sensemaking 0.15.3 → 0.16.0

package/dist/cjs/db/reconcile.js.map

@@ -1 +1 @@
-{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/db/reconcile.ts"],"sourcesContent":["import type { DatabaseSync, StatementSync } from 'node:sqlite';\nimport type { Config } from '../config/index.ts';\nimport { contentTokenize } from '../config/index.ts';\nimport { SenseError } from '../errors.ts';\nimport { activeFeatures } from '../features/index.ts';\nimport type { ReconcileDelta } from '../features/types.ts';\nimport { progress } from '../progress.ts';\nimport type { ParsedDoc } from '../scan.ts';\nimport { listFiles, parseFile, RESERVED_COLUMNS } from '../scan.ts';\nimport { segmentField } from '../segment.ts';\nimport { getColumns, getMeta, quoteIdent, setMeta } from './shared.ts';\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.\nconst CORE_FRONTMATTER_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_parse_error']);\n\n// SQLite's compile-time SQLITE_MAX_COLUMN, default 2000 (https://www.sqlite.org/limits.html).\nconst MAX_FRONTMATTER_COLUMNS = 2000;\n\n// Shared by reconcile() and the tokenize-only rebuild in open(), so both prepare the same\n// literal instead of two copies drifting apart.\nconst INSERT_CONTENT_SQL = `INSERT INTO content (rowid, title, summary, text, \"path\", title_seg, summary_seg, text_seg) VALUES ((SELECT rowid FROM frontmatter WHERE \"path\" = ?), ?, ?, ?, ?, ?, ?, ?)`;\n\n// Shared by reconcile() and the tokenize-only rebuild in open(), so content population is\n// defined once. Assumes the frontmatter row for doc.relPath already exists (rowid lookup).\nfunction insertContentRow(insertBody: StatementSync, doc: ParsedDoc, segmenting: boolean): void {\n  insertBody.run(doc.relPath, doc.search.title, doc.search.summary, doc.search.text, doc.relPath, segmenting ? segmentField(doc.search.title) : '', segmenting ? segmentField(doc.search.summary) : '', segmenting ? segmentField(doc.search.text) : '');\n}\n\nexport function reconcile(db: DatabaseSync, cfg: Config, baseDir: string): { parsed: number; warnings: string[] } {\n  const files = listFiles(cfg, baseDir);\n  const currentSet = new Set(files.map((f) => f.relPath));\n\n  const existingRows = db.prepare(`SELECT \"path\", \"_mtime\", \"_size\" FROM frontmatter`).all() as Array<{\n    path: string;\n    _mtime: number;\n    _size: number;\n  }>;\n  const existing = new Map(existingRows.map((r) => [r.path, r]));\n  const vanished = existingRows.filter((r) => !currentSet.has(r.path)).map((r) => r.path);\n\n  const toReparse = files.filter((f) => {\n    const row = existing.get(f.relPath);\n    return !row || row._mtime !== f.mtimeMs || row._size !== f.size;\n  });\n\n  if (vanished.length === 0 && toReparse.length === 0) return { parsed: 0, warnings: [] };\n\n  const features = activeFeatures(cfg);\n  const seenColumns = getColumns(db);\n  const newColumns: string[] = [];\n  const parsedDocs: ParsedDoc[] = [];\n  const warnings: string[] = [];\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  let parsedCount = 0;\n  for (const file of toReparse) {\n    // A doc only gets extract/store from features that apply to it (currently: embed, via\n    // FileStat.embed -- true iff the config names an embedding model).\n    const fileFeatures = features.filter((feature) => !feature.enabledForFile || feature.enabledForFile(cfg, file));\n    const { doc, warnings: fileWarnings } = parseFile(file, fileFeatures);\n    report.tick(++parsedCount);\n    warnings.push(...fileWarnings);\n    for (const key of Object.keys(doc.data)) {\n      if (!seenColumns.has(key)) {\n        seenColumns.add(key);\n        newColumns.push(key);\n      }\n    }\n    parsedDocs.push(doc);\n  }\n  report.finish();\n\n  const allColumns = [...seenColumns];\n  // Fence before ALTERing: SQLite's own failure past this point is a raw\n  // \"too many columns on sqlite_altertab_frontmatter\" with no indication of the boundary or the levers.\n  if (allColumns.length > MAX_FRONTMATTER_COLUMNS) {\n    throw new SenseError(\n      'COLUMN_LIMIT',\n      `frontmatter would need ${allColumns.length} columns, crossing SQLite's compile-time SQLITE_MAX_COLUMN limit (default ${MAX_FRONTMATTER_COLUMNS}; see https://www.sqlite.org/limits.html). Narrow the presets' include globs so fewer/other files are indexed, or fix whatever is generating unbounded frontmatter keys.`\n    );\n  }\n  // Columns the frontmatter upsert actually writes: core + parsed frontmatter keys, never a\n  // feature-owned reserved column (see CORE_FRONTMATTER_COLUMNS above).\n  const writableColumns = allColumns.filter((c) => CORE_FRONTMATTER_COLUMNS.has(c) || !RESERVED_COLUMNS.has(c));\n  // ON CONFLICT UPDATE (not OR REPLACE) keeps the row's rowid stable across reparses --\n  // content rows are coupled to that rowid below.\n  const insertSql = `INSERT INTO frontmatter (${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\n  const txStart = Date.now();\n  db.exec('BEGIN');\n  try {\n    for (const col of newColumns) db.exec(`ALTER TABLE frontmatter ADD COLUMN ${quoteIdent(col)}`);\n    // FTS5 has no upsert, so delete-before-insert into `content`; coupled to the\n    // frontmatter rowid (indexed via its PRIMARY KEY) instead of the UNINDEXED `path`\n    // column, which a per-row DELETE would otherwise scan the whole table to find.\n    const delBody = db.prepare(`DELETE FROM content WHERE rowid = (SELECT rowid FROM frontmatter WHERE \"path\" = ?)`);\n    const delPresetFiles = db.prepare(`DELETE FROM preset_files WHERE \"path\" = ?`);\n    const insertPresetFile = db.prepare(`INSERT INTO preset_files (\"path\", preset) VALUES (?, ?)`);\n    if (vanished.length > 0) {\n      const del = db.prepare(`DELETE FROM frontmatter WHERE \"path\" = ?