAM,CAACY,SAASM,GAAG,CAAClB,EAAEO,OAAO,GAAGR,GAAG,CAAC,CAACC,IAAMA,EAAEO,OAAO;IACpF,MAAMwC,QAAwB;QAAE3C;QAAO4C,UAAUf,WAAWlC,GAAG,CAAC,CAACkD,IAAMA,EAAE1C,OAAO;QAAGuC;QAAO9B;IAAS;IACnG,MAAMkC,WAAW,IAAIjE,IAAI6D;IACzB,MAAMK,mBAAmBlB,WAAWlC,GAAG,CAAC,CAACkD,IAAMA,EAAE1C,OAAO,EAAEU,MAAM,CAAC,CAACmC,IAAM,CAACF,SAAShC,GAAG,CAACkC;IACtF,+FAA+F;IAC/F,0EAA0E;IAC1E,MAAMC,UAAU;WAAIrC;WAAamC;KAAiB;IAElD,MAAMG,UAAUC,KAAKC,GAAG;IACxB,MAAMzE,gBACJI,MACA;QACE,iGAAiG;QACjG,sFAAsF;QACtF,MAAMsE,UAAU,MAAM9E,WAAWQ;QACjC,MAAMuE,iBAAiBxB,WAAWjB,MAAM,CAAC,CAACqB,MAAQ,CAACmB,QAAQvC,GAAG,CAACoB;QAC/D,IAAIoB,eAAehC,MAAM,GAAG,GAAG,MAAM5B,OAAOO,IAAI,CAAC,SAAS,IAAMf,QAAQqE,UAAU,CAACxE,MAAMuE;QAEzF,yFAAyF;QACzF,sFAAsF;QACtF,sFAAsF;QACtF,sFAAsF;QACtF,4FAA4F;QAC5F,IAAIE,iBAAiBP;QACrB,IAAIP,MAAMpB,MAAM,GAAG,GACjBkC,iBAAiB,MAAM9D,OAAOO,IAAI,CAAC,iBAAiB;YAClD,MAAMwD,mBAAmB,MAAM1E,KAAKuB,OAAO,CAAC;YAC5C,MAAMoD,eAAe,IAAI7E,IAAI,AAAE,CAAA,MAAM4E,iBAAiBlD,GAAG,EAAC,EAA+BZ,GAAG,CAAC,CAACe,IAAMA,EAAEC,IAAI;YAC1G,MAAMgD,aAAajB,MAAM7B,MAAM,CAAC,CAACmC,IAAMU,aAAa5C,GAAG,CAACkC;YACxD,OAAOW,WAAWrC,MAAM,GAAG,IAAI;mBAAI2B;mBAAYU;aAAW,GAAGV;QAC/D;QAEF,IAAIpB,WAAWP,MAAM,GAAG,GAAG;YACzB,MAAMsC,QAAQ,CAACC,MACbvB,gBAAgB3C,GAAG,CAAC,CAACuC;wBAOZ2B;oBANP,IAAI3B,QAAQ,QAAQ,OAAO2B,IAAI1D,OAAO;oBACtC,IAAI+B,QAAQ,UAAU,OAAO2B,IAAI1C,OAAO;oBACxC,IAAIe,QAAQ,UAAU,OAAO2B,IAAIC,OAAO;oBACxC,IAAI5B,QAAQ,SAAS,OAAO2B,IAAIxC,IAAI;oBACpC,mFAAmF;oBACnF,IAAIa,QAAQ,gBAAgB,OAAO2B,IAAIE,UAAU;oBACjD,QAAOF,gBAAAA,IAAIG,IAAI,CAAC9B,IAAI,cAAb2B,2BAAAA,gBAAiB;gBAC1B;YACF,wFAAwF;YACxF,sCAAsC;YACtC,MAAMnE,OAAOO,IAAI,CAAC,aAAa;gBAC7B,MAAMgE,UAAUpC,WAAWhB,MAAM,CAAC,CAACgC,IAAMC,SAAShC,GAAG,CAAC+B,EAAE1C,OAAO;gBAC/D,MAAM+D,aAAarC,WAAWhB,MAAM,CAAC,CAACgC,IAAM,CAACC,SAAShC,GAAG,CAAC+B,EAAE1C,OAAO;gBACnE,MAAM7B,WAAWS,MAAM,eAAeuD,iBAAiBE,WAAWyB,QAAQtE,GAAG,CAACiE;gBAC9E,IAAIM,WAAW5C,MAAM,GAAG,GAAG,MAAMvC,KAAKoF,QAAQ,CAAC3B,WAAW0B,WAAWvE,GAAG,CAACiE;YAC3E;QACF;QAEA,wFAAwF;QACxF,2FAA2F;QAC3F,MAAMlE,OAAOO,IAAI,CAAC,cAAc,IAAMf,QAAQkF,gBAAgB,CAACrF,MAAMyE,gBAAgB3B,YAAYc,OAAO3D;QAExG,0FAA0F;QAC1F,kEAAkE;QAClE,MAAMU,OAAOO,IAAI,CAAC,WAAW;YAC3B,IAAIgD,QAAQ3B,MAAM,GAAG,GACnB,MAAMvC,KAAKoF,QAAQ,CACjB,6CACAlB,QAAQtD,GAAG,CAAC,CAACqD,IAAM;oBAACA;iBAAE;YAE1B,MAAMqB,aAA0B,EAAE;YAClC,KAAK,MAAMR,OAAOhC,WAAY,KAAK,MAAMyC,cAAcT,IAAIU,OAAO,CAAEF,WAAWG,IAAI,CAAC;gBAACX,IAAI1D,OAAO;gBAAEmE;aAAW;YAC7G,uFAAuF;YACvF,wFAAwF;YACxF,uGAAuG;YACvG,MAAMhG,WAAWS,MAAM,gBAAgB;gBAAC;gBAAQ;aAAS,EAAE,kGAAkGsF;QAC/J;QAEA,+FAA+F;QAC/F,4FAA4F;QAC5F,IAAIzD,SAASU,MAAM,GAAG,GACpB,MAAM5B,OAAOO,IAAI,CAAC,YAAY,IAC5BlB,KAAKoF,QAAQ,CACX,4CACAvD,SAASjB,GAAG,CAAC,CAACqD,IAAM;oBAACA;iBAAE;QAI7B,uFAAuF;QACvF,0FAA0F;QAC1F,IAAIC,QAAQ3B,MAAM,GAAG,GAAG,KAAK,MAAMmD,WAAWhF,SAAU,MAAMC,OAAOO,IAAI,CAACxB,aAAagG,QAAQ5E,IAAI,EAAE,WAAW;gBAAM4E;oBAAAA,kBAAAA,QAAQC,MAAM,cAAdD,sCAAAA,qBAAAA,SAAiB1F,MAAMkE,SAASN;;QACtJ,KAAK,MAAM8B,WAAWhF,SAAU;YAC9B,MAAMkF,iBAAiC9C,WAAWlC,GAAG,CAAC,CAACkE,MAAS,CAAA;oBAAElD,MAAMkD,IAAI1D,OAAO;oBAAEyE,WAAWf,IAAIe,SAAS,CAACH,QAAQ5E,IAAI,CAAC;gBAAC,CAAA;YAC5H,MAAMH,OAAOO,IAAI,CAACxB,aAAagG,QAAQ5E,IAAI,EAAE,UAAU;oBAAM4E;wBAAAA,iBAAAA,QAAQI,KAAK,cAAbJ,qCAAAA,oBAAAA,SAAgB1F,MAAM4F,gBAAgBhC;;QACrG;QACA,KAAK,MAAM8B,WAAWhF,SAAU,MAAMC,OAAOO,IAAI,CAACxB,aAAagG,QAAQ5E,IAAI,EAAE,UAAU;gBAAM4E;oBAAAA,0BAAAA,QAAQK,cAAc,cAAtBL,8CAAAA,6BAAAA,SAAyB1F,MAAM4D;;IAC9H,GACAzD,QAAQ6F,SAAS;IAGnB,MAAMC,aAAa7B,KAAKC,GAAG,KAAKF;IAChC,2GAA2G;IAC3G7E,iBAAiBY,SAAS+F;IAC1B,IAAI9F,QAAQ+F,cAAc,EAAE,MAAMvF,OAAOO,IAAI,CAAC,QAAQ;YAAMf;gBAAAA,0BAAAA,QAAQ+F,cAAc,cAAtB/F,8CAAAA,6BAAAA,SAAyBH,MAAMiG;;IAE3F,OAAO;QAAEzD,QAAQM,WAAWP,MAAM;QAAEE;QAAU9B,QAAQA,OAAO+B,IAAI,CAAC3B,WAAWkF,YAAYjD;IAAe;AAC1G"}
1
+ {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/reconcile.ts"],"sourcesContent":["import type { Config } from '../config/index.ts';\nimport { activeFeatures } from '../features/index.ts';\nimport type { ExtractedDoc, ReconcileDelta } from '../features/types.ts';\nimport { progress } from '../output/progress.ts';\nimport { listFiles, RESERVED_COLUMNS } from '../scan/index.ts';\nimport type { ParsePool } from '../scan/pool.ts';\nimport { reparseFiles } from '../scan/reparse.ts';\nimport { recordLockWaitMs } from './lock-wait.ts';\nimport { appendRows, getColumns, quoteIdent } from './shared.ts';\nimport { featureStage, type Stages, stageRecorder } from './stages.ts';\nimport { withTransaction } from './transaction.ts';\nimport type { Connection, ReconcileDialect } from './types.ts';\n\n// One reconcile algorithm shared by every store, parameterised by a per-engine ReconcileDialect\n// (types.ts). Ordering is universal, not a dialect concern: ALTER, then the frontmatter upsert,\n// then reconcileContent, then preset_files, then the vanished-frontmatter delete, then feature hooks\n// -- a vanished path's content delete (inside reconcileContent) must precede its frontmatter\n// delete, since sqlite's delete SQL resolves the row via its frontmatter rowid.\n\n// Feature-owned columns (`_rank`) must stay out of the upsert: a reparse would null the last\n// computed value on every touch, not just the reconciles that recompute it.\nexport const CORE_FRONTMATTER_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_parse_error']);\n\nexport interface ReconcilePaths {\n rootDir: string;\n configDir: string;\n}\n\nexport async function reconcile(conn: Connection, cfg: Config, paths: ReconcilePaths, dialect: ReconcileDialect, pool?: ParsePool, forcedPaths?: ReadonlySet<string>): Promise<{ parsed: number; warnings: string[]; stages: Stages }> {\n const { rootDir, configDir } = paths;\n const start = process.hrtime.bigint();\n const features = activeFeatures(cfg);\n const stages = stageRecorder(features.map((f) => f.name));\n const elapsed = () => Number(process.hrtime.bigint() - start) / 1e6;\n const files = await stages.time('list', () => listFiles(cfg, rootDir));\n const currentSet = new Set(files.map((f) => f.relPath));\n\n const existingRows = await stages.time('existing', async () => {\n const existingStmt = await conn.prepare('SELECT \"path\", \"_mtime\", \"_size\" FROM frontmatter');\n return (await existingStmt.all()) as Array<{ path: string; _mtime: number; _size: number }>;\n });\n const existing = new Map(existingRows.map((r) => [r.path, r]));\n // A path whose coverage moved between presets (forcedPaths) but is no longer covered at all is\n // already caught below by !currentSet.has, since it can only be forced by having existed under\n // an old preset's match, which means it was reconciled into `existing` already.\n const vanished = existingRows.filter((r) => !currentSet.has(r.path)).map((r) => r.path);\n\n // forcedPaths treats an unchanged file as touched because its preset coverage moved, not its\n // stamp -- reconcile still owns add/update/remove and every cross-feature cascade for it.\n const toReparse = files.filter((f) => {\n const row = existing.get(f.relPath);\n return !row || row._mtime !== f.mtimeMs || row._size !== f.size || (forcedPaths?.has(f.relPath) ?? false);\n });\n\n if (vanished.length === 0 && toReparse.length === 0) return { parsed: 0, warnings: [], stages: stages.take(elapsed(), 0) };\n\n const seenColumns = await getColumns(conn);\n\n // Bulk reparses (a sync, a cold build) are the long silences a query can hit; short\n // reconciles stay silent (progress() has a threshold).\n const report = progress('reparsing files', toReparse.length);\n // Pool wall time, dispatch to drain, so this stage shares a clock with every other one.\n const { docs: parsedDocs, warnings, newColumns, workerParseMs } = await stages.time('parse', () => reparseFiles(toReparse, features, cfg, seenColumns, report.tick, { pool }));\n report.finish();\n for (const col of newColumns) seenColumns.add(col);\n\n const allColumns = [...seenColumns];\n // Fence before ALTERing: a store's own failure past this point is a raw, engine-specific\n // error with no indication of the boundary or the levers -- dialect.checkColumnLimit names both.\n dialect.checkColumnLimit(allColumns.length);\n // Columns the frontmatter upsert actually writes: core + parsed frontmatter keys, never a\n // feature-owned reserved column (see CORE_FRONTMATTER_COLUMNS above).\n const writableColumns = allColumns.filter((c) => CORE_FRONTMATTER_COLUMNS.has(c) || !RESERVED_COLUMNS.has(c));\n // ON CONFLICT DO UPDATE (not OR REPLACE) keeps the row's rowid stable across reparses --\n // sqlite's content rows are coupled to that rowid.\n const insertSql = `INSERT INTO frontmatter (${writableColumns.map(quoteIdent).join(', ')}) VALUES (${writableColumns.map(() => '?').join(', ')}) ON CONFLICT(\"path\") DO UPDATE SET ${writableColumns\n .filter((c) => c !== 'path')\n .map((c) => `${quoteIdent(c)} = excluded.${quoteIdent(c)}`)\n .join(', ')}`;\n\n const added = toReparse.filter((f) => !existing.has(f.relPath)).map((f) => f.relPath);\n const delta: ReconcileDelta = { files, reparsed: parsedDocs.map((d) => d.relPath), added, vanished };\n const addedSet = new Set(added);\n const reparsedExisting = parsedDocs.map((d) => d.relPath).filter((p) => !addedSet.has(p));\n // Paths whose content (and, per feature, other rows) need clearing: gone entirely, or about to\n // be reinserted fresh. Disjoint from `added`, which has nothing to clear.\n const touched = [...vanished, ...reparsedExisting];\n\n const txStart = Date.now();\n await withTransaction(\n conn,\n async () => {\n // Re-read inside the write transaction: newColumns came from a read taken before it opened, so a\n // concurrent reconcile may have added some of them since. ALTER has no IF NOT EXISTS.\n const present = await getColumns(conn);\n const missingColumns = newColumns.filter((col) => !present.has(col));\n if (missingColumns.length > 0) await stages.time('alter', () => dialect.addColumns(conn, missingColumns));\n\n // Revalidated once the lock is held, before this process's own frontmatter upsert below:\n // `added` came from a path read taken before this transaction's lock, so a path still\n // called \"added\" here may already have a content row from a concurrent reconcile that\n // committed while this one waited. One SELECT, not a DELETE per added file -- with no\n // contention it returns the same set `existing` already ruled out, so nothing extra clears.