`);\n      for (const path of vanished) {\n        // content delete must run first: it looks up the frontmatter rowid by path, which\n        // the frontmatter delete below would otherwise have already removed.\n        delBody.run(path);\n        del.run(path);\n        delPresetFiles.run(path);\n        for (const feature of features) feature.remove?.(db, path, delta);\n      }\n    }\n    if (parsedDocs.length > 0) {\n      const insert = db.prepare(insertSql);\n      const insertBody = db.prepare(INSERT_CONTENT_SQL);\n      // A non-default tokenizer means the tree has chosen its own scheme; a phrase query over\n      // grapheme runs would be nonsense against trigram, so the sidecars stay empty.\n      const segmenting = contentTokenize(cfg) === undefined;\n      for (const doc of parsedDocs) {\n        const values = 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        // Frontmatter upsert first: content's rowid lookup below depends on this row existing.\n        insert.run(...values);\n        // Only for docs that have rows: an unconditional delete made cold crawls quadratic.\n        if (existing.has(doc.relPath)) {\n          delBody.run(doc.relPath);\n          for (const feature of features) feature.remove?.(db, doc.relPath, delta);\n        }\n        insertContentRow(insertBody, doc, segmenting);\n        // A preset edit forces a full rebuild, so an unchanged doc's coverage is already\n        // correct; new docs have nothing to clear, which keeps cold builds linear.\n        if (existing.has(doc.relPath)) delPresetFiles.run(doc.relPath);\n        for (const presetName of doc.presets) insertPresetFile.run(doc.relPath, presetName);\n        for (const feature of features) feature.store?.(db, doc.relPath, doc.extracted[feature.name], delta);\n      }\n    }\n    for (const feature of features) feature.afterReconcile?.(db, delta);\n    db.exec('COMMIT');\n  } catch (err) {\n    db.exec('ROLLBACK');\n    throw err;\n  }\n\n  // Reconcile's own write-transaction duration, for open()'s derived busy_timeout (F):\n  // keep the observed max so a big watcher reconcile's lock hold is what the next open bounds its wait against.\n  const durationMs = Date.now() - txStart;\n  const prevRaw = getMeta(db, 'reconcile_max_ms');\n  // -1, not 0, so a genuinely 0ms first reconcile (sub-millisecond, common on a tiny tree)\n  // still gets recorded instead of losing to the \"nothing recorded yet\" default.\n  const prevMax = prevRaw === null ? -1 : Number(prevRaw);\n  if (durationMs > prevMax) setMeta(db, 'reconcile_max_ms', String(durationMs));\n\n  return { parsed: parsedDocs.length, warnings };\n}\n\n// Segment keys that moved between two feature signatures (see config.featureSignature's\n// format: global features, embed provider, tokenize, then one segment per preset).\nexport function changedSignatureKeys(before: string, after: string): Set<string> {\n  const keyOf = (part: string) => (part.startsWith('preset:') ? part.split(':').slice(0, 2).join(':') : part.split(':')[0]);\n  const parse = (sig: string) => new Map(sig.split('|').map((part) => [keyOf(part), part]));\n  const a = parse(before);\n  const b = parse(after);\n  const changed = new Set<string>();\n  for (const [key, val] of b) if (a.get(key) !== val) changed.add(key);\n  for (const key of a.keys()) if (!b.has(key)) changed.add(key);\n  return changed;\n}\n\n// Names what moved, for the rebuild notice.\nexport function signatureDiff(before: string, after: string): string {\n  const changed = changedSignatureKeys(before, after);\n  const label = (key: string) => (key === 'embed' ? 'embed settings' : key === 'tokenize' ? 'content tokenizer' : key.startsWith('preset:') ? `preset \"${key.slice(7)}\"` : 'features');\n  return changed.size === 0 ? 'features' : [...changed].map(label).join(', ');\n}\n\n// The tokenize-only rebuild in open(): content is dropped and repopulated from files already\n// listed in frontmatter, which itself is untouched. Frontmatter, links, sections, and\n// embeddings are file-derived and tokenizer-independent, so they survive. No feature extractors\n// run here -- doc.search (title/summary/text) is all content population needs. Returns\n// parseFile's per-file warnings (e.g. a bad date) so open() can surface them -- mtimes are\n// untouched, so reconcile() never reparses these files and would otherwise never emit them again.\nexport function rebuildContentTable(db: DatabaseSync, cfg: Config, baseDir: string): string[] {\n  const known = new Set((db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path));\n  const files = listFiles(cfg, baseDir).filter((f) => known.has(f.relPath));\n  const segmenting = contentTokenize(cfg) === undefined;\n  const insertBody = db.prepare(INSERT_CONTENT_SQL);\n  const warnings: string[] = [];\n  db.exec('BEGIN');\n  try {\n    for (const file of files) {\n      const { doc, warnings: fileWarnings } = parseFile(file);\n      warnings.push(...fileWarnings);\n      insertContentRow(insertBody, doc, segmenting);\n    }\n    db.exec('COMMIT');\n  } catch (err) {\n    db.exec('ROLLBACK');\n    throw err;\n  }\n  return warnings;\n}\n"],"names":["changedSignatureKeys","rebuildContentTable","reconcile","signatureDiff","CORE_FRONTMATTER_COLUMNS","Set","MAX_FRONTMATTER_COLUMNS","INSERT_CONTENT_SQL","insertContentRow","insertBody","doc","segmenting","run","relPath","search","title","summary","text","segmentField","db","cfg","baseDir","files","listFiles","currentSet","map","f","existingRows","prepare","all","existing","Map","r","path","vanished","filter","has","toReparse","row","get","_mtime","mtimeMs","_size","size","length","parsed","warnings","features","activeFeatures","seenColumns","getColumns","newColumns","parsedDocs","report","progress","parsedCount","file","fileFeatures","feature","enabledForFile","parseFile","fileWarnings","tick","push","Object","keys","data","key","add","finish","allColumns","SenseError","writableColumns","c","RESERVED_COLUMNS","insertSql","quoteIdent","join","added","delta","reparsed","d","txStart","Date","now","exec","col","delBody","delPresetFiles","insertPresetFile","del","remove","insert","contentTokenize","undefined","values","ctimeMs","parseError","presets","presetName","store","extracted","name","afterReconcile","err","durationMs","prevRaw","getMeta","prevMax","Number","setMeta","String","before","after","keyOf","part","startsWith","split","slice","parse","sig","a","b","changed","val","label","known"],"mappings":";;;;;;;;;;;QA0KgBA;eAAAA;;QAwBAC;eAAAA;;QArKAC;eAAAA;;QAyJAC;eAAAA;;;uBApLgB;wBACL;wBACI;0BAEN;sBAE8B;yBAC1B;wBAC4B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAEzD,6FAA6F;AAC7F,4EAA4E;AAC5E,IAAMC,2BAA2B,IAAIC,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;CAAe;AAE9F,8FAA8F;AAC9F,IAAMC,0BAA0B;AAEhC,0FAA0F;AAC1F,gDAAgD;AAChD,IAAMC,qBAAqB;AAE3B,0FAA0F;AAC1F,2FAA2F;AAC3F,SAASC,iBAAiBC,UAAyB,EAAEC,GAAc,EAAEC,UAAmB;IACtFF,WAAWG,GAAG,CAACF,IAAIG,OAAO,EAAEH,IAAII,MAAM,CAACC,KAAK,EAAEL,IAAII,MAAM,CAACE,OAAO,EAAEN,IAAII,MAAM,CAACG,IAAI,EAAEP,IAAIG,OAAO,EAAEF,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACC,KAAK,IAAI,IAAIJ,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACE,OAAO,IAAI,IAAIL,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACG,IAAI,IAAI;AACrP;AAEO,SAASf,UAAUiB,EAAgB,EAAEC,GAAW,EAAEC,OAAe;IACtE,IAAMC,QAAQC,IAAAA,iBAAS,EAACH,KAAKC;IAC7B,IAAMG,aAAa,IAAInB,IAAIiB,MAAMG,GAAG,CAAC,SAACC;eAAMA,EAAEb,OAAO;;IAErD,IAAMc,eAAeR,GAAGS,OAAO,CAAC,qDAAqDC,GAAG;IAKxF,IAAMC,WAAW,IAAIC,IAAIJ,aAAaF,GAAG,CAAC,SAACO;eAAM;YAACA,EAAEC,IAAI;YAAED;SAAE;;IAC5D,IAAME,WAAWP,aAAaQ,MAAM,CAAC,SAACH;eAAM,CAACR,WAAWY,GAAG,CAACJ,EAAEC,IAAI;OAAGR,GAAG,CAAC,SAACO;eAAMA,EAAEC,IAAI;;IAEtF,IAAMI,YAAYf,MAAMa,MAAM,CAAC,SAACT;QAC9B,IAAMY,MAAMR,SAASS,GAAG,CAACb,EAAEb,OAAO;QAClC,OAAO,CAACyB,OAAOA,IAAIE,MAAM,KAAKd,EAAEe,OAAO,IAAIH,IAAII,KAAK,KAAKhB,EAAEiB,IAAI;IACjE;IAEA,IAAIT,SAASU,MAAM,KAAK,KAAKP,UAAUO,MAAM,KAAK,GAAG,OAAO;QAAEC,QAAQ;QAAGC,UAAU,EAAE;IAAC;IAEtF,IAAMC,WAAWC,IAAAA,wBAAc,EAAC5B;IAChC,IAAM6B,cAAcC,IAAAA,oBAAU,EAAC/B;IAC/B,IAAMgC,aAAuB,EAAE;IAC/B,IAAMC,aAA0B,EAAE;IAClC,IAAMN,WAAqB,EAAE;IAE7B,oFAAoF;IACpF,uDAAuD;IACvD,IAAMO,SAASC,IAAAA,oBAAQ,EAAC,mBAAmBjB,UAAUO,MAAM;IAC3D,IAAIW,cAAc;QACb,kCAAA,2BAAA;;;YAAA,IAAMC,OAAN;gBAMHV;YALA,sFAAsF;YACtF,mEAAmE;YACnE,IAAMW,eAAeV,SAASZ,MAAM,CAAC,SAACuB;uBAAY,CAACA,QAAQC,cAAc,IAAID,QAAQC,cAAc,CAACvC,KAAKoC;;YACzG,IAAwCI,aAAAA,IAAAA,iBAAS,EAACJ,MAAMC,eAAhD/C,MAAgCkD,WAAhClD,KAAKoC,AAAUe,eAAiBD,WAA3Bd;YACbO,OAAOS,IAAI,CAAC,EAAEP;YACdT,CAAAA,YAAAA,UAASiB,IAAI,OAAbjB,WAAc,qBAAGe;gBACZ,kCAAA,2BAAA;;gBAAL,QAAK,YAAaG,OAAOC,IAAI,CAACvD,IAAIwD,IAAI,sBAAjC,UAAA,6BAAA,SAAA,yBAAA,iCAAoC;oBAApC,IAAMC,MAAN;oBACH,IAAI,CAAClB,YAAYb,GAAG,CAAC+B,MAAM;wBACzBlB,YAAYmB,GAAG,CAACD;wBAChBhB,WAAWY,IAAI,CAACI;oBAClB;gBACF;;gBALK;gBAAA;;;yBAAA,6BAAA;wBAAA;;;wBAAA;8BAAA;;;;YAMLf,WAAWW,IAAI,CAACrD;QAClB;QAdA,QAAK,YAAc2B,8BAAd,SAAA,6BAAA,QAAA,yBAAA;;QAAA;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;IAeLgB,OAAOgB,MAAM;IAEb,IAAMC,aAAc,qBAAGrB;IACvB,uEAAuE;IACvE,sGAAsG;IACtG,IAAIqB,WAAW1B,MAAM,GAAGtC,yBAAyB;QAC/C,MAAM,IAAIiE,oBAAU,CAClB,gBACA,AAAC,0BAAuHjE,OAA9FgE,WAAW1B,MAAM,EAAC,8EAAoG,OAAxBtC,yBAAwB;IAEpJ;IACA,0FAA0F;IAC1F,sEAAsE;IACtE,IAAMkE,kBAAkBF,WAAWnC,MAAM,CAAC,SAACsC;eAAMrE,yBAAyBgC,GAAG,CAACqC,MAAM,CAACC,wBAAgB,CAACtC,GAAG,CAACqC;;IAC1G,sFAAsF;IACtF,gDAAgD;IAChD,IAAME,YAAY,AAAC,4BAAkFH,OAAvDA,gBAAgB/C,GAAG,CAACmD,oBAAU,EAAEC,IAAI,CAAC,OAAM,cAA4FL,OAAhFA,gBAAgB/C,GAAG,CAAC;eAAM;OAAKoD,IAAI,CAAC,OAAM,wCAGjI,OAHuKL,gBAClLrC,MAAM,CAAC,SAACsC;eAAMA,MAAM;OACpBhD,GAAG,CAAC,SAACgD;eAAM,AAAC,GAA8BG,OAA5BA,IAAAA,oBAAU,EAACH,IAAG,gBAA4B,OAAdG,IAAAA,oBAAU,EAACH;OACrDI,IAAI,CAAC;IAER,IAAMC,QAAQzC,UAAUF,MAAM,CAAC,SAACT;eAAM,CAACI,SAASM,GAAG,CAACV,EAAEb,OAAO;OAAGY,GAAG,CAAC,SAACC;eAAMA,EAAEb,OAAO;;IACpF,IAAMkE,QAAwB;QAAEzD,OAAAA;QAAO0D,UAAU5B,WAAW3B,GAAG,CAAC,SAACwD;mBAAMA,EAAEpE,OAAO;;QAAGiE,OAAAA;QAAO5C,UAAAA;IAAS;IAEnG,IAAMgD,UAAUC,KAAKC,GAAG;IACxBjE,GAAGkE,IAAI,CAAC;IACR,IAAI;YAkD8B3B;YAjD3B,mCAAA,4BAAA;;YAAL,QAAK,aAAaP,+BAAb,UAAA,8BAAA,SAAA,0BAAA;gBAAA,IAAMmC,MAAN;gBAAyBnE,GAAGkE,IAAI,CAAC,AAAC,sCAAqD,OAAhBT,IAAAA,oBAAU,EAACU;;;YAAlF;YAAA;;;qBAAA,8BAAA;oBAAA;;;oBAAA;0BAAA;;;;QACL,6EAA6E;QAC7E,kFAAkF;QAClF,+EAA+E;QAC/E,IAAMC,UAAUpE,GAAGS,OAAO,CAAC;QAC3B,IAAM4D,iBAAiBrE,GAAGS,OAAO,CAAC;QAClC,IAAM6D,mBAAmBtE,GAAGS,OAAO,CAAC;QACpC,IAAIM,SAASU,MAAM,GAAG,GAAG;YACvB,IAAM8C,MAAMvE,GAAGS,OAAO,CAAC;gBAClB,mCAAA,4BAAA;;gBAAL,QAAK,aAAcM,6BAAd,UAAA,8BAAA,SAAA,0BAAA,kCAAwB;oBAAxB,IAAMD,OAAN;wBAM6ByB;oBALhC,kFAAkF;oBAClF,qEAAqE;oBACrE6B,QAAQ3E,GAAG,CAACqB;oBACZyD,IAAI9E,GAAG,CAACqB;oBACRuD,eAAe5E,GAAG,CAACqB;wBACd,mCAAA,4BAAA;;wBAAL,QAAK,aAAiBc,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMW,UAAN;6BAA2BA,kBAAAA,QAAQiC,MAAM,cAAdjC,sCAAAA,qBAAAA,SAAiBvC,IAAIc,MAAM8C;;;wBAAtD;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;gBACP;;gBAPK;gBAAA;;;yBAAA,8BAAA;wBAAA;;;wBAAA;8BAAA;;;;QAQP;QACA,IAAI3B,WAAWR,MAAM,GAAG,GAAG;YACzB,IAAMgD,SAASzE,GAAGS,OAAO,CAAC+C;YAC1B,IAAMlE,aAAaU,GAAGS,OAAO,CAACrB;YAC9B,wFAAwF;YACxF,+EAA+E;YAC/E,IAAMI,aAAakF,IAAAA,wBAAe,EAACzE,SAAS0E;gBACvC,mCAAA,4BAAA;;;oBAAA,IAAMpF,MAAN;wBAUH,uFAAuF;oBACvFkF;wBAWgClC;oBArBhC,IAAMqC,SAASvB,gBAAgB/C,GAAG,CAAC,SAAC6D;4BAO3B5E;wBANP,IAAI4E,QAAQ,QAAQ,OAAO5E,IAAIG,OAAO;wBACtC,IAAIyE,QAAQ,UAAU,OAAO5E,IAAI+B,OAAO;wBACxC,IAAI6C,QAAQ,UAAU,OAAO5E,IAAIsF,OAAO;wBACxC,IAAIV,QAAQ,SAAS,OAAO5E,IAAIiC,IAAI;wBACpC,mFAAmF;wBACnF,IAAI2C,QAAQ,gBAAgB,OAAO5E,IAAIuF,UAAU;wBACjD,QAAOvF,gBAAAA,IAAIwD,IAAI,CAACoB,IAAI,cAAb5E,2BAAAA,gBAAiB;oBAC1B;oBAEAkF,CAAAA,UAAAA,QAAOhF,GAAG,OAAVgF,SAAW,qBAAGG;oBACd,oFAAoF;