\n let contentTouched = touched;\n if (added.length > 0)\n contentTouched = await stages.time('added-recheck', async () => {\n const currentPathsStmt = await conn.prepare('SELECT \"path\" FROM frontmatter');\n const currentPaths = new Set(((await currentPathsStmt.all()) as Array<{ path: string }>).map((r) => r.path));\n const staleAdded = added.filter((p) => currentPaths.has(p));\n return staleAdded.length > 0 ? [...touched, ...staleAdded] : touched;\n });\n\n if (parsedDocs.length > 0) {\n const toRow = (doc: (typeof parsedDocs)[number]) =>\n writableColumns.map((col) => {\n if (col === 'path') return doc.relPath;\n if (col === '_mtime') return doc.mtimeMs;\n if (col === '_ctime') return doc.ctimeMs;\n if (col === '_size') return doc.size;\n // Written per parse, unlike _rank, which a feature pass owns and the upsert skips.\n if (col === '_parse_error') return doc.parseError;\n return doc.data[col] ?? null;\n });\n // A path in `added` has no existing frontmatter row, so it can never conflict; the rest\n // genuinely can, and keep the upsert.\n await stages.time('fm-upsert', async () => {\n const newDocs = parsedDocs.filter((d) => addedSet.has(d.relPath));\n const updateDocs = parsedDocs.filter((d) => !addedSet.has(d.relPath));\n await appendRows(conn, 'frontmatter', writableColumns, insertSql, newDocs.map(toRow));\n if (updateDocs.length > 0) await conn.runBatch(insertSql, updateDocs.map(toRow));\n });\n }\n\n // After the upsert, so every doc already has the frontmatter row sqlite's content rowid\n // couples to. ON CONFLICT DO UPDATE preserves that rowid, so a reparse keeps its identity.\n await stages.time('text-index', () => dialect.reconcileContent(conn, contentTouched, parsedDocs, delta, cfg));\n\n // A preset edit forces a full rebuild, so an unchanged doc's coverage is already correct;\n // new docs have nothing to clear, which keeps cold builds linear.\n await stages.time('presets', async () => {\n if (touched.length > 0)\n await conn.runBatch(\n 'DELETE FROM preset_files WHERE \"path\" = ?',\n touched.map((p) => [p])\n );\n const presetRows: unknown[][] = [];\n for (const doc of parsedDocs) for (const presetName of doc.presets) presetRows.push([doc.relPath, presetName]);\n // DO NOTHING, not a bare INSERT: the added/touched split above comes from a read taken\n // before this transaction's lock, so a path this process calls \"added\" can already have\n // its (path, preset) row committed by a concurrent reconcile -- the row would be identical either way.\n await appendRows(conn, 'preset_files', ['path', 'preset'], 'INSERT INTO preset_files (\"path\", preset) VALUES (?, ?) ON CONFLICT(\"path\", preset) DO NOTHING', presetRows);\n });\n\n // Before the feature hooks, never after: rank's afterReconcile reads frontmatter as PageRank's\n // node set, so a lingering vanished row would dilute rank mass across every surviving note.\n if (vanished.length > 0)\n await stages.time('vanished', () =>\n conn.runBatch(\n 'DELETE FROM frontmatter WHERE \"path\" = ?',\n vanished.map((p) => [p])\n )\n );\n\n // Timed per feature per hook, so link resolution and PageRank are named stages without\n // links.ts or rank.ts knowing anything about this, and a new feature is visible for free.\n if (touched.length > 0) for (const feature of features) await stages.time(featureStage(feature.name, 'remove'), () => feature.remove?.(conn, touched, delta));\n for (const feature of features) {\n const docsForFeature: ExtractedDoc[] = parsedDocs.map((doc) => ({ path: doc.relPath, extracted: doc.extracted[feature.name] }));\n await stages.time(featureStage(feature.name, 'store'), () => feature.store?.(conn, docsForFeature, delta));\n }\n for (const feature of features) await stages.time(featureStage(feature.name, 'after'), () => feature.afterReconcile?.(conn, delta));\n },\n dialect.beginMode()\n );\n\n const durationMs = Date.now() - txStart;\n // Every store, not just the ones with a PRAGMA to derive: connectUnlocked's lock-wait budget needs it too.\n recordLockWaitMs(configDir, durationMs);\n if (dialect.recordDuration) await stages.time('meta', () => dialect.recordDuration?.(conn, durationMs));\n\n return { parsed: parsedDocs.length, warnings, stages: stages.take(elapsed(), durationMs, workerParseMs) };\n}\n"],"names":["activeFeatures","progress","listFiles","RESERVED_COLUMNS","reparseFiles","recordLockWaitMs","appendRows","getColumns","quoteIdent","featureStage","stageRecorder","withTransaction","CORE_FRONTMATTER_COLUMNS","Set","reconcile","conn","cfg","paths","dialect","pool","forcedPaths","rootDir","configDir","start","process","hrtime","bigint","features","stages","map","f","name","elapsed","Number","files","time","currentSet","relPath","existingRows","existingStmt","prepare","all","existing","Map","r","path","vanished","filter","has","toReparse","row","get","_mtime","mtimeMs","_size","size","length","parsed","warnings","take","seenColumns","report","docs","parsedDocs","newColumns","workerParseMs","tick","finish","col","add","allColumns","checkColumnLimit","writableColumns","c","insertSql","join","added","delta","reparsed","d","addedSet","reparsedExisting","p","touched","txStart","Date","now","present","missingColumns","addColumns","contentTouched","currentPathsStmt","currentPaths","staleAdded","toRow","doc","ctimeMs","parseError","data","newDocs","updateDocs","runBatch","reconcileContent","presetRows","presetName","presets","push","feature","remove","docsForFeature","extracted","store","afterReconcile","beginMode","durationMs","recordDuration"],"mappings":"AACA,SAASA,cAAc,QAAQ,uBAAuB;AAEtD,SAASC,QAAQ,QAAQ,wBAAwB;AACjD,SAASC,SAAS,EAAEC,gBAAgB,QAAQ,mBAAmB;AAE/D,SAASC,YAAY,QAAQ,qBAAqB;AAClD,SAASC,gBAAgB,QAAQ,iBAAiB;AAClD,SAASC,UAAU,EAAEC,UAAU,EAAEC,UAAU,QAAQ,cAAc;AACjE,SAASC,YAAY,EAAeC,aAAa,QAAQ,cAAc;AACvE,SAASC,eAAe,QAAQ,mBAAmB;AAGnD,gGAAgG;AAChG,gGAAgG;AAChG,qGAAqG;AACrG,6FAA6F;AAC7F,gFAAgF;AAEhF,6FAA6F;AAC7F,4EAA4E;AAC5E,OAAO,MAAMC,2BAA2B,IAAIC,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;CAAe,EAAE;AAOvG,OAAO,eAAeC,UAAUC,IAAgB,EAAEC,GAAW,EAAEC,KAAqB,EAAEC,OAAyB,EAAEC,IAAgB,EAAEC,WAAiC;IAClK,MAAM,EAAEC,OAAO,EAAEC,SAAS,EAAE,GAAGL;IAC/B,MAAMM,QAAQC,QAAQC,MAAM,CAACC,MAAM;IACnC,MAAMC,WAAW3B,eAAegB;IAChC,MAAMY,SAASlB,cAAciB,SAASE,GAAG,CAAC,CAACC,IAAMA,EAAEC,IAAI;IACvD,MAAMC,UAAU,IAAMC,OAAOT,QAAQC,MAAM,CAACC,MAAM,KAAKH,SAAS;IAChE,MAAMW,QAAQ,MAAMN,OAAOO,IAAI,CAAC,QAAQ,IAAMjC,UAAUc,KAAKK;IAC7D,MAAMe,aAAa,IAAIvB,IAAIqB,MAAML,GAAG,CAAC,CAACC,IAAMA,EAAEO,OAAO;IAErD,MAAMC,eAAe,MAAMV,OAAOO,IAAI,CAAC,YAAY;QACjD,MAAMI,eAAe,MAAMxB,KAAKyB,OAAO,CAAC;QACxC,OAAQ,MAAMD,aAAaE,GAAG;IAChC;IACA,MAAMC,WAAW,IAAIC,IAAIL,aAAaT,GAAG,CAAC,CAACe,IAAM;YAACA,EAAEC,IAAI;YAAED;SAAE;IAC5D,+FAA+F;IAC/F,+FAA+F;IAC/F,gFAAgF;IAChF,MAAME,WAAWR,aAAaS,MAAM,CAAC,CAACH,IAAM,CAACR,WAAWY,GAAG,CAACJ,EAAEC,IAAI,GAAGhB,GAAG,CAAC,CAACe,IAAMA,EAAEC,IAAI;IAEtF,6FAA6F;IAC7F,0FAA0F;IAC1F,MAAMI,YAAYf,MAAMa,MAAM,CAAC,CAACjB;;QAC9B,MAAMoB,MAAMR,SAASS,GAAG,CAACrB,EAAEO,OAAO;QAClC,OAAO,CAACa,OAAOA,IAAIE,MAAM,KAAKtB,EAAEuB,OAAO,IAAIH,IAAII,KAAK,KAAKxB,EAAEyB,IAAI,aAAKnC,wBAAAA,kCAAAA,YAAa4B,GAAG,CAAClB,EAAEO,OAAO,wCAAK;IACrG;IAEA,IAAIS,SAASU,MAAM,KAAK,KAAKP,UAAUO,MAAM,KAAK,GAAG,OAAO;QAAEC,QAAQ;QAAGC,UAAU,EAAE;QAAE9B,QAAQA,OAAO+B,IAAI,CAAC3B,WAAW;IAAG;IAEzH,MAAM4B,cAAc,MAAMrD,WAAWQ;IAErC,oFAAoF;IACpF,uDAAuD;IACvD,MAAM8C,SAAS5D,SAAS,mBAAmBgD,UAAUO,MAAM;IAC3D,wFAAwF;IACxF,MAAM,EAAEM,MAAMC,UAAU,EAAEL,QAAQ,EAAEM,UAAU,EAAEC,aAAa,EAAE,GAAG,MAAMrC,OAAOO,IAAI,CAAC,SAAS,IAAM/B,aAAa6C,WAAWtB,UAAUX,KAAK4C,aAAaC,OAAOK,IAAI,EAAE;YAAE/C;QAAK;IAC3K0C,OAAOM,MAAM;IACb,KAAK,MAAMC,OAAOJ,WAAYJ,YAAYS,GAAG,CAACD;IAE9C,MAAME,aAAa;WAAIV;KAAY;IACnC,yFAAyF;IACzF,iGAAiG;IACjG1C,QAAQqD,gBAAgB,CAACD,WAAWd,MAAM;IAC1C,0FAA0F;IAC1F,sEAAsE;IACtE,MAAMgB,kBAAkBF,WAAWvB,MAAM,CAAC,CAAC0B,IAAM7D,yBAAyBoC,GAAG,CAACyB,MAAM,CAACtE,iBAAiB6C,GAAG,CAACyB;IAC1G,yFAAyF;IACzF,mDAAmD;IACnD,MAAMC,YAAY,CAAC,yBAAyB,EAAEF,gBAAgB3C,GAAG,CAACrB,YAAYmE,IAAI,CAAC,MAAM,UAAU,EAAEH,gBAAgB3C,GAAG,CAAC,IAAM,KAAK8C,IAAI,CAAC,MAAM,oCAAoC,EAAEH,gBAClLzB,MAAM,CAAC,CAAC0B,IAAMA,MAAM,QACpB5C,GAAG,CAAC,CAAC4C,IAAM,GAAGjE,WAAWiE,GAAG,YAAY,EAAEjE,WAAWiE,IAAI,EACzDE,IAAI,CAAC,OAAO;IAEf,MAAMC,QAAQ3B,UAAUF,MAAM,CAAC,CAACjB,IAAM,CAACY,SAASM,GAAG,CAAClB,EAAEO,OAAO,GAAGR,GAAG,CAAC,CAACC,IAAMA,EAAEO,OAAO;IACpF,MAAMwC,QAAwB;QAAE3C;QAAO4C,UAAUf,WAAWlC,GAAG,CAAC,CAACkD,IAAMA,EAAE1C,OAAO;QAAGuC;QAAO9B;IAAS;IACnG,MAAMkC,WAAW,IAAInE,IAAI+D;IACzB,MAAMK,mBAAmBlB,WAAWlC,GAAG,CAAC,CAACkD,IAAMA,EAAE1C,OAAO,EAAEU,MAAM,CAAC,CAACmC,IAAM,CAACF,SAAShC,GAAG,CAACkC;IACtF,+FAA+F;IAC/F,0EAA0E;IAC1E,MAAMC,UAAU;WAAIrC;WAAamC;KAAiB;IAElD,MAAMG,UAAUC,KAAKC,GAAG;IACxB,MAAM3E,gBACJI,MACA;QACE,iGAAiG;QACjG,sFAAsF;QACtF,MAAMwE,UAAU,MAAMhF,WAAWQ;QACjC,MAAMyE,iBAAiBxB,WAAWjB,MAAM,CAAC,CAACqB,MAAQ,CAACmB,QAAQvC,GAAG,CAACoB;QAC/D,IAAIoB,eAAehC,MAAM,GAAG,GAAG,MAAM5B,OAAOO,IAAI,CAAC,SAAS,IAAMjB,QAAQuE,UAAU,CAAC1E,MAAMyE;QAEzF,yFAAyF;QACzF,sFAAsF;QACtF,sFAAsF;QACtF,sFAAsF;QACtF,4FAA4F;QAC5F,IAAIE,iBAAiBP;QACrB,IAAIP,MAAMpB,MAAM,GAAG,GACjBkC,iBAAiB,MAAM9D,OAAOO,IAAI,CAAC,iBAAiB;YAClD,MAAMwD,mBAAmB,MAAM5E,KAAKyB,OAAO,CAAC;YAC5C,MAAMoD,eAAe,IAAI/E,IAAI,AAAE,CAAA,MAAM8E,iBAAiBlD,GAAG,EAAC,EAA+BZ,GAAG,CAAC,CAACe,IAAMA,EAAEC,IAAI;YAC1G,MAAMgD,aAAajB,MAAM7B,MAAM,CAAC,CAACmC,IAAMU,aAAa5C,GAAG,CAACkC;YACxD,OAAOW,WAAWrC,MAAM,GAAG,IAAI;mBAAI2B;mBAAYU;aAAW,GAAGV;QAC/D;QAEF,IAAIpB,WAAWP,MAAM,GAAG,GAAG;YACzB,MAAMsC,QAAQ,CAACC,MACbvB,gBAAgB3C,GAAG,CAAC,CAACuC;wBAOZ2B;oBANP,IAAI3B,QAAQ,QAAQ,OAAO2B,IAAI1D,OAAO;oBACtC,IAAI+B,QAAQ,UAAU,OAAO2B,IAAI1C,OAAO;oBACxC,IAAIe,QAAQ,UAAU,OAAO2B,IAAIC,OAAO;oBACxC,IAAI5B,QAAQ,SAAS,OAAO2B,IAAIxC,IAAI;oBACpC,mFAAmF;oBACnF,IAAIa,QAAQ,gBAAgB,OAAO2B,IAAIE,UAAU;oBACjD,QAAOF,gBAAAA,IAAIG,IAAI,CAAC9B,IAAI,cAAb2B,2BAAAA,gBAAiB;gBAC1B;YACF,wFAAwF;YACxF,sCAAsC;YACtC,MAAMnE,OAAOO,IAAI,CAAC,aAAa;gBAC7B,MAAMgE,UAAUpC,WAAWhB,MAAM,CAAC,CAACgC,IAAMC,SAAShC,GAAG,CAAC+B,EAAE1C,OAAO;gBAC/D,MAAM+D,aAAarC,WAAWhB,MAAM,CAAC,CAACgC,IAAM,CAACC,SAAShC,GAAG,CAAC+B,EAAE1C,OAAO;gBACnE,MAAM/B,WAAWS,MAAM,eAAeyD,iBAAiBE,WAAWyB,QAAQtE,GAAG,CAACiE;gBAC9E,IAAIM,WAAW5C,MAAM,GAAG,GAAG,MAAMzC,KAAKsF,QAAQ,CAAC3B,WAAW0B,WAAWvE,GAAG,CAACiE;YAC3E;QACF;QAEA,wFAAwF;QACxF,2FAA2F;QAC3F,MAAMlE,OAAOO,IAAI,CAAC,cAAc,IAAMjB,QAAQoF,gBAAgB,CAACvF,MAAM2E,gBAAgB3B,YAAYc,OAAO7D;QAExG,0FAA0F;QAC1F,kEAAkE;QAClE,MAAMY,OAAOO,IAAI,CAAC,WAAW;YAC3B,IAAIgD,QAAQ3B,MAAM,GAAG,GACnB,MAAMzC,KAAKsF,QAAQ,CACjB,6CACAlB,QAAQtD,GAAG,CAAC,CAACqD,IAAM;oBAACA;iBAAE;YAE1B,MAAMqB,aAA0B,EAAE;YAClC,KAAK,MAAMR,OAAOhC,WAAY,KAAK,MAAMyC,cAAcT,IAAIU,OAAO,CAAEF,WAAWG,IAAI,CAAC;gBAACX,IAAI1D,OAAO;gBAAEmE;aAAW;YAC7G,uFAAuF;YACvF,wFAAwF;YACxF,uGAAuG;YACvG,MAAMlG,WAAWS,MAAM,gBAAgB;gBAAC;gBAAQ;aAAS,EAAE,kGAAkGwF;QAC/J;QAEA,+FAA+F;QAC/F,4FAA4F;QAC5F,IAAIzD,SAASU,MAAM,GAAG,GACpB,MAAM5B,OAAOO,IAAI,CAAC,YAAY,IAC5BpB,KAAKsF,QAAQ,CACX,4CACAvD,SAASjB,GAAG,CAAC,CAACqD,IAAM;oBAACA;iBAAE;QAI7B,uFAAuF;QACvF,0FAA0F;QAC1F,IAAIC,QAAQ3B,MAAM,GAAG,GAAG,KAAK,MAAMmD,WAAWhF,SAAU,MAAMC,OAAOO,IAAI,CAAC1B,aAAakG,QAAQ5E,IAAI,EAAE,WAAW;gBAAM4E;oBAAAA,kBAAAA,QAAQC,MAAM,cAAdD,sCAAAA,qBAAAA,SAAiB5F,MAAMoE,SAASN;;QACtJ,KAAK,MAAM8B,WAAWhF,SAAU;YAC9B,MAAMkF,iBAAiC9C,WAAWlC,GAAG,CAAC,CAACkE,MAAS,CAAA;oBAAElD,MAAMkD,IAAI1D,OAAO;oBAAEyE,WAAWf,IAAIe,SAAS,CAACH,QAAQ5E,IAAI,CAAC;gBAAC,CAAA;YAC5H,MAAMH,OAAOO,IAAI,CAAC1B,aAAakG,QAAQ5E,IAAI,EAAE,UAAU;oBAAM4E;wBAAAA,iBAAAA,QAAQI,KAAK,cAAbJ,qCAAAA,oBAAAA,SAAgB5F,MAAM8F,gBAAgBhC;;QACrG;QACA,KAAK,MAAM8B,WAAWhF,SAAU,MAAMC,OAAOO,IAAI,CAAC1B,aAAakG,QAAQ5E,IAAI,EAAE,UAAU;gBAAM4E;oBAAAA,0BAAAA,QAAQK,cAAc,cAAtBL,8CAAAA,6BAAAA,SAAyB5F,MAAM8D;;IAC9H,GACA3D,QAAQ+F,SAAS;IAGnB,MAAMC,aAAa7B,KAAKC,GAAG,KAAKF;IAChC,2GAA2G;IAC3G/E,iBAAiBiB,WAAW4F;IAC5B,IAAIhG,QAAQiG,cAAc,EAAE,MAAMvF,OAAOO,IAAI,CAAC,QAAQ;YAAMjB;gBAAAA,0BAAAA,QAAQiG,cAAc,cAAtBjG,8CAAAA,6BAAAA,SAAyBH,MAAMmG;;IAE3F,OAAO;QAAEzD,QAAQM,WAAWP,MAAM;QAAEE;QAAU9B,QAAQA,OAAO+B,IAAI,CAAC3B,WAAWkF,YAAYjD;IAAe;AAC1G"}
@@ -106,7 +106,7 @@ export function createConnection(db) {
106
106
  // can hold across many batches, and nothing else owns this statement's lifetime.
107
107
  const stmt = await db.prepare(sql);
108
108
  try {
109
- for (const row of paramRows)await stmt.run(...row);
109
+ for (const row of paramRows)await stmt.run(row);
110
110
  } finally{
111
111
  await stmt.close();
112
112
  }
@@ -1 +1 @@
1
- {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/connection.ts"],"sourcesContent":["import { stat } from 'node:fs/promises';\nimport type { Database } from '@tursodatabase/database';\nimport { rewriteInsert } from '../batch.ts';\nimport { setMeta } from '../shared.ts';\nimport { BEGIN_WRITE, withTransaction } from '../transaction.ts';\nimport type { Connection, RunResult, Statement } from '../types.ts';\nimport { CONNECT_OPTS, tursoApi } from './native.ts';\nimport { rewriteFunctions } from './sql-functions.ts';\n\n// The client's own Statement class isn't re-exported by name from '@tursodatabase/database',\n// so its type is derived structurally from Database.prepare()'s return type instead.\ntype TursoStatement = Awaited<ReturnType<Database['prepare']>>;\n\n// @tursodatabase/database/compat offers a synchronous escape hatch, but it measured only 4-12%\n// faster than this promise client, so this wraps the async client in the same Connection/Statement shape 1:1.\nclass TursoStatementWrapper implements Statement {\n private stmt: TursoStatement;\n\n constructor(stmt: TursoStatement) {\n this.stmt = stmt;\n }\n\n async run(...params: unknown[]): Promise<RunResult> {\n return this.stmt.run(...params);\n }\n\n async get(...params: unknown[]): Promise<unknown> {\n return this.stmt.get(...params);\n }\n\n async all(...params: unknown[]): Promise<unknown[]> {\n return this.stmt.all(...params);\n }\n\n async *iterate(...params: unknown[]): AsyncIterable<unknown> {\n yield* this.stmt.iterate(...params);\n }\n\n columns(): Array<{ name: string }> {\n return this.stmt.columns();\n }\n\n setReadBigInts(enabled: boolean): void {\n this.stmt.safeIntegers(enabled);\n }\n}\n\n// This client leaves the WAL behind for the next opener, where node:sqlite checkpoints on the last\n// close, so a tree reconciled over and over grows one without bound. Best-effort: close must not throw.\nexport async function checkpointWal(db: Database): Promise<void> {\n try {\n await db.exec('PRAGMA wal_checkpoint(TRUNCATE)');\n } catch (err) {\n console.error(`sense: turso WAL checkpoint failed, the -wal file will keep growing until one succeeds: ${(err as Error).message}`);\n }\n}\n\n// The main cache file's path, or null when unresolvable. The store holds a Database and\n// Connection, not a path (PLAN 3.52), so the path is read from the database itself.\nexport async function cacheFilePath(conn: Connection): Promise<string | null> {\n const rows = (await (await conn.prepare('PRAGMA database_list')).all()) as Array<{ name: string; file: string }>;\n return rows.find((r) => r.name === 'main')?.file || null;\n}\n\nexport async function fileSize(path: string): Promise<number | null> {\n try {\n return (await stat(path)).size;\n } catch {\n return null;\n }\n}\n\n// Reclaims turso#8170's FTS space amplification (PLAN 3.41), on a throwaway connection after the\n// store's own has closed: `vacuum` doubles incremental reconcile cost merely by being enabled (3.54).\nexport async function reclaimSpace(path: string): Promise<void> {\n try {\n const turso = await tursoApi();\n const db = await turso.connect(path, { ...CONNECT_OPTS, experimental: [...CONNECT_OPTS.experimental, 'vacuum'] });\n try {\n await db.exec('VACUUM');\n const size = await fileSize(path);\n // Recorded here, not by the caller: this connection is the only one still open on the file.\n if (size !== null) await setMeta(createConnection(db), 'compact_size', String(size));\n await checkpointWal(db);\n } finally {\n await db.close();\n }\n } catch (err) {\n console.error(`sense: turso VACUUM failed, the cache will keep the space until one succeeds: ${(err as Error).message}`);\n }\n}\n\nexport function createConnection(db: Database): Connection {\n const conn: Connection = {\n async exec(sql: string): Promise<void> {\n await db.exec(sql);\n },\n async prepare(sql: string): Promise<Statement> {\n return new TursoStatementWrapper(await db.prepare(rewriteFunctions(sql)));\n },\n // Folds a plain INSERT into one multi-row VALUES statement (shared rewriteInsert, ../batch.ts):\n // no bind-variable ceiling to chunk against, measured empirically. UPDATE/DELETE keep the per-row loop.\n async runBatch(sql: string, paramRows: unknown[][]): Promise<void> {\n if (paramRows.length === 0) return;\n await withTransaction(\n conn,\n async () => {\n const rewritten = rewriteInsert(sql, paramRows.length);\n if (rewritten) {\n const stmt = await db.prepare(rewritten.sql);\n try {\n // One flat array, not spread: spreading tens of thousands of args hits a JS call-stack\n // limit well before turso enforces any variable-count ceiling (measured; see PLAN.md).\n await stmt.run(paramRows.flat());\n } finally {\n await stmt.close();\n }\n return;\n }\n // Finalized here because nothing else will: open() hands back a connection the caller\n // can hold across many batches, and nothing else owns this statement's lifetime.\n const stmt = await db.prepare(sql);\n try {\n for (const row of paramRows) await stmt.run(...row);\n } finally {\n await stmt.close();\n }\n },\n BEGIN_WRITE\n );\n },\n };\n return conn;\n}\n"],"names":["stat","rewriteInsert","setMeta","BEGIN_WRITE","withTransaction","CONNECT_OPTS","tursoApi","rewriteFunctions","TursoStatementWrapper","run","params","stmt","get","all","iterate","columns","setReadBigInts","enabled","safeIntegers","checkpointWal","db","exec","err","console","error","message","cacheFilePath","conn","rows","prepare","find","r","name","file","fileSize","path","size","reclaimSpace","turso","connect","experimental","createConnection","String","close","sql","runBatch","paramRows","length","rewritten","flat","row"],"mappings":"AAAA,SAASA,IAAI,QAAQ,mBAAmB;AAExC,SAASC,aAAa,QAAQ,cAAc;AAC5C,SAASC,OAAO,QAAQ,eAAe;AACvC,SAASC,WAAW,EAAEC,eAAe,QAAQ,oBAAoB;AAEjE,SAASC,YAAY,EAAEC,QAAQ,QAAQ,cAAc;AACrD,SAASC,gBAAgB,QAAQ,qBAAqB;AAMtD,+FAA+F;AAC/F,8GAA8G;AAC9G,IAAA,AAAMC,wBAAN,MAAMA;IAOJ,MAAMC,IAAI,GAAGC,MAAiB,EAAsB;QAClD,OAAO,IAAI,CAACC,IAAI,CAACF,GAAG,IAAIC;IAC1B;IAEA,MAAME,IAAI,GAAGF,MAAiB,EAAoB;QAChD,OAAO,IAAI,CAACC,IAAI,CAACC,GAAG,IAAIF;IAC1B;IAEA,MAAMG,IAAI,GAAGH,MAAiB,EAAsB;QAClD,OAAO,IAAI,CAACC,IAAI,CAACE,GAAG,IAAIH;IAC1B;IAEA,OAAOI,QAAQ,GAAGJ,MAAiB,EAA0B;QAC3D,OAAO,IAAI,CAACC,IAAI,CAACG,OAAO,IAAIJ;IAC9B;IAEAK,UAAmC;QACjC,OAAO,IAAI,CAACJ,IAAI,CAACI,OAAO;IAC1B;IAEAC,eAAeC,OAAgB,EAAQ;QACrC,IAAI,CAACN,IAAI,CAACO,YAAY,CAACD;IACzB;IA1BA,YAAYN,IAAoB,CAAE;QAChC,IAAI,CAACA,IAAI,GAAGA;IACd;AAyBF;AAEA,mGAAmG;AACnG,wGAAwG;AACxG,OAAO,eAAeQ,cAAcC,EAAY;IAC9C,IAAI;QACF,MAAMA,GAAGC,IAAI,CAAC;IAChB,EAAE,OAAOC,KAAK;QACZC,QAAQC,KAAK,CAAC,CAAC,wFAAwF,EAAE,AAACF,IAAcG,OAAO,EAAE;IACnI;AACF;AAEA,wFAAwF;AACxF,oFAAoF;AACpF,OAAO,eAAeC,cAAcC,IAAgB;QAE3CC;IADP,MAAMA,OAAQ,MAAM,AAAC,CAAA,MAAMD,KAAKE,OAAO,CAAC,uBAAsB,EAAGhB,GAAG;IACpE,OAAOe,EAAAA,aAAAA,KAAKE,IAAI,CAAC,CAACC,IAAMA,EAAEC,IAAI,KAAK,qBAA5BJ,iCAAAA,WAAqCK,IAAI,KAAI;AACtD;AAEA,OAAO,eAAeC,SAASC,IAAY;IACzC,IAAI;QACF,OAAO,AAAC,CAAA,MAAMnC,KAAKmC,KAAI,EAAGC,IAAI;IAChC,EAAE,OAAM;QACN,OAAO;IACT;AACF;AAEA,iGAAiG;AACjG,sGAAsG;AACtG,OAAO,eAAeC,aAAaF,IAAY;IAC7C,IAAI;QACF,MAAMG,QAAQ,MAAMhC;QACpB,MAAMc,KAAK,MAAMkB,MAAMC,OAAO,CAACJ,MAAM;YAAE,GAAG9B,YAAY;YAAEmC,cAAc;mBAAInC,aAAamC,YAAY;gBAAE;aAAS;QAAC;QAC/G,IAAI;YACF,MAAMpB,GAAGC,IAAI,CAAC;YACd,MAAMe,OAAO,MAAMF,SAASC;YAC5B,4FAA4F;YAC5F,IAAIC,SAAS,MAAM,MAAMlC,QAAQuC,iBAAiBrB,KAAK,gBAAgBsB,OAAON;YAC9E,MAAMjB,cAAcC;QACtB,SAAU;YACR,MAAMA,GAAGuB,KAAK;QAChB;IACF,EAAE,OAAOrB,KAAK;QACZC,QAAQC,KAAK,CAAC,CAAC,8EAA8E,EAAE,AAACF,IAAcG,OAAO,EAAE;IACzH;AACF;AAEA,OAAO,SAASgB,iBAAiBrB,EAAY;IAC3C,MAAMO,OAAmB;QACvB,MAAMN,MAAKuB,GAAW;YACpB,MAAMxB,GAAGC,IAAI,CAACuB;QAChB;QACA,MAAMf,SAAQe,GAAW;YACvB,OAAO,IAAIpC,sBAAsB,MAAMY,GAAGS,OAAO,CAACtB,iBAAiBqC;QACrE;QACA,gGAAgG;QAChG,wGAAwG;QACxG,MAAMC,UAASD,GAAW,EAAEE,SAAsB;YAChD,IAAIA,UAAUC,MAAM,KAAK,GAAG;YAC5B,MAAM3C,gBACJuB,MACA;gBACE,MAAMqB,YAAY/C,cAAc2C,KAAKE,UAAUC,MAAM;gBACrD,IAAIC,WAAW;oBACb,MAAMrC,OAAO,MAAMS,GAAGS,OAAO,CAACmB,UAAUJ,GAAG;oBAC3C,IAAI;wBACF,uFAAuF;wBACvF,uFAAuF;wBACvF,MAAMjC,KAAKF,GAAG,CAACqC,UAAUG,IAAI;oBAC/B,SAAU;wBACR,MAAMtC,KAAKgC,KAAK;oBAClB;oBACA;gBACF;gBACA,sFAAsF;gBACtF,iFAAiF;gBACjF,MAAMhC,OAAO,MAAMS,GAAGS,OAAO,CAACe;gBAC9B,IAAI;oBACF,KAAK,MAAMM,OAAOJ,UAAW,MAAMnC,KAAKF,GAAG,IAAIyC;gBACjD,SAAU;oBACR,MAAMvC,KAAKgC,KAAK;gBAClB;YACF,GACAxC;QAEJ;IACF;IACA,OAAOwB;AACT"}
1