oBACpF,IAAIjE,SAASM,GAAG,CAAC1B,IAAIG,OAAO,GAAG;4BAEG6C;wBADhC6B,QAAQ3E,GAAG,CAACF,IAAIG,OAAO;4BAClB,kCAAA,2BAAA;;4BAAL,QAAK,YAAiBkC,6BAAjB,SAAA,6BAAA,QAAA,yBAAA;gCAAA,IAAMW,UAAN;iCAA2BA,kBAAAA,QAAQiC,MAAM,cAAdjC,sCAAAA,qBAAAA,SAAiBvC,IAAIT,IAAIG,OAAO,EAAEkE;;;4BAA7D;4BAAA;;;qCAAA,6BAAA;oCAAA;;;oCAAA;0CAAA;;;;oBACP;oBACAvE,iBAAiBC,YAAYC,KAAKC;oBAClC,iFAAiF;oBACjF,2EAA2E;oBAC3E,IAAImB,SAASM,GAAG,CAAC1B,IAAIG,OAAO,GAAG2E,eAAe5E,GAAG,CAACF,IAAIG,OAAO;wBACxD,mCAAA,4BAAA;;wBAAL,QAAK,aAAoBH,IAAIwF,OAAO,qBAA/B,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMC,aAAN;4BAAiCV,iBAAiB7E,GAAG,CAACF,IAAIG,OAAO,EAAEsF;;;wBAAnE;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;wBACA,mCAAA,4BAAA;;wBAAL,QAAK,aAAiBpD,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMW,WAAN;6BAA2BA,iBAAAA,SAAQ0C,KAAK,cAAb1C,qCAAAA,oBAAAA,UAAgBvC,IAAIT,IAAIG,OAAO,EAAEH,IAAI2F,SAAS,CAAC3C,SAAQ4C,IAAI,CAAC,EAAEvB;;;wBAAzF;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;gBACP;gBAvBA,QAAK,aAAa3B,+BAAb,UAAA,8BAAA,SAAA,0BAAA;;gBAAA;gBAAA;;;yBAAA,8BAAA;wBAAA;;;wBAAA;8BAAA;;;;QAwBP;YACK,mCAAA,4BAAA;;YAAL,QAAK,aAAiBL,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;gBAAA,IAAMW,WAAN;iBAA2BA,0BAAAA,SAAQ6C,cAAc,cAAtB7C,8CAAAA,6BAAAA,UAAyBvC,IAAI4D;;;YAAxD;YAAA;;;qBAAA,8BAAA;oBAAA;;;oBAAA;0BAAA;;;;QACL5D,GAAGkE,IAAI,CAAC;IACV,EAAE,OAAOmB,KAAK;QACZrF,GAAGkE,IAAI,CAAC;QACR,MAAMmB;IACR;IAEA,qFAAqF;IACrF,8GAA8G;IAC9G,IAAMC,aAAatB,KAAKC,GAAG,KAAKF;IAChC,IAAMwB,UAAUC,IAAAA,iBAAO,EAACxF,IAAI;IAC5B,yFAAyF;IACzF,+EAA+E;IAC/E,IAAMyF,UAAUF,YAAY,OAAO,CAAC,IAAIG,OAAOH;IAC/C,IAAID,aAAaG,SAASE,IAAAA,iBAAO,EAAC3F,IAAI,oBAAoB4F,OAAON;IAEjE,OAAO;QAAE5D,QAAQO,WAAWR,MAAM;QAAEE,UAAAA;IAAS;AAC/C;AAIO,SAAS9C,qBAAqBgH,MAAc,EAAEC,KAAa;IAChE,IAAMC,QAAQ,eAACC;eAAkBA,KAAKC,UAAU,CAAC,aAAaD,KAAKE,KAAK,CAAC,KAAKC,KAAK,CAAC,GAAG,GAAGzC,IAAI,CAAC,OAAOsC,KAAKE,KAAK,CAAC,IAAI,CAAC,EAAE;;IACxH,IAAME,QAAQ,eAACC;eAAgB,IAAIzF,IAAIyF,IAAIH,KAAK,CAAC,KAAK5F,GAAG,CAAC,SAAC0F;mBAAS;gBAACD,MAAMC;gBAAOA;aAAK;;;IACvF,IAAMM,IAAIF,MAAMP;IAChB,IAAMU,IAAIH,MAAMN;IAChB,IAAMU,UAAU,IAAItH;QACf,kCAAA,2BAAA;;QAAL,QAAK,YAAoBqH,sBAApB,SAAA,6BAAA,QAAA,yBAAA;YAAA,mCAAA,iBAAOvD,sBAAKyD;YAAW,IAAIH,EAAElF,GAAG,CAAC4B,SAASyD,KAAKD,QAAQvD,GAAG,CAACD;;;QAA3D;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;QACA,mCAAA,4BAAA;;QAAL,QAAK,aAAasD,EAAExD,IAAI,uBAAnB,UAAA,8BAAA,SAAA,0BAAA;YAAA,IAAME,OAAN;YAAuB,IAAI,CAACuD,EAAEtF,GAAG,CAAC+B,OAAMwD,QAAQvD,GAAG,CAACD;;;QAApD;QAAA;;;iBAAA,8BAAA;gBAAA;;;gBAAA;sBAAA;;;;IACL,OAAOwD;AACT;AAGO,SAASxH,cAAc6G,MAAc,EAAEC,KAAa;IACzD,IAAMU,UAAU3H,qBAAqBgH,QAAQC;IAC7C,IAAMY,QAAQ,eAAC1D;eAAiBA,QAAQ,UAAU,mBAAmBA,QAAQ,aAAa,sBAAsBA,IAAIiD,UAAU,CAAC,aAAa,AAAC,WAAuB,OAAbjD,IAAImD,KAAK,CAAC,IAAG,OAAK;;IACzK,OAAOK,QAAQhF,IAAI,KAAK,IAAI,aAAa,AAAC,qBAAGgF,SAASlG,GAAG,CAACoG,OAAOhD,IAAI,CAAC;AACxE;AAQO,SAAS5E,oBAAoBkB,EAAgB,EAAEC,GAAW,EAAEC,OAAe;IAChF,IAAMyG,QAAQ,IAAIzH,IAAI,AAACc,GAAGS,OAAO,CAAC,kCAAkCC,GAAG,GAA+BJ,GAAG,CAAC,SAACO;eAAMA,EAAEC,IAAI;;IACvH,IAAMX,QAAQC,IAAAA,iBAAS,EAACH,KAAKC,SAASc,MAAM,CAAC,SAACT;eAAMoG,MAAM1F,GAAG,CAACV,EAAEb,OAAO;;IACvE,IAAMF,aAAakF,IAAAA,wBAAe,EAACzE,SAAS0E;IAC5C,IAAMrF,aAAaU,GAAGS,OAAO,CAACrB;IAC9B,IAAMuC,WAAqB,EAAE;IAC7B3B,GAAGkE,IAAI,CAAC;IACR,IAAI;YACG,kCAAA,2BAAA;;YAAL,QAAK,YAAc/D,0BAAd,SAAA,6BAAA,QAAA,yBAAA,iCAAqB;gBAArB,IAAMkC,OAAN;oBAEHV;gBADA,IAAwCc,aAAAA,IAAAA,iBAAS,EAACJ,OAA1C9C,MAAgCkD,WAAhClD,KAAKoC,AAAUe,eAAiBD,WAA3Bd;gBACbA,CAAAA,YAAAA,UAASiB,IAAI,OAAbjB,WAAc,qBAAGe;gBACjBrD,iBAAiBC,YAAYC,KAAKC;YACpC;;YAJK;YAAA;;;qBAAA,6BAAA;oBAAA;;;oBAAA;0BAAA;;;;QAKLQ,GAAGkE,IAAI,CAAC;IACV,EAAE,OAAOmB,KAAK;QACZrF,GAAGkE,IAAI,CAAC;QACR,MAAMmB;IACR;IACA,OAAO1D;AACT"}
+{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/db/reconcile.ts"],"sourcesContent":["import type { DatabaseSync, StatementSync } from 'node:sqlite';\nimport type { Config } from '../config/index.ts';\nimport { contentTokenize } from '../config/index.ts';\nimport { SenseError } from '../errors.ts';\nimport { activeFeatures } from '../features/index.ts';\nimport type { ReconcileDelta } from '../features/types.ts';\nimport { progress } from '../progress.ts';\nimport type { ParsedDoc } from '../scan.ts';\nimport { listFiles, parseFile, RESERVED_COLUMNS } from '../scan.ts';\nimport { segmentField } from '../segment.ts';\nimport { getColumns, getMeta, quoteIdent, setMeta } from './shared.ts';\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.\nconst CORE_FRONTMATTER_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_parse_error']);\n\n// SQLite's compile-time SQLITE_MAX_COLUMN, default 2000 (https://www.sqlite.org/limits.html).\nconst MAX_FRONTMATTER_COLUMNS = 2000;\n\n// Shared by reconcile() and the tokenize-only rebuild in open(), so both prepare the same\n// literal instead of two copies drifting apart.\nconst INSERT_CONTENT_SQL = `INSERT INTO content (rowid, title, summary, text, \"path\", title_seg, summary_seg, text_seg) VALUES ((SELECT rowid FROM frontmatter WHERE \"path\" = ?), ?, ?, ?, ?, ?, ?, ?)`;\n\n// Shared by reconcile() and the tokenize-only rebuild in open(), so content population is\n// defined once. Assumes the frontmatter row for doc.relPath already exists (rowid lookup).\nfunction insertContentRow(insertBody: StatementSync, doc: ParsedDoc, segmenting: boolean): void {\n  insertBody.run(doc.relPath, doc.search.title, doc.search.summary, doc.search.text, doc.relPath, segmenting ? segmentField(doc.search.title) : '', segmenting ? segmentField(doc.search.summary) : '', segmenting ? segmentField(doc.search.text) : '');\n}\n\nexport function reconcile(db: DatabaseSync, cfg: Config, baseDir: string): { parsed: number; warnings: string[] } {\n  const files = listFiles(cfg, baseDir);\n  const currentSet = new Set(files.map((f) => f.relPath));\n\n  const existingRows = db.prepare(`SELECT \"path\", \"_mtime\", \"_size\" FROM frontmatter`).all() as Array<{\n    path: string;\n    _mtime: number;\n    _size: number;\n  }>;\n  const existing = new Map(existingRows.map((r) => [r.path, r]));\n  const vanished = existingRows.filter((r) => !currentSet.has(r.path)).map((r) => r.path);\n\n  const toReparse = files.filter((f) => {\n    const row = existing.get(f.relPath);\n    return !row || row._mtime !