+ {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/connection.ts"],"sourcesContent":["import { stat } from 'node:fs/promises';\nimport type { Database } from '@tursodatabase/database';\nimport { rewriteInsert } from '../batch.ts';\nimport { setMeta } from '../shared.ts';\nimport { BEGIN_WRITE, withTransaction } from '../transaction.ts';\nimport type { Connection, RunResult, Statement } from '../types.ts';\nimport { CONNECT_OPTS, tursoApi } from './native.ts';\nimport { rewriteFunctions } from './sql-functions.ts';\n\n// The client's own Statement class isn't re-exported by name from '@tursodatabase/database',\n// so its type is derived structurally from Database.prepare()'s return type instead.\ntype TursoStatement = Awaited<ReturnType<Database['prepare']>>;\n\n// @tursodatabase/database/compat offers a synchronous escape hatch, but it measured only 4-12%\n// faster than this promise client, so this wraps the async client in the same Connection/Statement shape 1:1.\nclass TursoStatementWrapper implements Statement {\n private stmt: TursoStatement;\n\n constructor(stmt: TursoStatement) {\n this.stmt = stmt;\n }\n\n async run(...params: unknown[]): Promise<RunResult> {\n return this.stmt.run(...params);\n }\n\n async get(...params: unknown[]): Promise<unknown> {\n return this.stmt.get(...params);\n }\n\n async all(...params: unknown[]): Promise<unknown[]> {\n return this.stmt.all(...params);\n }\n\n async *iterate(...params: unknown[]): AsyncIterable<unknown> {\n yield* this.stmt.iterate(...params);\n }\n\n columns(): Array<{ name: string }> {\n return this.stmt.columns();\n }\n\n setReadBigInts(enabled: boolean): void {\n this.stmt.safeIntegers(enabled);\n }\n}\n\n// This client leaves the WAL behind for the next opener, where node:sqlite checkpoints on the last\n// close, so a tree reconciled over and over grows one without bound. Best-effort: close must not throw.\nexport async function checkpointWal(db: Database): Promise<void> {\n try {\n await db.exec('PRAGMA wal_checkpoint(TRUNCATE)');\n } catch (err) {\n console.error(`sense: turso WAL checkpoint failed, the -wal file will keep growing until one succeeds: ${(err as Error).message}`);\n }\n}\n\n// The main cache file's path, or null when unresolvable. The store holds a Database and\n// Connection, not a path (PLAN 3.52), so the path is read from the database itself.\nexport async function cacheFilePath(conn: Connection): Promise<string | null> {\n const rows = (await (await conn.prepare('PRAGMA database_list')).all()) as Array<{ name: string; file: string }>;\n return rows.find((r) => r.name === 'main')?.file || null;\n}\n\nexport async function fileSize(path: string): Promise<number | null> {\n try {\n return (await stat(path)).size;\n } catch {\n return null;\n }\n}\n\n// Reclaims turso#8170's FTS space amplification (PLAN 3.41), on a throwaway connection after the\n// store's own has closed: `vacuum` doubles incremental reconcile cost merely by being enabled (3.54).\nexport async function reclaimSpace(path: string): Promise<void> {\n try {\n const turso = await tursoApi();\n const db = await turso.connect(path, { ...CONNECT_OPTS, experimental: [...CONNECT_OPTS.experimental, 'vacuum'] });\n try {\n await db.exec('VACUUM');\n const size = await fileSize(path);\n // Recorded here, not by the caller: this connection is the only one still open on the file.\n if (size !== null) await setMeta(createConnection(db), 'compact_size', String(size));\n await checkpointWal(db);\n } finally {\n await db.close();\n }\n } catch (err) {\n console.error(`sense: turso VACUUM failed, the cache will keep the space until one succeeds: ${(err as Error).message}`);\n }\n}\n\nexport function createConnection(db: Database): Connection {\n const conn: Connection = {\n async exec(sql: string): Promise<void> {\n await db.exec(sql);\n },\n async prepare(sql: string): Promise<Statement> {\n return new TursoStatementWrapper(await db.prepare(rewriteFunctions(sql)));\n },\n // Folds a plain INSERT into one multi-row VALUES statement (shared rewriteInsert, ../batch.ts):\n // no bind-variable ceiling to chunk against, measured empirically. UPDATE/DELETE keep the per-row loop.\n async runBatch(sql: string, paramRows: unknown[][]): Promise<void> {\n if (paramRows.length === 0) return;\n await withTransaction(\n conn,\n async () => {\n const rewritten = rewriteInsert(sql, paramRows.length);\n if (rewritten) {\n const stmt = await db.prepare(rewritten.sql);\n try {\n // One flat array, not spread: spreading tens of thousands of args hits a JS call-stack\n // limit well before turso enforces any variable-count ceiling (measured; see PLAN.md).\n await stmt.run(paramRows.flat());\n } finally {\n await stmt.close();\n }\n return;\n }\n // Finalized here because nothing else will: open() hands back a connection the caller\n // can hold across many batches, and nothing else owns this statement's lifetime.\n const stmt = await db.prepare(sql);\n try {\n for (const row of paramRows) await stmt.run(row);\n } finally {\n await stmt.close();\n }\n },\n BEGIN_WRITE\n );\n },\n };\n return conn;\n}\n"],"names":["stat","rewriteInsert","setMeta","BEGIN_WRITE","withTransaction","CONNECT_OPTS","tursoApi","rewriteFunctions","TursoStatementWrapper","run","params","stmt","get","all","iterate","columns","setReadBigInts","enabled","safeIntegers","checkpointWal","db","exec","err","console","error","message","cacheFilePath","conn","rows","prepare","find","r","name","file","fileSize","path","size","reclaimSpace","turso","connect","experimental","createConnection","String","close","sql","runBatch","paramRows","length","rewritten","flat","row"],"mappings":"AAAA,SAASA,IAAI,QAAQ,mBAAmB;AAExC,SAASC,aAAa,QAAQ,cAAc;AAC5C,SAASC,OAAO,QAAQ,eAAe;AACvC,SAASC,WAAW,EAAEC,eAAe,QAAQ,oBAAoB;AAEjE,SAASC,YAAY,EAAEC,QAAQ,QAAQ,cAAc;AACrD,SAASC,gBAAgB,QAAQ,qBAAqB;AAMtD,+FAA+F;AAC/F,8GAA8G;AAC9G,IAAA,AAAMC,wBAAN,MAAMA;IAOJ,MAAMC,IAAI,GAAGC,MAAiB,EAAsB;QAClD,OAAO,IAAI,CAACC,IAAI,CAACF,GAAG,IAAIC;IAC1B;IAEA,MAAME,IAAI,GAAGF,MAAiB,EAAoB;QAChD,OAAO,IAAI,CAACC,IAAI,CAACC,GAAG,IAAIF;IAC1B;IAEA,MAAMG,IAAI,GAAGH,MAAiB,EAAsB;QAClD,OAAO,IAAI,CAACC,IAAI,CAACE,GAAG,IAAIH;IAC1B;IAEA,OAAOI,QAAQ,GAAGJ,MAAiB,EAA0B;QAC3D,OAAO,IAAI,CAACC,IAAI,CAACG,OAAO,IAAIJ;IAC9B;IAEAK,UAAmC;QACjC,OAAO,IAAI,CAACJ,IAAI,CAACI,OAAO;IAC1B;IAEAC,eAAeC,OAAgB,EAAQ;QACrC,IAAI,CAACN,IAAI,CAACO,YAAY,CAACD;IACzB;IA1BA,YAAYN,IAAoB,CAAE;QAChC,IAAI,CAACA,IAAI,GAAGA;IACd;AAyBF;AAEA,mGAAmG;AACnG,wGAAwG;AACxG,OAAO,eAAeQ,cAAcC,EAAY;IAC9C,IAAI;QACF,MAAMA,GAAGC,IAAI,CAAC;IAChB,EAAE,OAAOC,KAAK;QACZC,QAAQC,KAAK,CAAC,CAAC,wFAAwF,EAAE,AAACF,IAAcG,OAAO,EAAE;IACnI;AACF;AAEA,wFAAwF;AACxF,oFAAoF;AACpF,OAAO,eAAeC,cAAcC,IAAgB;QAE3CC;IADP,MAAMA,OAAQ,MAAM,AAAC,CAAA,MAAMD,KAAKE,OAAO,CAAC,uBAAsB,EAAGhB,GAAG;IACpE,OAAOe,EAAAA,aAAAA,KAAKE,IAAI,CAAC,CAACC,IAAMA,EAAEC,IAAI,KAAK,qBAA5BJ,iCAAAA,WAAqCK,IAAI,KAAI;AACtD;AAEA,OAAO,eAAeC,SAASC,IAAY;IACzC,IAAI;QACF,OAAO,AAAC,CAAA,MAAMnC,KAAKmC,KAAI,EAAGC,IAAI;IAChC,EAAE,OAAM;QACN,OAAO;IACT;AACF;AAEA,iGAAiG;AACjG,sGAAsG;AACtG,OAAO,eAAeC,aAAaF,IAAY;IAC7C,IAAI;QACF,MAAMG,QAAQ,MAAMhC;QACpB,MAAMc,KAAK,MAAMkB,MAAMC,OAAO,CAACJ,MAAM;YAAE,GAAG9B,YAAY;YAAEmC,cAAc;mBAAInC,aAAamC,YAAY;gBAAE;aAAS;QAAC;QAC/G,IAAI;YACF,MAAMpB,GAAGC,IAAI,CAAC;YACd,MAAMe,OAAO,MAAMF,SAASC;YAC5B,4FAA4F;YAC5F,IAAIC,SAAS,MAAM,MAAMlC,QAAQuC,iBAAiBrB,KAAK,gBAAgBsB,OAAON;YAC9E,MAAMjB,cAAcC;QACtB,SAAU;YACR,MAAMA,GAAGuB,KAAK;QAChB;IACF,EAAE,OAAOrB,KAAK;QACZC,QAAQC,KAAK,CAAC,CAAC,8EAA8E,EAAE,AAACF,IAAcG,OAAO,EAAE;IACzH;AACF;AAEA,OAAO,SAASgB,iBAAiBrB,EAAY;IAC3C,MAAMO,OAAmB;QACvB,MAAMN,MAAKuB,GAAW;YACpB,MAAMxB,GAAGC,IAAI,CAACuB;QAChB;QACA,MAAMf,SAAQe,GAAW;YACvB,OAAO,IAAIpC,sBAAsB,MAAMY,GAAGS,OAAO,CAACtB,iBAAiBqC;QACrE;QACA,gGAAgG;QAChG,wGAAwG;QACxG,MAAMC,UAASD,GAAW,EAAEE,SAAsB;YAChD,IAAIA,UAAUC,MAAM,KAAK,GAAG;YAC5B,MAAM3C,gBACJuB,MACA;gBACE,MAAMqB,YAAY/C,cAAc2C,KAAKE,UAAUC,MAAM;gBACrD,IAAIC,WAAW;oBACb,MAAMrC,OAAO,MAAMS,GAAGS,OAAO,CAACmB,UAAUJ,GAAG;oBAC3C,IAAI;wBACF,uFAAuF;wBACvF,uFAAuF;wBACvF,MAAMjC,KAAKF,GAAG,CAACqC,UAAUG,IAAI;oBAC/B,SAAU;wBACR,MAAMtC,KAAKgC,KAAK;oBAClB;oBACA;gBACF;gBACA,sFAAsF;gBACtF,iFAAiF;gBACjF,MAAMhC,OAAO,MAAMS,GAAGS,OAAO,CAACe;gBAC9B,IAAI;oBACF,KAAK,MAAMM,OAAOJ,UAAW,MAAMnC,KAAKF,GAAG,CAACyC;gBAC9C,SAAU;oBACR,MAAMvC,KAAKgC,KAAK;gBAClB;YACF,GACAxC;QAEJ;IACF;IACA,OAAOwB;AACT"}
@@ -40,10 +40,14 @@ const FTS5_OPERATORS = [
40
40
  {
41
41
  label: 'column filter',
42
42
  re: /(?:^|\s)[\p{L}_]\w*\s*:/u
43
+ },
44
+ {
45
+ label: 'grouping operator',
46
+ re: /[()]/
43
47
  }
44
48
  ];
45
49
  function unsupportedOperator(terms) {
46
- const withoutPhrases = terms.replace(/"[^"]*"/g, ' ');
50
+ const withoutPhrases = terms.replace(/"[^"]*"/g, 'phrase');
47
51
  for (const { label, re } of FTS5_OPERATORS){
48
52
  const m = withoutPhrases.match(re);
49
53
  if (m) return {
@@ -1 +1 @@
1
- {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/lexical.ts"],"sourcesContent":["// T1: Tantivy FTS via fts_match/fts_score on the default-tokenizer index for bare words and\n// quoted phrases, plus a second ngram-tokenized index over \"_ngram\" sidecars for unspaced-script runs.\nimport { SenseError } from '../../errors.ts';\nimport { matchesSearchPhrase, type SearchToken, searchTokens } from '../../text/segment.ts';\nimport type { Connection, LexicalHit, LexicalQueryOptions } from '../types.ts';\nimport { stemFolded } from './lexical-text.ts';\n\nconst FIELDS = ['title_stem', 'summary_stem', 'text_stem'] as const;\nconst NGRAM_FIELDS = ['title_ngram', 'summary_ngram', 'text_ngram'] as const;\nconst AUTHORED_FIELDS = ['title', 'summary', 'text'] as const;\ntype AuthoredField = (typeof AUTHORED_FIELDS)[number];\n\n// FTS5 operator syntax a bare query would otherwise treat as literal text (PRINCIPLES:\n// no-silent-modes). `^` is rejected either way: boost to Tantivy, initial-token match to FTS5.\nconst FTS5_OPERATORS: Array<{ label: string; re: RegExp }> = [\n { label: 'prefix query', re: /[\\p{L}\\p{N}_]+\\*(?=\\s|$)/u },\n { label: 'boolean operator', re: /(?:^|\\s)(?:AND|OR|NOT)(?=\\s|$)/ },\n { label: 'NEAR operator', re: /(?:^|\\s)NEAR\\b/ },\n { label: 'boost/initial-token operator', re: /\\S*\\^\\S*/ },\n { label: 'column filter', re: /(?:^|\\s)[\\p{L}_]\\w*\\s*:/u },\n];\n\nfunction unsupportedOperator(terms: string): { label: string; token: string } | null {\n const withoutPhrases = terms.replace(/\"[^\"]*\"/g, ' ');\n for (const { label, re } of FTS5_OPERATORS) {\n const m = withoutPhrases.match(re);\n if (m) return { label, token: m[0].trim() };\n }\n return null;\n}\n\n// A run whose script marks no word boundaries indexes as one opaque token under every tokenizer\n// but ngram, so such runs -- bare or quoted -- go to the ngram sidecar index instead.\nfunction splitTerms(terms: string): { words: string[]; phrases: SearchToken[][]; unspaced: string[]; emptyPhrase: boolean } {\n const words: string[] = [];\n const phrases: SearchToken[][] = [];\n const unspaced: string[] = [];\n let emptyPhrase = false;\n const withoutPhrases = terms.replace(/\"([^\"]*)\"/g, (_m, inner: string) => {\n const phrase = inner.trim();\n const tokens = searchTokens(phrase);\n if (tokens.length === 0) {\n emptyPhrase = true;\n return ' ';\n }\n unspaced.push(...tokens.filter((token) => token.unspaced).map((token) => token.text));\n phrases.push(tokens);\n return ' ';\n });\n for (const tok of withoutPhrases.split(/\\s+/)) {\n if (tok.length === 0) continue;\n for (const token of searchTokens(tok)) {\n if (token.unspaced) unspaced.push(token.text);\n else words.push(token.text);\n }\n }\n return { words, phrases, unspaced, emptyPhrase };\n}\n\n// fts_match/fts_score's query argument must be a SQL literal: bound as `?`, the row set stays\n// correct but fts_score silently returns 0 for every row (PRINCIPLES: no-silent-modes).\nfunction escapeFtsLiteral(query: string): string {\n return query.replace(/'/g, \"''\");\n}\n\n// Terms are quoted (\"O'Brien's\" is a parse error bare) and AND-joined: Tantivy's default for\n// bare terms is disjunctive, so space-joining would loosen \"apple banana\" into \"apple OR banana\".\nfunction buildQueries(words: string[], phrases: SearchToken[][], unspaced: string[]): { defaultQuery: string; ngramQuery: string } {\n const phraseParts = phrases.flatMap((phrase) => {\n const latin = phrase.filter((token) => !token.unspaced).map((token) => stemFolded(token.text));\n if (latin.length === 0) return [];\n return phrase.some((token) => token.unspaced) ? latin : [latin.join(' ')];\n });\n const defaultParts = [...words.map(stemFolded), ...phraseParts].map((t) => `\"${t}\"`);\n const unspacedParts = unspaced.map((t) => `\"${t}\"`);\n return { defaultQuery: defaultParts.join(' AND '), ngramQuery: unspacedParts.join(' AND ') };\n}\n\n// fts_score() is real only as a SELECT's bare projected expression, and only for the index that\n// SELECT's fts_match() chose; every other form silently returns 0. Hence one SELECT per index.\nfunction ftsBranch(cols: readonly string[], query: string, opts: LexicalQueryOptions): string {\n const colList = cols.map((c) => `content.${c}`).join(', ');\n const lit = escapeFtsLiteral(query);\n return `SELECT content.\"path\" AS path, fts_score(${colList}, '${lit}') AS score\n FROM content ${opts.whereJoin} WHERE fts_match(${colList}, '${lit}') ${opts.whereCond} ${opts.scopeCond}`;\n}\n\n// Ranked query, scoped by the caller-built SQL fragments (same shape sqlite's/duckdb's\n// queryLexical take). `snippets` is always NULL: fts_highlight returns the full column text, not a bounded snippet, so every row goes through the caller's JS snippet fallback (commands/search.ts).\nexport async function queryLexical(conn: Connection, terms: string, opts: LexicalQueryOptions): Promise<LexicalHit[]> {\n const unsupported = unsupportedOperator(terms);\n if (unsupported !== null) {\n throw new SenseError('STORE_CAPABILITY_MISSING', `store \"turso\" does not implement FTS5's ${unsupported.label} (\"${unsupported.token}\") in this build; rephrase \"${terms.trim()}\" without it, or set \"store\" to \"sqlite\" in this tree's config to search it as written`);\n }\n const { words, phrases, unspaced, emptyPhrase } = splitTerms(terms);\n if (emptyPhrase) return [];\n const { defaultQuery, ngramQuery } = buildQueries(words, phrases, unspaced);\n if (defaultQuery === '' && ngramQuery === '') return [];\n\n const branches: string[] = [];\n if (defaultQuery !== '') branches.push(ftsBranch(FIELDS, defaultQuery, opts));\n if (ngramQuery !== '') branches.push(ftsBranch(NGRAM_FIELDS, ngramQuery, opts));\n\n // HAVING COUNT(*) = branches.length is the AND across indexes; with one branch it is a no-op,\n // `path` being content's primary key. fts_score is higher-is-better BM25, so DESC. Phrase\n // candidates carry authored fields through the join below so the shared verifier can reject\n // cross-field, reordered, and non-adjacent matches before applying the caller's limit.\n const ranked = `SELECT path, SUM(score) AS score FROM (\n ${branches.join(' UNION ALL ')}\n ) parts GROUP BY path HAVING COUNT(*) = ${branches.length}`;\n const phraseSelect = phrases.length > 0 ? ', content.title, content.summary, content.text' : '';\n const phraseJoin = phrases.length > 0 ? ' JOIN content ON content.path = sq.path' : '';\n const sql = `SELECT sq.path, sq.score${phraseSelect} FROM (\n ${ranked}\n ) sq${phraseJoin} ORDER BY sq.score DESC, sq.path${phrases.length > 0 ? '' : ' LIMIT ?'}`;\n const stmt = await conn.prepare(sql);\n const rows = (await stmt.all(...(phrases.length > 0 ? [] : [opts.limit]))) as Array<LexicalHit & Partial<Record<AuthoredField, string>>>;\n // A concurrent FTS read during an open write transaction can yield a {path: null} row. Raised,\n // not filtered: filtering returns a silently short result (PRINCIPLES: no-silent-modes).\n if (rows.some((r) => r.path === null)) {\n throw new SenseError('LEXICAL_NULL_PATH', 'store \"turso\" returned a search hit with no path (a known engine anomaly under concurrent FTS reads); retry the query');\n }\n const matches = phrases.length === 0 ? rows : rows.filter((row) => phrases.every((phrase) => AUTHORED_FIELDS.some((field) => matchesSearchPhrase(row[field] ?? '', phrase))));\n return matches.slice(0, opts.limit).map(({ path }) => ({ path }));\n}\n"],"names":["SenseError","matchesSearchPhrase","searchTokens","stemFolded","FIELDS","NGRAM_FIELDS","AUTHORED_FIELDS","FTS5_OPERATORS","label","re","unsupportedOperator","terms","withoutPhrases","replace","m","match","token","trim","splitTerms","words","phrases","unspaced","emptyPhrase","_m","inner","phrase","tokens","length","push","filter","map","text","tok","split","escapeFtsLiteral","query","buildQueries","phraseParts","flatMap","latin","some","join","defaultParts","t","unspacedParts","defaultQuery","ngramQuery","ftsBranch","cols","opts","colList","c","lit","whereJoin","whereCond","scopeCond","queryLexical","conn","unsupported","branches","ranked","phraseSelect","phraseJoin","sql","stmt","prepare","rows","all","limit","r","path","matches","row","every","field","slice"],"mappings":"AAAA,4FAA4F;AAC5F,uGAAuG;AACvG,SAASA,UAAU,QAAQ,kBAAkB;AAC7C,SAASC,mBAAmB,EAAoBC,YAAY,QAAQ,wBAAwB;AAE5F,SAASC,UAAU,QAAQ,oBAAoB;AAE/C,MAAMC,SAAS;IAAC;IAAc;IAAgB;CAAY;AAC1D,MAAMC,eAAe;IAAC;IAAe;IAAiB;CAAa;AACnE,MAAMC,kBAAkB;IAAC;IAAS;IAAW;CAAO;AAGpD,uFAAuF;AACvF,+FAA+F;AAC/F,MAAMC,iBAAuD;IAC3D;QAAEC,OAAO;QAAgBC,IAAI;IAA4B;IACzD;QAAED,OAAO;QAAoBC,IAAI;IAAiC;IAClE;QAAED,OAAO;QAAiBC,IAAI;IAAiB;IAC/C;QAAED,OAAO;QAAgCC,IAAI;IAAW;IACxD;QAAED,OAAO;QAAiBC,IAAI;IAA2B;CAC1D;AAED,SAASC,oBAAoBC,KAAa;IACxC,MAAMC,iBAAiBD,MAAME,OAAO,CAAC,YAAY;IACjD,KAAK,MAAM,EAAEL,KAAK,EAAEC,EAAE,EAAE,IAAIF,eAAgB;QAC1C,MAAMO,IAAIF,eAAeG,KAAK,CAACN;QAC/B,IAAIK,GAAG,OAAO;YAAEN;YAAOQ,OAAOF,CAAC,CAAC,EAAE,CAACG,IAAI;QAAG;IAC5C;IACA,OAAO;AACT;AAEA,gGAAgG;AAChG,sFAAsF;AACtF,SAASC,WAAWP,KAAa;IAC/B,MAAMQ,QAAkB,EAAE;IAC1B,MAAMC,UAA2B,EAAE;IACnC,MAAMC,WAAqB,EAAE;IAC7B,IAAIC,cAAc;IAClB,MAAMV,iBAAiBD,MAAME,OAAO,CAAC,cAAc,CAACU,IAAIC;QACtD,MAAMC,SAASD,MAAMP,IAAI;QACzB,MAAMS,SAASxB,aAAauB;QAC5B,IAAIC,OAAOC,MAAM,KAAK,GAAG;YACvBL,cAAc;YACd,OAAO;QACT;QACAD,SAASO,IAAI,IAAIF,OAAOG,MAAM,CAAC,CAACb,QAAUA,MAAMK,QAAQ,EAAES,GAAG,CAAC,CAACd,QAAUA,MAAMe,IAAI;QACnFX,QAAQQ,IAAI,CAACF;QACb,OAAO;IACT;IACA,KAAK,MAAMM,OAAOpB,eAAeqB,KAAK,CAAC,OAAQ;QAC7C,IAAID,IAAIL,MAAM,KAAK,GAAG;QACtB,KAAK,MAAMX,SAASd,aAAa8B,KAAM;YACrC,IAAIhB,MAAMK,QAAQ,EAAEA,SAASO,IAAI,CAACZ,MAAMe,IAAI;iBACvCZ,MAAMS,IAAI,CAACZ,MAAMe,IAAI;QAC5B;IACF;IACA,OAAO;QAAEZ;QAAOC;QAASC;QAAUC;IAAY;AACjD;AAEA,8FAA8F;AAC9F,wFAAwF;AACxF,SAASY,iBAAiBC,KAAa;IACrC,OAAOA,MAAMtB,OAAO,CAAC,MAAM;AAC7B;AAEA,6FAA6F;AAC7F,kGAAkG;AAClG,SAASuB,aAAajB,KAAe,EAAEC,OAAwB,EAAEC,QAAkB;IACjF,MAAMgB,cAAcjB,QAAQkB,OAAO,CAAC,CAACb;QACnC,MAAMc,QAAQd,OAAOI,MAAM,CAAC,CAACb,QAAU,CAACA,MAAMK,QAAQ,EAAES,GAAG,CAAC,CAACd,QAAUb,WAAWa,MAAMe,IAAI;QAC5F,IAAIQ,MAAMZ,MAAM,KAAK,GAAG,OAAO,EAAE;QACjC,OAAOF,OAAOe,IAAI,CAAC,CAACxB,QAAUA,MAAMK,QAAQ,IAAIkB,QAAQ;YAACA,MAAME,IAAI,CAAC;SAAK;IAC3E;IACA,MAAMC,eAAe;WAAIvB,MAAMW,GAAG,CAAC3B;WAAgBkC;KAAY,CAACP,GAAG,CAAC,CAACa,IAAM,CAAC,CAAC,EAAEA,EAAE,CAAC,CAAC;IACnF,MAAMC,gBAAgBvB,SAASS,GAAG,CAAC,CAACa,IAAM,CAAC,CAAC,EAAEA,EAAE,CAAC,CAAC;IAClD,OAAO;QAAEE,cAAcH,aAAaD,IAAI,CAAC;QAAUK,YAAYF,cAAcH,IAAI,CAAC;IAAS;AAC7F;AAEA,gGAAgG;AAChG,+FAA+F;AAC/F,SAASM,UAAUC,IAAuB,EAAEb,KAAa,EAAEc,IAAyB;IAClF,MAAMC,UAAUF,KAAKlB,GAAG,CAAC,CAACqB,IAAM,CAAC,QAAQ,EAAEA,GAAG,EAAEV,IAAI,CAAC;IACrD,MAAMW,MAAMlB,iBAAiBC;IAC7B,OAAO,CAAC,yCAAyC,EAAEe,QAAQ,GAAG,EAAEE,IAAI;iBACrD,EAAEH,KAAKI,SAAS,CAAC,iBAAiB,EAAEH,QAAQ,GAAG,EAAEE,IAAI,GAAG,EAAEH,KAAKK,SAAS,CAAC,CAAC,EAAEL,KAAKM,SAAS,EAAE;AAC7G;AAEA,uFAAuF;AACvF,qMAAqM;AACrM,OAAO,eAAeC,aAAaC,IAAgB,EAAE9C,KAAa,EAAEsC,IAAyB;IAC3F,MAAMS,cAAchD,oBAAoBC;IACxC,IAAI+C,gBAAgB,MAAM;QACxB,MAAM,IAAI1D,WAAW,4BAA4B,CAAC,wCAAwC,EAAE0D,YAAYlD,KAAK,CAAC,GAAG,EAAEkD,YAAY1C,KAAK,CAAC,4BAA4B,EAAEL,MAAMM,IAAI,GAAG,sFAAsF,CAAC;IACzQ;IACA,MAAM,EAAEE,KAAK,EAAEC,OAAO,EAAEC,QAAQ,EAAEC,WAAW,EAAE,GAAGJ,WAAWP;IAC7D,IAAIW,aAAa,OAAO,EAAE;IAC1B,MAAM,EAAEuB,YAAY,EAAEC,UAAU,EAAE,GAAGV,aAAajB,OAAOC,SAASC;IAClE,IAAIwB,iBAAiB,MAAMC,eAAe,IAAI,OAAO,EAAE;IAEvD,MAAMa,WAAqB,EAAE;IAC7B,IAAId,iBAAiB,IAAIc,SAAS/B,IAAI,CAACmB,UAAU3C,QAAQyC,cAAcI;IACvE,IAAIH,eAAe,IAAIa,SAAS/B,IAAI,CAACmB,UAAU1C,cAAcyC,YAAYG;IAEzE,8FAA8F;IAC9F,0FAA0F;IAC1F,4FAA4F;IAC5F,uFAAuF;IACvF,MAAMW,SAAS,CAAC;MACZ,EAAED,SAASlB,IAAI,CAAC,eAAe;4CACO,EAAEkB,SAAShC,MAAM,EAAE;IAC7D,MAAMkC,eAAezC,QAAQO,MAAM,GAAG,IAAI,mDAAmD;IAC7F,MAAMmC,aAAa1C,QAAQO,MAAM,GAAG,IAAI,4CAA4C;IACpF,MAAMoC,MAAM,CAAC,wBAAwB,EAAEF,aAAa;IAClD,EAAED,OAAO;MACP,EAAEE,WAAW,gCAAgC,EAAE1C,QAAQO,MAAM,GAAG,IAAI,KAAK,YAAY;IACzF,MAAMqC,OAAO,MAAMP,KAAKQ,OAAO,CAACF;IAChC,MAAMG,OAAQ,MAAMF,KAAKG,GAAG,IAAK/C,QAAQO,MAAM,GAAG,IAAI,EAAE,GAAG;QAACsB,KAAKmB,KAAK;KAAC;IACvE,+FAA+F;IAC/F,yFAAyF;IACzF,IAAIF,KAAK1B,IAAI,CAAC,CAAC6B,IAAMA,EAAEC,IAAI,KAAK,OAAO;QACrC,MAAM,IAAItE,WAAW,qBAAqB;IAC5C;IACA,MAAMuE,UAAUnD,QAAQO,MAAM,KAAK,IAAIuC,OAAOA,KAAKrC,MAAM,CAAC,CAAC2C,MAAQpD,QAAQqD,KAAK,CAAC,CAAChD,SAAWnB,gBAAgBkC,IAAI,CAAC,CAACkC;oBAA8BF;uBAApBvE,qBAAoBuE,aAAAA,GAAG,CAACE,MAAM,cAAVF,wBAAAA,aAAc,IAAI/C;;IACnK,OAAO8C,QAAQI,KAAK,CAAC,GAAG1B,KAAKmB,KAAK,EAAEtC,GAAG,CAAC,CAAC,EAAEwC,IAAI,EAAE,GAAM,CAAA;YAAEA;QAAK,CAAA;AAChE"}