== f.mtimeMs || row._size !== f.size;\n  });\n\n  if (vanished.length === 0 && toReparse.length === 0) return { parsed: 0, warnings: [] };\n\n  const features = activeFeatures(cfg);\n  const seenColumns = getColumns(db);\n  const newColumns: string[] = [];\n  const parsedDocs: ParsedDoc[] = [];\n  const warnings: string[] = [];\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  let parsedCount = 0;\n  for (const file of toReparse) {\n    // A doc only gets extract/store from features that apply to it (currently: embed, via\n    // FileStat.embed -- true iff the config names an embedding model).\n    const fileFeatures = features.filter((feature) => !feature.enabledForFile || feature.enabledForFile(cfg, file));\n    const { doc, warnings: fileWarnings } = parseFile(file, fileFeatures);\n    report.tick(++parsedCount);\n    warnings.push(...fileWarnings);\n    for (const key of Object.keys(doc.data)) {\n      if (!seenColumns.has(key)) {\n        seenColumns.add(key);\n        newColumns.push(key);\n      }\n    }\n    parsedDocs.push(doc);\n  }\n  report.finish();\n\n  const allColumns = [...seenColumns];\n  // Fence before ALTERing: SQLite's own failure past this point is a raw\n  // \"too many columns on sqlite_altertab_frontmatter\" with no indication of the boundary or the levers.\n  if (allColumns.length > MAX_FRONTMATTER_COLUMNS) {\n    throw new SenseError(\n      'COLUMN_LIMIT',\n      `frontmatter would need ${allColumns.length} columns, crossing SQLite's compile-time SQLITE_MAX_COLUMN limit (default ${MAX_FRONTMATTER_COLUMNS}; see https://www.sqlite.org/limits.html). Narrow the presets' include globs so fewer/other files are indexed, or fix whatever is generating unbounded frontmatter keys.`\n    );\n  }\n  // Columns the frontmatter upsert actually writes: core + parsed frontmatter keys, never a\n  // feature-owned reserved column (see CORE_FRONTMATTER_COLUMNS above).\n  const writableColumns = allColumns.filter((c) => CORE_FRONTMATTER_COLUMNS.has(c) || !RESERVED_COLUMNS.has(c));\n  // ON CONFLICT UPDATE (not OR REPLACE) keeps the row's rowid stable across reparses --\n  // content rows are coupled to that rowid below.\n  const insertSql = `INSERT INTO frontmatter (${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\n  const txStart = Date.now();\n  db.exec('BEGIN');\n  try {\n    for (const col of newColumns) db.exec(`ALTER TABLE frontmatter ADD COLUMN ${quoteIdent(col)}`);\n    // FTS5 has no upsert, so delete-before-insert into `content`; coupled to the\n    // frontmatter rowid (indexed via its PRIMARY KEY) instead of the UNINDEXED `path`\n    // column, which a per-row DELETE would otherwise scan the whole table to find.\n    const delBody = db.prepare(`DELETE FROM content WHERE rowid = (SELECT rowid FROM frontmatter WHERE \"path\" = ?)`);\n    const delPresetFiles = db.prepare(`DELETE FROM preset_files WHERE \"path\" = ?`);\n    const insertPresetFile = db.prepare(`INSERT INTO preset_files (\"path\", preset) VALUES (?, ?)`);\n    if (vanished.length > 0) {\n      const del = db.prepare(`DELETE FROM frontmatter WHERE \"path\" = ?`);\n      for (const path of vanished) {\n        // content delete must run first: it looks up the frontmatter rowid by path, which\n        // the frontmatter delete below would otherwise have already removed.\n        delBody.run(path);\n        del.run(path);\n        delPresetFiles.run(path);\n        for (const feature of features) feature.remove?.(db, path, delta);\n      }\n    }\n    if (parsedDocs.length > 0) {\n      const insert = db.prepare(insertSql);\n      const insertBody = db.prepare(INSERT_CONTENT_SQL);\n      // A non-default tokenizer means the tree has chosen its own scheme; a phrase query over\n      // grapheme runs would be nonsense against trigram, so the sidecars stay empty.\n      const segmenting = contentTokenize(cfg) === undefined;\n      for (const doc of parsedDocs) {\n        const values = 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        // Frontmatter upsert first: content's rowid lookup below depends on this row existing.\n        insert.run(...values);\n        // Only for docs that have rows: an unconditional delete made cold crawls quadratic.\n        if (existing.has(doc.relPath)) {\n          delBody.run(doc.relPath);\n          for (const feature of features) feature.remove?.(db, doc.relPath, delta);\n        }\n        insertContentRow(insertBody, doc, segmenting);\n        // A preset edit forces a full rebuild, so an unchanged doc's coverage is already\n        // correct; new docs have nothing to clear, which keeps cold builds linear.\n        if (existing.has(doc.relPath)) delPresetFiles.run(doc.relPath);\n        for (const presetName of doc.presets) insertPresetFile.run(doc.relPath, presetName);\n        for (const feature of features) feature.store?.(db, doc.relPath, doc.extracted[feature.name], delta);\n      }\n    }\n    for (const feature of features) feature.afterReconcile?.(db, delta);\n    db.exec('COMMIT');\n  } catch (err) {\n    db.exec('ROLLBACK');\n    throw err;\n  }\n\n  // Reconcile's own write-transaction duration, for open()'s derived busy_timeout (F):\n  // keep the observed max so a big watcher reconcile's lock hold is what the next open bounds its wait against.\n  const durationMs = Date.now() - txStart;\n  const prevRaw = getMeta(db, 'reconcile_max_ms');\n  // -1, not 0, so a genuinely 0ms first reconcile (sub-millisecond, common on a tiny tree)\n  // still gets recorded instead of losing to the \"nothing recorded yet\" default.\n  const prevMax = prevRaw === null ? -1 : Number(prevRaw);\n  if (durationMs > prevMax) setMeta(db, 'reconcile_max_ms', String(durationMs));\n\n  return { parsed: parsedDocs.length, warnings };\n}\n\n// Segment keys that moved between two feature signatures (see config.featureSignature's\n// format: global features, embed provider, tokenize, then one segment per preset).\nexport function changedSignatureKeys(before: string, after: string): Set<string> {\n  const keyOf = (part: string) => (part.startsWith('preset:') ? part.split(':').slice(0, 2).join(':') : part.split(':')[0]);\n  const parse = (sig: string) => new Map(sig.split('|').map((part) => [keyOf(part), part]));\n  const a = parse(before);\n  const b = parse(after);\n  const changed = new Set<string>();\n  for (const [key, val] of b) if (a.get(key) !== val) changed.add(key);\n  for (const key of a.keys()) if (!b.has(key)) changed.add(key);\n  return changed;\n}\n\n// Whether the embed segment only gained its resolved weight identity: same provider and\n// model, no identity recorded before, one now -- adopted into meta without a rebuild.\nexport function embedIdentityAdopted(before: string, after: string): boolean {\n  const embedPart = (sig: string) => sig.split('|').find((p) => p.startsWith('embed:'));\n  const b = embedPart(before);\n  const a = embedPart(after);\n  if (b === undefined || a === undefined) return false;\n  const at = a.indexOf('@');\n  return b.indexOf('@') === -1 && at !== -1 && a.slice(0, at) === b;\n}\n\n// Names what moved, for the rebuild notice.\nexport function signatureDiff(before: string, after: string): string {\n  const changed = changedSignatureKeys(before, after);\n  const label = (key: string) => (key === 'embed' ? 'embed settings' : key === 'tokenize' ? 'content tokenizer' : key.startsWith('preset:') ? `preset \"${key.slice(7)}\"` : 'features');\n  return changed.size === 0 ? 'features' : [...changed].map(label).join(', ');\n}\n\n// The tokenize-only rebuild in open(): content is dropped and repopulated from files already\n// listed in frontmatter, which itself is untouched. Frontmatter, links, sections, and\n// embeddings are file-derived and tokenizer-independent, so they survive. No feature extractors\n// run here -- doc.search (title/summary/text) is all content population needs. Returns\n// parseFile's per-file warnings (e.g. a bad date) so open() can surface them -- mtimes are\n// untouched, so reconcile() never reparses these files and would otherwise never emit them again.\nexport function rebuildContentTable(db: DatabaseSync, cfg: Config, baseDir: string): string[] {\n  const known = new Set((db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path));\n  const files = listFiles(cfg, baseDir).filter((f) => known.has(f.relPath));\n  const segmenting = contentTokenize(cfg) === undefined;\n  const insertBody = db.prepare(INSERT_CONTENT_SQL);\n  const warnings: string[] = [];\n  db.exec('BEGIN');\n  try {\n    for (const file of files) {\n      const { doc, warnings: fileWarnings } = parseFile(file);\n      warnings.push(...fileWarnings);\n      insertContentRow(insertBody, doc, segmenting);\n    }\n    db.exec('COMMIT');\n  } catch (err) {\n    db.exec('ROLLBACK');\n    throw err;\n  }\n  return warnings;\n}\n"],"names":["changedSignatureKeys","embedIdentityAdopted","rebuildContentTable","reconcile","signatureDiff","CORE_FRONTMATTER_COLUMNS","Set","MAX_FRONTMATTER_COLUMNS","INSERT_CONTENT_SQL","insertContentRow","insertBody","doc","segmenting","run","relPath","search","title","summary","text","segmentField","db","cfg","baseDir","files","listFiles","currentSet","map","f","existingRows","prepare","all","existing","Map","r","path","vanished","filter","has","toReparse","row","get","_mtime","mtimeMs","_size","size","length","parsed","warnings","features","activeFeatures","seenColumns","getColumns","newColumns","parsedDocs","report","progress","parsedCount","file","fileFeatures","feature","enabledForFile","parseFile","fileWarnings","tick","push","Object","keys","data","key","add","finish","allColumns","SenseError","writableColumns","c","RESERVED_COLUMNS","insertSql","quoteIdent","join","added","delta","reparsed","d","txStart","Date","now","exec","col","delBody","delPresetFiles","insertPresetFile","del","remove","insert","contentTokenize","undefined","values","ctimeMs","parseError","presets","presetName","store","extracted","name","afterReconcile","err","durationMs","prevRaw","getMeta","prevMax","Number","setMeta","String","before","after","keyOf","part","startsWith","split","slice","parse","sig","a","b","changed","val","embedPart","find","p","at","indexOf","label","known"],"mappings":";;;;;;;;;;;QA0KgBA;eAAAA;;QAaAC;eAAAA;;QAsBAC;eAAAA;;QAhLAC;eAAAA;;QAoKAC;eAAAA;;;uBA/LgB;wBACL;wBACI;0BAEN;sBAE8B;yBAC1B;wBAC4B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAEzD,6FAA6F;AAC7F,4EAA4E;AAC5E,IAAMC,2BAA2B,IAAIC,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;CAAe;AAE9F,8FAA8F;AAC9F,IAAMC,0BAA0B;AAEhC,0FAA0F;AAC1F,gDAAgD;AAChD,IAAMC,qBAAqB;AAE3B,0FAA0F;AAC1F,2FAA2F;AAC3F,SAASC,iBAAiBC,UAAyB,EAAEC,GAAc,EAAEC,UAAmB;IACtFF,WAAWG,GAAG,CAACF,IAAIG,OAAO,EAAEH,IAAII,MAAM,CAACC,KAAK,EAAEL,IAAII,MAAM,CAACE,OAAO,EAAEN,IAAII,MAAM,CAACG,IAAI,EAAEP,IAAIG,OAAO,EAAEF,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACC,KAAK,IAAI,IAAIJ,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACE,OAAO,IAAI,IAAIL,aAAaO,IAAAA,uBAAY,EAACR,IAAII,MAAM,CAACG,IAAI,IAAI;AACrP;AAEO,SAASf,UAAUiB,EAAgB,EAAEC,GAAW,EAAEC,OAAe;IACtE,IAAMC,QAAQC,IAAAA,iBAAS,EAACH,KAAKC;IAC7B,IAAMG,aAAa,