1
+ {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/lexical.ts"],"sourcesContent":["// T1: Tantivy FTS via fts_match/fts_score on the default-tokenizer index for bare words and\n// quoted phrases, plus a second ngram-tokenized index over \"_ngram\" sidecars for unspaced-script runs.\nimport { SenseError } from '../../errors.ts';\nimport { matchesSearchPhrase, type SearchToken, searchTokens } from '../../text/segment.ts';\nimport type { Connection, LexicalHit, LexicalQueryOptions } from '../types.ts';\nimport { stemFolded } from './lexical-text.ts';\n\nconst FIELDS = ['title_stem', 'summary_stem', 'text_stem'] as const;\nconst NGRAM_FIELDS = ['title_ngram', 'summary_ngram', 'text_ngram'] as const;\nconst AUTHORED_FIELDS = ['title', 'summary', 'text'] as const;\ntype AuthoredField = (typeof AUTHORED_FIELDS)[number];\n\n// FTS5 operator syntax a bare query would otherwise treat as literal text (PRINCIPLES:\n// no-silent-modes). `^` is rejected either way: boost to Tantivy, initial-token match to FTS5.\nconst FTS5_OPERATORS: Array<{ label: string; re: RegExp }> = [\n { label: 'prefix query', re: /[\\p{L}\\p{N}_]+\\*(?=\\s|$)/u },\n { label: 'boolean operator', re: /(?:^|\\s)(?:AND|OR|NOT)(?=\\s|$)/ },\n { label: 'NEAR operator', re: /(?:^|\\s)NEAR\\b/ },\n { label: 'boost/initial-token operator', re: /\\S*\\^\\S*/ },\n { label: 'column filter', re: /(?:^|\\s)[\\p{L}_]\\w*\\s*:/u },\n { label: 'grouping operator', re: /[()]/ },\n];\n\nfunction unsupportedOperator(terms: string): { label: string; token: string } | null {\n const withoutPhrases = terms.replace(/\"[^\"]*\"/g, 'phrase');\n for (const { label, re } of FTS5_OPERATORS) {\n const m = withoutPhrases.match(re);\n if (m) return { label, token: m[0].trim() };\n }\n return null;\n}\n\n// A run whose script marks no word boundaries indexes as one opaque token under every tokenizer\n// but ngram, so such runs -- bare or quoted -- go to the ngram sidecar index instead.\nfunction splitTerms(terms: string): { words: string[]; phrases: SearchToken[][]; unspaced: string[]; emptyPhrase: boolean } {\n const words: string[] = [];\n const phrases: SearchToken[][] = [];\n const unspaced: string[] = [];\n let emptyPhrase = false;\n const withoutPhrases = terms.replace(/\"([^\"]*)\"/g, (_m, inner: string) => {\n const phrase = inner.trim();\n const tokens = searchTokens(phrase);\n if (tokens.length === 0) {\n emptyPhrase = true;\n return ' ';\n }\n unspaced.push(...tokens.filter((token) => token.unspaced).map((token) => token.text));\n phrases.push(tokens);\n return ' ';\n });\n for (const tok of withoutPhrases.split(/\\s+/)) {\n if (tok.length === 0) continue;\n for (const token of searchTokens(tok)) {\n if (token.unspaced) unspaced.push(token.text);\n else words.push(token.text);\n }\n }\n return { words, phrases, unspaced, emptyPhrase };\n}\n\n// fts_match/fts_score's query argument must be a SQL literal: bound as `?`, the row set stays\n// correct but fts_score silently returns 0 for every row (PRINCIPLES: no-silent-modes).\nfunction escapeFtsLiteral(query: string): string {\n return query.replace(/'/g, \"''\");\n}\n\n// Terms are quoted (\"O'Brien's\" is a parse error bare) and AND-joined: Tantivy's default for\n// bare terms is disjunctive, so space-joining would loosen \"apple banana\" into \"apple OR banana\".\nfunction buildQueries(words: string[], phrases: SearchToken[][], unspaced: string[]): { defaultQuery: string; ngramQuery: string } {\n const phraseParts = phrases.flatMap((phrase) => {\n const latin = phrase.filter((token) => !token.unspaced).map((token) => stemFolded(token.text));\n if (latin.length === 0) return [];\n return phrase.some((token) => token.unspaced) ? latin : [latin.join(' ')];\n });\n const defaultParts = [...words.map(stemFolded), ...phraseParts].map((t) => `\"${t}\"`);\n const unspacedParts = unspaced.map((t) => `\"${t}\"`);\n return { defaultQuery: defaultParts.join(' AND '), ngramQuery: unspacedParts.join(' AND ') };\n}\n\n// fts_score() is real only as a SELECT's bare projected expression, and only for the index that\n// SELECT's fts_match() chose; every other form silently returns 0. Hence one SELECT per index.\nfunction ftsBranch(cols: readonly string[], query: string, opts: LexicalQueryOptions): string {\n const colList = cols.map((c) => `content.${c}`).join(', ');\n const lit = escapeFtsLiteral(query);\n return `SELECT content.\"path\" AS path, fts_score(${colList}, '${lit}') AS score\n FROM content ${opts.whereJoin} WHERE fts_match(${colList}, '${lit}') ${opts.whereCond} ${opts.scopeCond}`;\n}\n\n// Ranked query, scoped by the caller-built SQL fragments (same shape sqlite's/duckdb's\n// queryLexical take). `snippets` is always NULL: fts_highlight returns the full column text, not a bounded snippet, so every row goes through the caller's JS snippet fallback (commands/search.ts).\nexport async function queryLexical(conn: Connection, terms: string, opts: LexicalQueryOptions): Promise<LexicalHit[]> {\n const unsupported = unsupportedOperator(terms);\n if (unsupported !== null) {\n throw new SenseError('STORE_CAPABILITY_MISSING', `store \"turso\" does not implement FTS5's ${unsupported.label} (\"${unsupported.token}\") in this build; rephrase \"${terms.trim()}\" without it, or set \"store\" to \"sqlite\" in this tree's config to search it as written`);\n }\n const { words, phrases, unspaced, emptyPhrase } = splitTerms(terms);\n if (emptyPhrase) return [];\n const { defaultQuery, ngramQuery } = buildQueries(words, phrases, unspaced);\n if (defaultQuery === '' && ngramQuery === '') return [];\n\n const branches: string[] = [];\n if (defaultQuery !== '') branches.push(ftsBranch(FIELDS, defaultQuery, opts));\n if (ngramQuery !== '') branches.push(ftsBranch(NGRAM_FIELDS, ngramQuery, opts));\n\n // HAVING COUNT(*) = branches.length is the AND across indexes; with one branch it is a no-op,\n // `path` being content's primary key. fts_score is higher-is-better BM25, so DESC. Phrase\n // candidates carry authored fields through the join below so the shared verifier can reject\n // cross-field, reordered, and non-adjacent matches before applying the caller's limit.\n const ranked = `SELECT path, SUM(score) AS score FROM (\n ${branches.join(' UNION ALL ')}\n ) parts GROUP BY path HAVING COUNT(*) = ${branches.length}`;\n const phraseSelect = phrases.length > 0 ? ', content.title, content.summary, content.text' : '';\n const phraseJoin = phrases.length > 0 ? ' JOIN content ON content.path = sq.path' : '';\n const sql = `SELECT sq.path, sq.score${phraseSelect} FROM (\n ${ranked}\n ) sq${phraseJoin} ORDER BY sq.score DESC, sq.path${phrases.length > 0 ? '' : ' LIMIT ?'}`;\n const stmt = await conn.prepare(sql);\n const rows = (await stmt.all(...(phrases.length > 0 ? [] : [opts.limit]))) as Array<LexicalHit & Partial<Record<AuthoredField, string>>>;\n // A concurrent FTS read during an open write transaction can yield a {path: null} row. Raised,\n // not filtered: filtering returns a silently short result (PRINCIPLES: no-silent-modes).\n if (rows.some((r) => r.path === null)) {\n throw new SenseError('LEXICAL_NULL_PATH', 'store \"turso\" returned a search hit with no path (a known engine anomaly under concurrent FTS reads); retry the query');\n }\n const matches = phrases.length === 0 ? rows : rows.filter((row) => phrases.every((phrase) => AUTHORED_FIELDS.some((field) => matchesSearchPhrase(row[field] ?? '', phrase))));\n return matches.slice(0, opts.limit).map(({ path }) => ({ path }));\n}\n"],"names":["SenseError","matchesSearchPhrase","searchTokens","stemFolded","FIELDS","NGRAM_FIELDS","AUTHORED_FIELDS","FTS5_OPERATORS","label","re","unsupportedOperator","terms","withoutPhrases","replace","m","match","token","trim","splitTerms","words","phrases","unspaced","emptyPhrase","_m","inner","phrase","tokens","length","push","filter","map","text","tok","split","escapeFtsLiteral","query","buildQueries","phraseParts","flatMap","latin","some","join","defaultParts","t","unspacedParts","defaultQuery","ngramQuery","ftsBranch","cols","opts","colList","c","lit","whereJoin","whereCond","scopeCond","queryLexical","conn","unsupported","branches","ranked","phraseSelect","phraseJoin","sql","stmt","prepare","rows","all","limit","r","path","matches","row","every","field","slice"],"mappings":"AAAA,4FAA4F;AAC5F,uGAAuG;AACvG,SAASA,UAAU,QAAQ,kBAAkB;AAC7C,SAASC,mBAAmB,EAAoBC,YAAY,QAAQ,wBAAwB;AAE5F,SAASC,UAAU,QAAQ,oBAAoB;AAE/C,MAAMC,SAAS;IAAC;IAAc;IAAgB;CAAY;AAC1D,MAAMC,eAAe;IAAC;IAAe;IAAiB;CAAa;AACnE,MAAMC,kBAAkB;IAAC;IAAS;IAAW;CAAO;AAGpD,uFAAuF;AACvF,+FAA+F;AAC/F,MAAMC,iBAAuD;IAC3D;QAAEC,OAAO;QAAgBC,IAAI;IAA4B;IACzD;QAAED,OAAO;QAAoBC,IAAI;IAAiC;IAClE;QAAED,OAAO;QAAiBC,IAAI;IAAiB;IAC/C;QAAED,OAAO;QAAgCC,IAAI;IAAW;IACxD;QAAED,OAAO;QAAiBC,IAAI;IAA2B;IACzD;QAAED,OAAO;QAAqBC,IAAI;IAAO;CAC1C;AAED,SAASC,oBAAoBC,KAAa;IACxC,MAAMC,iBAAiBD,MAAME,OAAO,CAAC,YAAY;IACjD,KAAK,MAAM,EAAEL,KAAK,EAAEC,EAAE,EAAE,IAAIF,eAAgB;QAC1C,MAAMO,IAAIF,eAAeG,KAAK,CAACN;QAC/B,IAAIK,GAAG,OAAO;YAAEN;YAAOQ,OAAOF,CAAC,CAAC,EAAE,CAACG,IAAI;QAAG;IAC5C;IACA,OAAO;AACT;AAEA,gGAAgG;AAChG,sFAAsF;AACtF,SAASC,WAAWP,KAAa;IAC/B,MAAMQ,QAAkB,EAAE;IAC1B,MAAMC,UAA2B,EAAE;IACnC,MAAMC,WAAqB,EAAE;IAC7B,IAAIC,cAAc;IAClB,MAAMV,iBAAiBD,MAAME,OAAO,CAAC,cAAc,CAACU,IAAIC;QACtD,MAAMC,SAASD,MAAMP,IAAI;QACzB,MAAMS,SAASxB,aAAauB;QAC5B,IAAIC,OAAOC,MAAM,KAAK,GAAG;YACvBL,cAAc;YACd,OAAO;QACT;QACAD,SAASO,IAAI,IAAIF,OAAOG,MAAM,CAAC,CAACb,QAAUA,MAAMK,QAAQ,EAAES,GAAG,CAAC,CAACd,QAAUA,MAAMe,IAAI;QACnFX,QAAQQ,IAAI,CAACF;QACb,OAAO;IACT;IACA,KAAK,MAAMM,OAAOpB,eAAeqB,KAAK,CAAC,OAAQ;QAC7C,IAAID,IAAIL,MAAM,KAAK,GAAG;QACtB,KAAK,MAAMX,SAASd,aAAa8B,KAAM;YACrC,IAAIhB,MAAMK,QAAQ,EAAEA,SAASO,IAAI,CAACZ,MAAMe,IAAI;iBACvCZ,MAAMS,IAAI,CAACZ,MAAMe,IAAI;QAC5B;IACF;IACA,OAAO;QAAEZ;QAAOC;QAASC;QAAUC;IAAY;AACjD;AAEA,8FAA8F;AAC9F,wFAAwF;AACxF,SAASY,iBAAiBC,KAAa;IACrC,OAAOA,MAAMtB,OAAO,CAAC,MAAM;AAC7B;AAEA,6FAA6F;AAC7F,kGAAkG;AAClG,SAASuB,aAAajB,KAAe,EAAEC,OAAwB,EAAEC,QAAkB;IACjF,MAAMgB,cAAcjB,QAAQkB,OAAO,CAAC,CAACb;QACnC,MAAMc,QAAQd,OAAOI,MAAM,CAAC,CAACb,QAAU,CAACA,MAAMK,QAAQ,EAAES,GAAG,CAAC,CAACd,QAAUb,WAAWa,MAAMe,IAAI;QAC5F,IAAIQ,MAAMZ,MAAM,KAAK,GAAG,OAAO,EAAE;QACjC,OAAOF,OAAOe,IAAI,CAAC,CAACxB,QAAUA,MAAMK,QAAQ,IAAIkB,QAAQ;YAACA,MAAME,IAAI,CAAC;SAAK;IAC3E;IACA,MAAMC,eAAe;WAAIvB,MAAMW,GAAG,CAAC3B;WAAgBkC;KAAY,CAACP,GAAG,CAAC,CAACa,IAAM,CAAC,CAAC,EAAEA,EAAE,CAAC,CAAC;IACnF,MAAMC,gBAAgBvB,SAASS,GAAG,CAAC,CAACa,IAAM,CAAC,CAAC,EAAEA,EAAE,CAAC,CAAC;IAClD,OAAO;QAAEE,cAAcH,aAAaD,IAAI,CAAC;QAAUK,YAAYF,cAAcH,IAAI,CAAC;IAAS;AAC7F;AAEA,gGAAgG;AAChG,+FAA+F;AAC/F,SAASM,UAAUC,IAAuB,EAAEb,KAAa,EAAEc,IAAyB;IAClF,MAAMC,UAAUF,KAAKlB,GAAG,CAAC,CAACqB,IAAM,CAAC,QAAQ,EAAEA,GAAG,EAAEV,IAAI,CAAC;IACrD,MAAMW,MAAMlB,iBAAiBC;IAC7B,OAAO,CAAC,yCAAyC,EAAEe,QAAQ,GAAG,EAAEE,IAAI;iBACrD,EAAEH,KAAKI,SAAS,CAAC,iBAAiB,EAAEH,QAAQ,GAAG,EAAEE,IAAI,GAAG,EAAEH,KAAKK,SAAS,CAAC,CAAC,EAAEL,KAAKM,SAAS,EAAE;AAC7G;AAEA,uFAAuF;AACvF,qMAAqM;AACrM,OAAO,eAAeC,aAAaC,IAAgB,EAAE9C,KAAa,EAAEsC,IAAyB;IAC3F,MAAMS,cAAchD,oBAAoBC;IACxC,IAAI+C,gBAAgB,MAAM;QACxB,MAAM,IAAI1D,WAAW,4BAA4B,CAAC,wCAAwC,EAAE0D,YAAYlD,KAAK,CAAC,GAAG,EAAEkD,YAAY1C,KAAK,CAAC,4BAA4B,EAAEL,MAAMM,IAAI,GAAG,sFAAsF,CAAC;IACzQ;IACA,MAAM,EAAEE,KAAK,EAAEC,OAAO,EAAEC,QAAQ,EAAEC,WAAW,EAAE,GAAGJ,WAAWP;IAC7D,IAAIW,aAAa,OAAO,EAAE;IAC1B,MAAM,EAAEuB,YAAY,EAAEC,UAAU,EAAE,GAAGV,aAAajB,OAAOC,SAASC;IAClE,IAAIwB,iBAAiB,MAAMC,eAAe,IAAI,OAAO,EAAE;IAEvD,MAAMa,WAAqB,EAAE;IAC7B,IAAId,iBAAiB,IAAIc,SAAS/B,IAAI,CAACmB,UAAU3C,QAAQyC,cAAcI;IACvE,IAAIH,eAAe,IAAIa,SAAS/B,IAAI,CAACmB,UAAU1C,cAAcyC,YAAYG;IAEzE,8FAA8F;IAC9F,0FAA0F;IAC1F,4FAA4F;IAC5F,uFAAuF;IACvF,MAAMW,SAAS,CAAC;MACZ,EAAED,SAASlB,IAAI,CAAC,eAAe;4CACO,EAAEkB,SAAShC,MAAM,EAAE;IAC7D,MAAMkC,eAAezC,QAAQO,MAAM,GAAG,IAAI,mDAAmD;IAC7F,MAAMmC,aAAa1C,QAAQO,MAAM,GAAG,IAAI,4CAA4C;IACpF,MAAMoC,MAAM,CAAC,wBAAwB,EAAEF,aAAa;IAClD,EAAED,OAAO;MACP,EAAEE,WAAW,gCAAgC,EAAE1C,QAAQO,MAAM,GAAG,IAAI,KAAK,YAAY;IACzF,MAAMqC,OAAO,MAAMP,KAAKQ,OAAO,CAACF;IAChC,MAAMG,OAAQ,MAAMF,KAAKG,GAAG,IAAK/C,QAAQO,MAAM,GAAG,IAAI,EAAE,GAAG;QAACsB,KAAKmB,KAAK;KAAC;IACvE,+FAA+F;IAC/F,yFAAyF;IACzF,IAAIF,KAAK1B,IAAI,CAAC,CAAC6B,IAAMA,EAAEC,IAAI,KAAK,OAAO;QACrC,MAAM,IAAItE,WAAW,qBAAqB;IAC5C;IACA,MAAMuE,UAAUnD,QAAQO,MAAM,KAAK,IAAIuC,OAAOA,KAAKrC,MAAM,CAAC,CAAC2C,MAAQpD,QAAQqD,KAAK,CAAC,CAAChD,SAAWnB,gBAAgBkC,IAAI,CAAC,CAACkC;oBAA8BF;uBAApBvE,qBAAoBuE,aAAAA,GAAG,CAACE,MAAM,cAAVF,wBAAAA,aAAc,IAAI/C;;IACnK,OAAO8C,QAAQI,KAAK,CAAC,GAAG1B,KAAKmB,KAAK,EAAEtC,GAAG,CAAC,CAAC,EAAEwC,IAAI,EAAE,GAAM,CAAA;YAAEA;QAAK,CAAA;AAChE"}
@@ -25,8 +25,8 @@ function ngramSidecar(text) {
25
25
  // Not a compile-time cap on ALTER TABLE ADD COLUMN (spike-measured: turso accepts 10,000 with
26
26
  // no error); the real fence is a SELECT projecting more than this many result columns, which fails to prepare.
27
27
  const MAX_FRONTMATTER_COLUMNS = 2000;
28
- // Changed files above which rebuilding the FTS index beats maintaining it per row. Measured
29
- // 2026-08-30; the derivation and its bounds are pinned in this file's spec.
28
+ // Retained policy boundary from 2026-08-30 corpus-scale measurements; the optimum after later
29
+ // update-path changes remains pending matched corpus-scale crossover evidence.
30
30
  export const FTS_REBUILD_THRESHOLD = 250;
31
31
  // No rowid coupling (unlike sqlite's FTS5 content): `path` is content's own primary key.
32
32
  const INSERT_CONTENT_SQL = `INSERT INTO content ("path", title, summary, text, title_stem, summary_stem, text_stem, title_ngram, summary_ngram, text_ngram) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`;
@@ -58,9 +58,12 @@ export async function reconcileTursoContentWithStrategy(conn, touched, docs, str
58
58
  if (bulk) for (const name of CONTENT_FTS_NAMES)await conn.exec(`DROP INDEX IF EXISTS ${name}`);
59
59
  // content is a plain table keyed by its own path (no rowid subquery, unlike sqlite's FTS5
60
60
  // content), so vanished and reparsed docs delete in one pass.
61
- if (touched.length > 0) await conn.runBatch('DELETE FROM content WHERE "path" = ?', touched.map((p)=>[
62
- p
63
- ]));
61
+ if (touched.length > 0) {
62
+ const placeholders = touched.map(()=>'?').join(', ');
63
+ await conn.runBatch(`DELETE FROM content WHERE "path" IN (${placeholders})`, [
64
+ touched
65
+ ]);
66
+ }
64
67
  if (docs.length > 0) await conn.runBatch(INSERT_CONTENT_SQL, docs.map(contentRow));
65
68
  if (bulk) for (const ddl of CONTENT_FTS_DDL)await conn.exec(ddl);
66
69
  }
@@ -1 +1 @@
1
- {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/reconcile.ts"],"sourcesContent":["import type { Config } from '../../config/index.ts';\nimport { SenseError } from '../../errors.ts';\nimport type { ReconcileDelta } from '../../features/types.ts';\nimport type { ParsedDoc } from '../../scan/index.ts';\nimport { hasUnspacedRun } from '../../text/segment.ts';\nimport { quoteIdent, recordReconcileDuration } from '../shared.ts';\nimport { BEGIN_WRITE } from '../transaction.ts';\nimport type { Connection, ReconcileDialect } from '../types.ts';\nimport { stemFolded } from './lexical-text.ts';\n\n// This store's dialect (types.ts's ReconcileDialect) for the shared orchestration in\n// store/reconcile.ts. content is a plain table with \"_ngram\" sidecars for lexical.ts.\n\n// The ngram index is scoped to disjoint \"_ngram\" sidecar columns: a second index over the same\n// columns makes a bare substring match a whole word, defeating the prefix-query rejection.\n// The two FTS indexes, named here (not open.ts) because reconcile drops and rebuilds them around\n// a bulk load: Tantivy maintains them per inserted row, which is quadratic in what is already indexed.\nexport const CONTENT_FTS_DDL = [\n `CREATE INDEX IF NOT EXISTS content_fts ON content USING fts (title_stem, summary_stem, text_stem) WITH (weights = 'title_stem=10.0,summary_stem=5.0,text_stem=1.0')`,\n `CREATE INDEX IF NOT EXISTS content_fts_ngram ON content USING fts (title_ngram, summary_ngram, text_ngram) WITH (tokenizer='ngram', weights='title_ngram=10.0,summary_ngram=5.0,text_ngram=1.0')`,\n] as const;\nexport const CONTENT_FTS_NAMES = ['content_fts', 'content_fts_ngram'] as const;\n\n// '' when the field has no unspaced-script run (the common case), so the ngram index carries\n// nothing for it -- same \"pay for nothing when absent\" shape as sqlite's segmentField sidecars.\nfunction ngramSidecar(text: string): string {\n return hasUnspacedRun(text) ? text : '';\n}\n\n// Not a compile-time cap on ALTER TABLE ADD COLUMN (spike-measured: turso accepts 10,000 with\n// no error); the real fence is a SELECT projecting more than this many result columns, which fails to prepare.\nconst MAX_FRONTMATTER_COLUMNS = 2000;\n\n// Changed files above which rebuilding the FTS index beats maintaining it per row. Measured\n// 2026-08-30; the derivation and its bounds are pinned in this file's spec.\nexport const FTS_REBUILD_THRESHOLD = 250;\n\nexport type TursoFtsStrategy = 'incremental' | 'rebuild';\n\n// No rowid coupling (unlike sqlite's FTS5 content): `path` is content's own primary key.\nconst INSERT_CONTENT_SQL = `INSERT INTO content (\"path\", title, summary, text, title_stem, summary_stem, text_stem, title_ngram, summary_ngram, text_ngram) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`;\n\nfunction contentRow(doc: ParsedDoc): unknown[] {\n const { title, summary, text } = doc.search;\n return [doc.relPath, title, summary, text, stemFolded(title), stemFolded(summary), stemFolded(text), ngramSidecar(title), ngramSidecar(summary), ngramSidecar(text)];\n}\n\nexport function tursoFtsStrategy(delta: ReconcileDelta): TursoFtsStrategy {\n const churn = delta.reparsed.length + delta.vanished.length;\n return delta.files.length === 0 || churn > FTS_REBUILD_THRESHOLD ? 'rebuild' : 'incremental';\n}\n\n// Kept transaction-free because the shared reconcile owns the one write transaction. The explicit\n// strategy is internal composition for real-engine diagnostics; production selects it from delta.\nexport async function reconcileTursoContentWithStrategy(conn: Connection, touched: string[], docs: ParsedDoc[], strategy: TursoFtsStrategy): Promise<void> {\n // Tantivy indexes per inserted row at a cost that grows with the batch, so a large insert is\n // superlinear and a rebuild wins past the threshold.\n const bulk = strategy === 'rebuild';\n if (bulk) for (const name of CONTENT_FTS_NAMES) await conn.exec(`DROP INDEX IF EXISTS ${name}`);\n\n // content is a plain table keyed by its own path (no rowid subquery, unlike sqlite's FTS5\n // content), so vanished and reparsed docs delete in one pass.\n if (touched.length > 0)\n await conn.runBatch(\n 'DELETE FROM content WHERE \"path\" = ?',\n touched.map((p) => [p])\n );\n if (docs.length > 0) await conn.runBatch(INSERT_CONTENT_SQL, docs.map(contentRow));\n if (bulk) for (const ddl of CONTENT_FTS_DDL) await conn.exec(ddl);\n}\n\nasync function reconcileTursoContent(conn: Connection, touched: string[], docs: ParsedDoc[], delta: ReconcileDelta, _cfg: Config): Promise<void> {\n await reconcileTursoContentWithStrategy(conn, touched, docs, tursoFtsStrategy(delta));\n}\n\n// turso's ADD COLUMN is metadata-only, so a loop costs nothing extra over one statement.\nasync function addColumns(conn: Connection, names: string[]): Promise<void> {\n for (const name of names) await conn.exec(`ALTER TABLE frontmatter ADD COLUMN ${quoteIdent(name)}`);\n}\n\nexport const tursoDialect: ReconcileDialect = {\n beginMode: () => BEGIN_WRITE,\n checkColumnLimit(count) {\n if (count > MAX_FRONTMATTER_COLUMNS) {\n throw new SenseError(\n 'COLUMN_LIMIT',\n `frontmatter would need ${count} columns, crossing turso's SELECT result-set column limit (${MAX_FRONTMATTER_COLUMNS}; ALTER TABLE ADD COLUMN itself accepts far more, but a query projecting past this many columns fails to prepare). Narrow the presets' include globs so fewer/other files are indexed, or fix whatever is generating unbounded frontmatter keys.