IAAInB,IAAIiB,MAAMG,GAAG,CAAC,SAACC;eAAMA,EAAEb,OAAO;;IAErD,IAAMc,eAAeR,GAAGS,OAAO,CAAC,qDAAqDC,GAAG;IAKxF,IAAMC,WAAW,IAAIC,IAAIJ,aAAaF,GAAG,CAAC,SAACO;eAAM;YAACA,EAAEC,IAAI;YAAED;SAAE;;IAC5D,IAAME,WAAWP,aAAaQ,MAAM,CAAC,SAACH;eAAM,CAACR,WAAWY,GAAG,CAACJ,EAAEC,IAAI;OAAGR,GAAG,CAAC,SAACO;eAAMA,EAAEC,IAAI;;IAEtF,IAAMI,YAAYf,MAAMa,MAAM,CAAC,SAACT;QAC9B,IAAMY,MAAMR,SAASS,GAAG,CAACb,EAAEb,OAAO;QAClC,OAAO,CAACyB,OAAOA,IAAIE,MAAM,KAAKd,EAAEe,OAAO,IAAIH,IAAII,KAAK,KAAKhB,EAAEiB,IAAI;IACjE;IAEA,IAAIT,SAASU,MAAM,KAAK,KAAKP,UAAUO,MAAM,KAAK,GAAG,OAAO;QAAEC,QAAQ;QAAGC,UAAU,EAAE;IAAC;IAEtF,IAAMC,WAAWC,IAAAA,wBAAc,EAAC5B;IAChC,IAAM6B,cAAcC,IAAAA,oBAAU,EAAC/B;IAC/B,IAAMgC,aAAuB,EAAE;IAC/B,IAAMC,aAA0B,EAAE;IAClC,IAAMN,WAAqB,EAAE;IAE7B,oFAAoF;IACpF,uDAAuD;IACvD,IAAMO,SAASC,IAAAA,oBAAQ,EAAC,mBAAmBjB,UAAUO,MAAM;IAC3D,IAAIW,cAAc;QACb,kCAAA,2BAAA;;;YAAA,IAAMC,OAAN;gBAMHV;YALA,sFAAsF;YACtF,mEAAmE;YACnE,IAAMW,eAAeV,SAASZ,MAAM,CAAC,SAACuB;uBAAY,CAACA,QAAQC,cAAc,IAAID,QAAQC,cAAc,CAACvC,KAAKoC;;YACzG,IAAwCI,aAAAA,IAAAA,iBAAS,EAACJ,MAAMC,eAAhD/C,MAAgCkD,WAAhClD,KAAKoC,AAAUe,eAAiBD,WAA3Bd;YACbO,OAAOS,IAAI,CAAC,EAAEP;YACdT,CAAAA,YAAAA,UAASiB,IAAI,OAAbjB,WAAc,qBAAGe;gBACZ,kCAAA,2BAAA;;gBAAL,QAAK,YAAaG,OAAOC,IAAI,CAACvD,IAAIwD,IAAI,sBAAjC,UAAA,6BAAA,SAAA,yBAAA,iCAAoC;oBAApC,IAAMC,MAAN;oBACH,IAAI,CAAClB,YAAYb,GAAG,CAAC+B,MAAM;wBACzBlB,YAAYmB,GAAG,CAACD;wBAChBhB,WAAWY,IAAI,CAACI;oBAClB;gBACF;;gBALK;gBAAA;;;yBAAA,6BAAA;wBAAA;;;wBAAA;8BAAA;;;;YAMLf,WAAWW,IAAI,CAACrD;QAClB;QAdA,QAAK,YAAc2B,8BAAd,SAAA,6BAAA,QAAA,yBAAA;;QAAA;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;IAeLgB,OAAOgB,MAAM;IAEb,IAAMC,aAAc,qBAAGrB;IACvB,uEAAuE;IACvE,sGAAsG;IACtG,IAAIqB,WAAW1B,MAAM,GAAGtC,yBAAyB;QAC/C,MAAM,IAAIiE,oBAAU,CAClB,gBACA,AAAC,0BAAuHjE,OAA9FgE,WAAW1B,MAAM,EAAC,8EAAoG,OAAxBtC,yBAAwB;IAEpJ;IACA,0FAA0F;IAC1F,sEAAsE;IACtE,IAAMkE,kBAAkBF,WAAWnC,MAAM,CAAC,SAACsC;eAAMrE,yBAAyBgC,GAAG,CAACqC,MAAM,CAACC,wBAAgB,CAACtC,GAAG,CAACqC;;IAC1G,sFAAsF;IACtF,gDAAgD;IAChD,IAAME,YAAY,AAAC,4BAAkFH,OAAvDA,gBAAgB/C,GAAG,CAACmD,oBAAU,EAAEC,IAAI,CAAC,OAAM,cAA4FL,OAAhFA,gBAAgB/C,GAAG,CAAC;eAAM;OAAKoD,IAAI,CAAC,OAAM,wCAGjI,OAHuKL,gBAClLrC,MAAM,CAAC,SAACsC;eAAMA,MAAM;OACpBhD,GAAG,CAAC,SAACgD;eAAM,AAAC,GAA8BG,OAA5BA,IAAAA,oBAAU,EAACH,IAAG,gBAA4B,OAAdG,IAAAA,oBAAU,EAACH;OACrDI,IAAI,CAAC;IAER,IAAMC,QAAQzC,UAAUF,MAAM,CAAC,SAACT;eAAM,CAACI,SAASM,GAAG,CAACV,EAAEb,OAAO;OAAGY,GAAG,CAAC,SAACC;eAAMA,EAAEb,OAAO;;IACpF,IAAMkE,QAAwB;QAAEzD,OAAAA;QAAO0D,UAAU5B,WAAW3B,GAAG,CAAC,SAACwD;mBAAMA,EAAEpE,OAAO;;QAAGiE,OAAAA;QAAO5C,UAAAA;IAAS;IAEnG,IAAMgD,UAAUC,KAAKC,GAAG;IACxBjE,GAAGkE,IAAI,CAAC;IACR,IAAI;YAkD8B3B;YAjD3B,mCAAA,4BAAA;;YAAL,QAAK,aAAaP,+BAAb,UAAA,8BAAA,SAAA,0BAAA;gBAAA,IAAMmC,MAAN;gBAAyBnE,GAAGkE,IAAI,CAAC,AAAC,sCAAqD,OAAhBT,IAAAA,oBAAU,EAACU;;;YAAlF;YAAA;;;qBAAA,8BAAA;oBAAA;;;oBAAA;0BAAA;;;;QACL,6EAA6E;QAC7E,kFAAkF;QAClF,+EAA+E;QAC/E,IAAMC,UAAUpE,GAAGS,OAAO,CAAC;QAC3B,IAAM4D,iBAAiBrE,GAAGS,OAAO,CAAC;QAClC,IAAM6D,mBAAmBtE,GAAGS,OAAO,CAAC;QACpC,IAAIM,SAASU,MAAM,GAAG,GAAG;YACvB,IAAM8C,MAAMvE,GAAGS,OAAO,CAAC;gBAClB,mCAAA,4BAAA;;gBAAL,QAAK,aAAcM,6BAAd,UAAA,8BAAA,SAAA,0BAAA,kCAAwB;oBAAxB,IAAMD,OAAN;wBAM6ByB;oBALhC,kFAAkF;oBAClF,qEAAqE;oBACrE6B,QAAQ3E,GAAG,CAACqB;oBACZyD,IAAI9E,GAAG,CAACqB;oBACRuD,eAAe5E,GAAG,CAACqB;wBACd,mCAAA,4BAAA;;wBAAL,QAAK,aAAiBc,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMW,UAAN;6BAA2BA,kBAAAA,QAAQiC,MAAM,cAAdjC,sCAAAA,qBAAAA,SAAiBvC,IAAIc,MAAM8C;;;wBAAtD;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;gBACP;;gBAPK;gBAAA;;;yBAAA,8BAAA;wBAAA;;;wBAAA;8BAAA;;;;QAQP;QACA,IAAI3B,WAAWR,MAAM,GAAG,GAAG;YACzB,IAAMgD,SAASzE,GAAGS,OAAO,CAAC+C;YAC1B,IAAMlE,aAAaU,GAAGS,OAAO,CAACrB;YAC9B,wFAAwF;YACxF,+EAA+E;YAC/E,IAAMI,aAAakF,IAAAA,wBAAe,EAACzE,SAAS0E;gBACvC,mCAAA,4BAAA;;;oBAAA,IAAMpF,MAAN;wBAUH,uFAAuF;oBACvFkF;wBAWgClC;oBArBhC,IAAMqC,SAASvB,gBAAgB/C,GAAG,CAAC,SAAC6D;4BAO3B5E;wBANP,IAAI4E,QAAQ,QAAQ,OAAO5E,IAAIG,OAAO;wBACtC,IAAIyE,QAAQ,UAAU,OAAO5E,IAAI+B,OAAO;wBACxC,IAAI6C,QAAQ,UAAU,OAAO5E,IAAIsF,OAAO;wBACxC,IAAIV,QAAQ,SAAS,OAAO5E,IAAIiC,IAAI;wBACpC,mFAAmF;wBACnF,IAAI2C,QAAQ,gBAAgB,OAAO5E,IAAIuF,UAAU;wBACjD,QAAOvF,gBAAAA,IAAIwD,IAAI,CAACoB,IAAI,cAAb5E,2BAAAA,gBAAiB;oBAC1B;oBAEAkF,CAAAA,UAAAA,QAAOhF,GAAG,OAAVgF,SAAW,qBAAGG;oBACd,oFAAoF;oBACpF,IAAIjE,SAASM,GAAG,CAAC1B,IAAIG,OAAO,GAAG;4BAEG6C;wBADhC6B,QAAQ3E,GAAG,CAACF,IAAIG,OAAO;4BAClB,kCAAA,2BAAA;;4BAAL,QAAK,YAAiBkC,6BAAjB,SAAA,6BAAA,QAAA,yBAAA;gCAAA,IAAMW,UAAN;iCAA2BA,kBAAAA,QAAQiC,MAAM,cAAdjC,sCAAAA,qBAAAA,SAAiBvC,IAAIT,IAAIG,OAAO,EAAEkE;;;4BAA7D;4BAAA;;;qCAAA,6BAAA;oCAAA;;;oCAAA;0CAAA;;;;oBACP;oBACAvE,iBAAiBC,YAAYC,KAAKC;oBAClC,iFAAiF;oBACjF,2EAA2E;oBAC3E,IAAImB,SAASM,GAAG,CAAC1B,IAAIG,OAAO,GAAG2E,eAAe5E,GAAG,CAACF,IAAIG,OAAO;wBACxD,mCAAA,4BAAA;;wBAAL,QAAK,aAAoBH,IAAIwF,OAAO,qBAA/B,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMC,aAAN;4BAAiCV,iBAAiB7E,GAAG,CAACF,IAAIG,OAAO,EAAEsF;;;wBAAnE;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;wBACA,mCAAA,4BAAA;;wBAAL,QAAK,aAAiBpD,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;4BAAA,IAAMW,WAAN;6BAA2BA,iBAAAA,SAAQ0C,KAAK,cAAb1C,qCAAAA,oBAAAA,UAAgBvC,IAAIT,IAAIG,OAAO,EAAEH,IAAI2F,SAAS,CAAC3C,SAAQ4C,IAAI,CAAC,EAAEvB;;;wBAAzF;wBAAA;;;iCAAA,8BAAA;gCAAA;;;gCAAA;sCAAA;;;;gBACP;gBAvBA,QAAK,aAAa3B,+BAAb,UAAA,8BAAA,SAAA,0BAAA;;gBAAA;gBAAA;;;yBAAA,8BAAA;wBAAA;;;wBAAA;8BAAA;;;;QAwBP;YACK,mCAAA,4BAAA;;YAAL,QAAK,aAAiBL,6BAAjB,UAAA,8BAAA,SAAA,0BAAA;gBAAA,IAAMW,WAAN;iBAA2BA,0BAAAA,SAAQ6C,cAAc,cAAtB7C,8CAAAA,6BAAAA,UAAyBvC,IAAI4D;;;YAAxD;YAAA;;;qBAAA,8BAAA;oBAAA;;;oBAAA;0BAAA;;;;QACL5D,GAAGkE,IAAI,CAAC;IACV,EAAE,OAAOmB,KAAK;QACZrF,GAAGkE,IAAI,CAAC;QACR,MAAMmB;IACR;IAEA,qFAAqF;IACrF,8GAA8G;IAC9G,IAAMC,aAAatB,KAAKC,GAAG,KAAKF;IAChC,IAAMwB,UAAUC,IAAAA,iBAAO,EAACxF,IAAI;IAC5B,yFAAyF;IACzF,+EAA+E;IAC/E,IAAMyF,UAAUF,YAAY,OAAO,CAAC,IAAIG,OAAOH;IAC/C,IAAID,aAAaG,SAASE,IAAAA,iBAAO,EAAC3F,IAAI,oBAAoB4F,OAAON;IAEjE,OAAO;QAAE5