`\n );\n }\n },\n addColumns,\n reconcileContent: reconcileTursoContent,\n recordDuration: recordReconcileDuration,\n};\n"],"names":["SenseError","hasUnspacedRun","quoteIdent","recordReconcileDuration","BEGIN_WRITE","stemFolded","CONTENT_FTS_DDL","CONTENT_FTS_NAMES","ngramSidecar","text","MAX_FRONTMATTER_COLUMNS","FTS_REBUILD_THRESHOLD","INSERT_CONTENT_SQL","contentRow","doc","title","summary","search","relPath","tursoFtsStrategy","delta","churn","reparsed","length","vanished","files","reconcileTursoContentWithStrategy","conn","touched","docs","strategy","bulk","name","exec","runBatch","map","p","ddl","reconcileTursoContent","_cfg","addColumns","names","tursoDialect","beginMode","checkColumnLimit","count","reconcileContent","recordDuration"],"mappings":"AACA,SAASA,UAAU,QAAQ,kBAAkB;AAG7C,SAASC,cAAc,QAAQ,wBAAwB;AACvD,SAASC,UAAU,EAAEC,uBAAuB,QAAQ,eAAe;AACnE,SAASC,WAAW,QAAQ,oBAAoB;AAEhD,SAASC,UAAU,QAAQ,oBAAoB;AAE/C,qFAAqF;AACrF,sFAAsF;AAEtF,+FAA+F;AAC/F,2FAA2F;AAC3F,iGAAiG;AACjG,uGAAuG;AACvG,OAAO,MAAMC,kBAAkB;IAC7B,CAAC,mKAAmK,CAAC;IACrK,CAAC,gMAAgM,CAAC;CACnM,CAAU;AACX,OAAO,MAAMC,oBAAoB;IAAC;IAAe;CAAoB,CAAU;AAE/E,6FAA6F;AAC7F,gGAAgG;AAChG,SAASC,aAAaC,IAAY;IAChC,OAAOR,eAAeQ,QAAQA,OAAO;AACvC;AAEA,8FAA8F;AAC9F,+GAA+G;AAC/G,MAAMC,0BAA0B;AAEhC,4FAA4F;AAC5F,4EAA4E;AAC5E,OAAO,MAAMC,wBAAwB,IAAI;AAIzC,yFAAyF;AACzF,MAAMC,qBAAqB,CAAC,qKAAqK,CAAC;AAElM,SAASC,WAAWC,GAAc;IAChC,MAAM,EAAEC,KAAK,EAAEC,OAAO,EAAEP,IAAI,EAAE,GAAGK,IAAIG,MAAM;IAC3C,OAAO;QAACH,IAAII,OAAO;QAAEH;QAAOC;QAASP;QAAMJ,WAAWU;QAAQV,WAAWW;QAAUX,WAAWI;QAAOD,aAAaO;QAAQP,aAAaQ;QAAUR,aAAaC;KAAM;AACtK;AAEA,OAAO,SAASU,iBAAiBC,KAAqB;IACpD,MAAMC,QAAQD,MAAME,QAAQ,CAACC,MAAM,GAAGH,MAAMI,QAAQ,CAACD,MAAM;IAC3D,OAAOH,MAAMK,KAAK,CAACF,MAAM,KAAK,KAAKF,QAAQV,wBAAwB,YAAY;AACjF;AAEA,kGAAkG;AAClG,kGAAkG;AAClG,OAAO,eAAee,kCAAkCC,IAAgB,EAAEC,OAAiB,EAAEC,IAAiB,EAAEC,QAA0B;IACxI,6FAA6F;IAC7F,qDAAqD;IACrD,MAAMC,OAAOD,aAAa;IAC1B,IAAIC,MAAM,KAAK,MAAMC,QAAQzB,kBAAmB,MAAMoB,KAAKM,IAAI,CAAC,CAAC,qBAAqB,EAAED,MAAM;IAE9F,0FAA0F;IAC1F,8DAA8D;IAC9D,IAAIJ,QAAQL,MAAM,GAAG,GACnB,MAAMI,KAAKO,QAAQ,CACjB,wCACAN,QAAQO,GAAG,CAAC,CAACC,IAAM;YAACA;SAAE;IAE1B,IAAIP,KAAKN,MAAM,GAAG,GAAG,MAAMI,KAAKO,QAAQ,CAACtB,oBAAoBiB,KAAKM,GAAG,CAACtB;IACtE,IAAIkB,MAAM,KAAK,MAAMM,OAAO/B,gBAAiB,MAAMqB,KAAKM,IAAI,CAACI;AAC/D;AAEA,eAAeC,sBAAsBX,IAAgB,EAAEC,OAAiB,EAAEC,IAAiB,EAAET,KAAqB,EAAEmB,IAAY;IAC9H,MAAMb,kCAAkCC,MAAMC,SAASC,MAAMV,iBAAiBC;AAChF;AAEA,yFAAyF;AACzF,eAAeoB,WAAWb,IAAgB,EAAEc,KAAe;IACzD,KAAK,MAAMT,QAAQS,MAAO,MAAMd,KAAKM,IAAI,CAAC,CAAC,mCAAmC,EAAE/B,WAAW8B,OAAO;AACpG;AAEA,OAAO,MAAMU,eAAiC;IAC5CC,WAAW,IAAMvC;IACjBwC,kBAAiBC,KAAK;QACpB,IAAIA,QAAQnC,yBAAyB;YACnC,MAAM,IAAIV,WACR,gBACA,CAAC,uBAAuB,EAAE6C,MAAM,2DAA2D,EAAEnC,wBAAwB,gPAAgP,CAAC;QAE1W;IACF;IACA8B;IACAM,kBAAkBR;IAClBS,gBAAgB5C;AAClB,EAAE"}
1
+ {"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/store/turso/reconcile.ts"],"sourcesContent":["import type { Config } from '../../config/index.ts';\nimport { SenseError } from '../../errors.ts';\nimport type { ReconcileDelta } from '../../features/types.ts';\nimport type { ParsedDoc } from '../../scan/index.ts';\nimport { hasUnspacedRun } from '../../text/segment.ts';\nimport { quoteIdent, recordReconcileDuration } from '../shared.ts';\nimport { BEGIN_WRITE } from '../transaction.ts';\nimport type { Connection, ReconcileDialect } from '../types.ts';\nimport { stemFolded } from './lexical-text.ts';\n\n// This store's dialect (types.ts's ReconcileDialect) for the shared orchestration in\n// store/reconcile.ts. content is a plain table with \"_ngram\" sidecars for lexical.ts.\n\n// The ngram index is scoped to disjoint \"_ngram\" sidecar columns: a second index over the same\n// columns makes a bare substring match a whole word, defeating the prefix-query rejection.\n// The two FTS indexes, named here (not open.ts) because reconcile drops and rebuilds them around\n// a bulk load: Tantivy maintains them per inserted row, which is quadratic in what is already indexed.\nexport const CONTENT_FTS_DDL = [\n `CREATE INDEX IF NOT EXISTS content_fts ON content USING fts (title_stem, summary_stem, text_stem) WITH (weights = 'title_stem=10.0,summary_stem=5.0,text_stem=1.0')`,\n `CREATE INDEX IF NOT EXISTS content_fts_ngram ON content USING fts (title_ngram, summary_ngram, text_ngram) WITH (tokenizer='ngram', weights='title_ngram=10.0,summary_ngram=5.0,text_ngram=1.0')`,\n] as const;\nexport const CONTENT_FTS_NAMES = ['content_fts', 'content_fts_ngram'] as const;\n\n// '' when the field has no unspaced-script run (the common case), so the ngram index carries\n// nothing for it -- same \"pay for nothing when absent\" shape as sqlite's segmentField sidecars.\nfunction ngramSidecar(text: string): string {\n return hasUnspacedRun(text) ? text : '';\n}\n\n// Not a compile-time cap on ALTER TABLE ADD COLUMN (spike-measured: turso accepts 10,000 with\n// no error); the real fence is a SELECT projecting more than this many result columns, which fails to prepare.\nconst MAX_FRONTMATTER_COLUMNS = 2000;\n\n// Retained policy boundary from 2026-08-30 corpus-scale measurements; the optimum after later\n// update-path changes remains pending matched corpus-scale crossover evidence.\nexport const FTS_REBUILD_THRESHOLD = 250;\n\nexport type TursoFtsStrategy = 'incremental' | 'rebuild';\n\n// No rowid coupling (unlike sqlite's FTS5 content): `path` is content's own primary key.\nconst INSERT_CONTENT_SQL = `INSERT INTO content (\"path\", title, summary, text, title_stem, summary_stem, text_stem, title_ngram, summary_ngram, text_ngram) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`;\n\nfunction contentRow(doc: ParsedDoc): unknown[] {\n const { title, summary, text } = doc.search;\n return [doc.relPath, title, summary, text, stemFolded(title), stemFolded(summary), stemFolded(text), ngramSidecar(title), ngramSidecar(summary), ngramSidecar(text)];\n}\n\nexport function tursoFtsStrategy(delta: ReconcileDelta): TursoFtsStrategy {\n const churn = delta.reparsed.length + delta.vanished.length;\n return delta.files.length === 0 || churn > FTS_REBUILD_THRESHOLD ? 'rebuild' : 'incremental';\n}\n\n// Kept transaction-free because the shared reconcile owns the one write transaction. The explicit\n// strategy is internal composition for real-engine diagnostics; production selects it from delta.\nexport async function reconcileTursoContentWithStrategy(conn: Connection, touched: string[], docs: ParsedDoc[], strategy: TursoFtsStrategy): Promise<void> {\n // Tantivy indexes per inserted row at a cost that grows with the batch, so a large insert is\n // superlinear and a rebuild wins past the threshold.\n const bulk = strategy === 'rebuild';\n if (bulk) for (const name of CONTENT_FTS_NAMES) await conn.exec(`DROP INDEX IF EXISTS ${name}`);\n\n // content is a plain table keyed by its own path (no rowid subquery, unlike sqlite's FTS5\n // content), so vanished and reparsed docs delete in one pass.\n if (touched.length > 0) {\n const placeholders = touched.map(() => '?').join(', ');\n await conn.runBatch(`DELETE FROM content WHERE \"path\" IN (${placeholders})`, [touched]);\n }\n if (docs.length > 0) await conn.runBatch(INSERT_CONTENT_SQL, docs.map(contentRow));\n if (bulk) for (const ddl of CONTENT_FTS_DDL) await conn.exec(ddl);\n}\n\nasync function reconcileTursoContent(conn: Connection, touched: string[], docs: ParsedDoc[], delta: ReconcileDelta, _cfg: Config): Promise<void> {\n await reconcileTursoContentWithStrategy(conn, touched, docs, tursoFtsStrategy(delta));\n}\n\n// turso's ADD COLUMN is metadata-only, so a loop costs nothing extra over one statement.\nasync function addColumns(conn: Connection, names: string[]): Promise<void> {\n for (const name of names) await conn.exec(`ALTER TABLE frontmatter ADD COLUMN ${quoteIdent(name)}`);\n}\n\nexport const tursoDialect: ReconcileDialect = {\n beginMode: () => BEGIN_WRITE,\n checkColumnLimit(count) {\n if (count > MAX_FRONTMATTER_COLUMNS) {\n throw new SenseError(\n 'COLUMN_LIMIT',\n `frontmatter would need ${count} columns, crossing turso's SELECT result-set column limit (${MAX_FRONTMATTER_COLUMNS}; ALTER TABLE ADD COLUMN itself accepts far more, but a query projecting past this many columns fails to prepare). Narrow the presets' include globs so fewer/other files are indexed, or fix whatever is generating unbounded frontmatter keys.`\n );\n }\n },\n addColumns,\n reconcileContent: reconcileTursoContent,\n recordDuration: recordReconcileDuration,\n};\n"],"names":["SenseError","hasUnspacedRun","quoteIdent","recordReconcileDuration","BEGIN_WRITE","stemFolded","CONTENT_FTS_DDL","CONTENT_FTS_NAMES","ngramSidecar","text","MAX_FRONTMATTER_COLUMNS","FTS_REBUILD_THRESHOLD","INSERT_CONTENT_SQL","contentRow","doc","title","summary","search","relPath","tursoFtsStrategy","delta","churn","reparsed","length","vanished","files","reconcileTursoContentWithStrategy","conn","touched","docs","strategy","bulk","name","exec","placeholders","map","join","runBatch","ddl","reconcileTursoContent","_cfg","addColumns","names","tursoDialect","beginMode","checkColumnLimit","count","reconcileContent","recordDuration"],"mappings":"AACA,SAASA,UAAU,QAAQ,kBAAkB;AAG7C,SAASC,cAAc,QAAQ,wBAAwB;AACvD,SAASC,UAAU,EAAEC,uBAAuB,QAAQ,eAAe;AACnE,SAASC,WAAW,QAAQ,oBAAoB;AAEhD,SAASC,UAAU,QAAQ,oBAAoB;AAE/C,qFAAqF;AACrF,sFAAsF;AAEtF,+FAA+F;AAC/F,2FAA2F;AAC3F,iGAAiG;AACjG,uGAAuG;AACvG,OAAO,MAAMC,kBAAkB;IAC7B,CAAC,mKAAmK,CAAC;IACrK,CAAC,gMAAgM,CAAC;CACnM,CAAU;AACX,OAAO,MAAMC,oBAAoB;IAAC;IAAe;CAAoB,CAAU;AAE/E,6FAA6F;AAC7F,gGAAgG;AAChG,SAASC,aAAaC,IAAY;IAChC,OAAOR,eAAeQ,QAAQA,OAAO;AACvC;AAEA,8FAA8F;AAC9F,+GAA+G;AAC/G,MAAMC,0BAA0B;AAEhC,8FAA8F;AAC9F,+EAA+E;AAC/E,OAAO,MAAMC,wBAAwB,IAAI;AAIzC,yFAAyF;AACzF,MAAMC,qBAAqB,CAAC,qKAAqK,CAAC;AAElM,SAASC,WAAWC,GAAc;IAChC,MAAM,EAAEC,KAAK,EAAEC,OAAO,EAAEP,IAAI,EAAE,GAAGK,IAAIG,MAAM;IAC3C,OAAO;QAACH,IAAII,OAAO;QAAEH;QAAOC;QAASP;QAAMJ,WAAWU;QAAQV,WAAWW;QAAUX,WAAWI;QAAOD,aAAaO;QAAQP,aAAaQ;QAAUR,aAAaC;KAAM;AACtK;AAEA,OAAO,SAASU,iBAAiBC,KAAqB;IACpD,MAAMC,QAAQD,MAAME,QAAQ,CAACC,MAAM,GAAGH,MAAMI,QAAQ,CAACD,MAAM;IAC3D,OAAOH,MAAMK,KAAK,CAACF,MAAM,KAAK,KAAKF,QAAQV,wBAAwB,YAAY;AACjF;AAEA,kGAAkG;AAClG,kGAAkG;AAClG,OAAO,eAAee,kCAAkCC,IAAgB,EAAEC,OAAiB,EAAEC,IAAiB,EAAEC,QAA0B;IACxI,6FAA6F;IAC7F,qDAAqD;IACrD,MAAMC,OAAOD,aAAa;IAC1B,IAAIC,MAAM,KAAK,MAAMC,QAAQzB,kBAAmB,MAAMoB,KAAKM,IAAI,CAAC,CAAC,qBAAqB,EAAED,MAAM;IAE9F,0FAA0F;IAC1F,8DAA8D;IAC9D,IAAIJ,QAAQL,MAAM,GAAG,GAAG;QACtB,MAAMW,eAAeN,QAAQO,GAAG,CAAC,IAAM,KAAKC,IAAI,CAAC;QACjD,MAAMT,KAAKU,QAAQ,CAAC,CAAC,qCAAqC,EAAEH,aAAa,CAAC,CAAC,EAAE;YAACN;SAAQ;IACxF;IACA,IAAIC,KAAKN,MAAM,GAAG,GAAG,MAAMI,KAAKU,QAAQ,CAACzB,oBAAoBiB,KAAKM,GAAG,CAACtB;IACtE,IAAIkB,MAAM,KAAK,MAAMO,OAAOhC,gBAAiB,MAAMqB,KAAKM,IAAI,CAACK;AAC/D;AAEA,eAAeC,sBAAsBZ,IAAgB,EAAEC,OAAiB,EAAEC,IAAiB,EAAET,KAAqB,EAAEoB,IAAY;IAC9H,MAAMd,kCAAkCC,MAAMC,SAASC,MAAMV,iBAAiBC;AAChF;AAEA,yFAAyF;AACzF,eAAeqB,WAAWd,IAAgB,EAAEe,KAAe;IACzD,KAAK,MAAMV,QAAQU,MAAO,MAAMf,KAAKM,IAAI,CAAC,CAAC,mCAAmC,EAAE/B,WAAW8B,OAAO;AACpG;AAEA,OAAO,MAAMW,eAAiC;IAC5CC,WAAW,IAAMxC;IACjByC,kBAAiBC,KAAK;QACpB,IAAIA,QAAQpC,yBAAyB;YACnC,MAAM,IAAIV,WACR,gBACA,CAAC,uBAAuB,EAAE8C,MAAM,2DAA2D,EAAEpC,wBAAwB,gPAAgP,CAAC;QAE1W;IACF;IACA+B;IACAM,kBAAkBR;IAClBS,gBAAgB7C;AAClB,EAAE"}