D,QAAQO,WAAWR,MAAM;QAAEE,UAAAA;IAAS;AAC/C;AAIO,SAAS/C,qBAAqBiH,MAAc,EAAEC,KAAa;IAChE,IAAMC,QAAQ,eAACC;eAAkBA,KAAKC,UAAU,CAAC,aAAaD,KAAKE,KAAK,CAAC,KAAKC,KAAK,CAAC,GAAG,GAAGzC,IAAI,CAAC,OAAOsC,KAAKE,KAAK,CAAC,IAAI,CAAC,EAAE;;IACxH,IAAME,QAAQ,eAACC;eAAgB,IAAIzF,IAAIyF,IAAIH,KAAK,CAAC,KAAK5F,GAAG,CAAC,SAAC0F;mBAAS;gBAACD,MAAMC;gBAAOA;aAAK;;;IACvF,IAAMM,IAAIF,MAAMP;IAChB,IAAMU,IAAIH,MAAMN;IAChB,IAAMU,UAAU,IAAItH;QACf,kCAAA,2BAAA;;QAAL,QAAK,YAAoBqH,sBAApB,SAAA,6BAAA,QAAA,yBAAA;YAAA,mCAAA,iBAAOvD,sBAAKyD;YAAW,IAAIH,EAAElF,GAAG,CAAC4B,SAASyD,KAAKD,QAAQvD,GAAG,CAACD;;;QAA3D;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;QACA,mCAAA,4BAAA;;QAAL,QAAK,aAAasD,EAAExD,IAAI,uBAAnB,UAAA,8BAAA,SAAA,0BAAA;YAAA,IAAME,OAAN;YAAuB,IAAI,CAACuD,EAAEtF,GAAG,CAAC+B,OAAMwD,QAAQvD,GAAG,CAACD;;;QAApD;QAAA;;;iBAAA,8BAAA;gBAAA;;;gBAAA;sBAAA;;;;IACL,OAAOwD;AACT;AAIO,SAAS3H,qBAAqBgH,MAAc,EAAEC,KAAa;IAChE,IAAMY,YAAY,mBAACL;eAAgBA,IAAIH,KAAK,CAAC,KAAKS,IAAI,CAAC,SAACC;mBAAMA,EAAEX,UAAU,CAAC;;;IAC3E,IAAMM,IAAIG,UAAUb;IACpB,IAAMS,IAAII,UAAUZ;IACpB,IAAIS,MAAM5B,aAAa2B,MAAM3B,WAAW,OAAO;IAC/C,IAAMkC,KAAKP,EAAEQ,OAAO,CAAC;IACrB,OAAOP,EAAEO,OAAO,CAAC,SAAS,CAAC,KAAKD,OAAO,CAAC,KAAKP,EAAEH,KAAK,CAAC,GAAGU,QAAQN;AAClE;AAGO,SAASvH,cAAc6G,MAAc,EAAEC,KAAa;IACzD,IAAMU,UAAU5H,qBAAqBiH,QAAQC;IAC7C,IAAMiB,QAAQ,eAAC/D;eAAiBA,QAAQ,UAAU,mBAAmBA,QAAQ,aAAa,sBAAsBA,IAAIiD,UAAU,CAAC,aAAa,AAAC,WAAuB,OAAbjD,IAAImD,KAAK,CAAC,IAAG,OAAK;;IACzK,OAAOK,QAAQhF,IAAI,KAAK,IAAI,aAAa,AAAC,qBAAGgF,SAASlG,GAAG,CAACyG,OAAOrD,IAAI,CAAC;AACxE;AAQO,SAAS5E,oBAAoBkB,EAAgB,EAAEC,GAAW,EAAEC,OAAe;IAChF,IAAM8G,QAAQ,IAAI9H,IAAI,AAACc,GAAGS,OAAO,CAAC,kCAAkCC,GAAG,GAA+BJ,GAAG,CAAC,SAACO;eAAMA,EAAEC,IAAI;;IACvH,IAAMX,QAAQC,IAAAA,iBAAS,EAACH,KAAKC,SAASc,MAAM,CAAC,SAACT;eAAMyG,MAAM/F,GAAG,CAACV,EAAEb,OAAO;;IACvE,IAAMF,aAAakF,IAAAA,wBAAe,EAACzE,SAAS0E;IAC5C,IAAMrF,aAAaU,GAAGS,OAAO,CAACrB;IAC9B,IAAMuC,WAAqB,EAAE;IAC7B3B,GAAGkE,IAAI,CAAC;IACR,IAAI;YACG,kCAAA,2BAAA;;YAAL,QAAK,YAAc/D,0BAAd,SAAA,6BAAA,QAAA,yBAAA,iCAAqB;gBAArB,IAAMkC,OAAN;oBAEHV;gBADA,IAAwCc,aAAAA,IAAAA,iBAAS,EAACJ,OAA1C9C,MAAgCkD,WAAhClD,KAAKoC,AAAUe,eAAiBD,WAA3Bd;gBACbA,CAAAA,YAAAA,UAASiB,IAAI,OAAbjB,WAAc,qBAAGe;gBACjBrD,iBAAiBC,YAAYC,KAAKC;YACpC;;YAJK;YAAA;;;qBAAA,6BAAA;oBAAA;;;oBAAA;0BAAA;;;;QAKLQ,GAAGkE,IAAI,CAAC;IACV,EAAE,OAAOmB,KAAK;QACZrF,GAAGkE,IAAI,CAAC;QACR,MAAMmB;IACR;IACA,OAAO1D;AACT"}

package/dist/cjs/embed/cohere.d.cts

@@ -0,0 +1,2 @@
+import type { EmbedProvider } from './types.js';
+export declare function cohereProvider(model: string, url: string | undefined, keyEnv: string | undefined): Promise<EmbedProvider>;

package/dist/cjs/embed/cohere.d.ts

@@ -0,0 +1,2 @@
+import type { EmbedProvider } from './types.js';
+export declare function cohereProvider(model: string, url: string | undefined, keyEnv: string | undefined): Promise<EmbedProvider>;

package/dist/cjs/embed/cohere.js

@@ -0,0 +1,238 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", {
+    value: true
+});
+Object.defineProperty(exports, "cohereProvider", {
+    enumerable: true,
+    get: function() {
+        return cohereProvider;
+    }
+});
+var _errorsts = require("../errors.js");
+var _httpts = require("./http.js");
+function asyncGeneratorStep(gen, resolve, reject, _next, _throw, key, arg) {
+    try {
+        var info = gen[key](arg);
+        var value = info.value;
+    } catch (error) {
+        reject(error);
+        return;
+    }
+    if (info.done) resolve(value);
+    else Promise.resolve(value).then(_next, _throw);
+}
+function _async_to_generator(fn) {
+    return function() {
+        var self = this, args = arguments;
+        return new Promise(function(resolve, reject) {
+            var gen = fn.apply(self, args);
+            function _next(value) {
+                asyncGeneratorStep(gen, resolve, reject, _next, _throw, "next", value);
+            }
+            function _throw(err) {
+                asyncGeneratorStep(gen, resolve, reject, _next, _throw, "throw", err);
+            }
+            _next(undefined);
+        });
+    };
+}
+function _ts_generator(thisArg, body) {
+    var f, y, t, _ = {
+        label: 0,
+        sent: function() {
+            if (t[0] & 1) throw t[1];
+            return t[1];
+        },
+        trys: [],
+        ops: []
+    }, g = Object.create((typeof Iterator === "function" ? Iterator : Object).prototype), d = Object.defineProperty;
+    return d(g, "next", {
+        value: verb(0)
+    }), d(g, "throw", {
+        value: verb(1)
+    }), d(g, "return", {
+        value: verb(2)
+    }), typeof Symbol === "function" && d(g, Symbol.iterator, {
+        value: function() {
+            return this;
+        }
+    }), g;
+    function verb(n) {
+        return function(v) {
+            return step([
+                n,
+                v
+            ]);
+        };
+    }
+    function step(op) {
+        if (f) throw new TypeError("Generator is already executing.");
+        while(g && (g = 0, op[0] && (_ = 0)), _)try {
+            if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
+            if (y = 0, t) op = [
+                op[0] & 2,
+                t.value
+            ];
+            switch(op[0]){
+                case 0:
+                case 1:
+                    t = op;
+                    break;
+                case 4:
+                    _.label++;
+                    return {
+                        value: op[1],
+                        done: false
+                    };
+                case 5:
+                    _.label++;
+                    y = op[1];
+                    op = [
+                        0
+                    ];
+                    continue;
+                case 7:
+                    op = _.ops.pop();
+                    _.trys.pop();
+                    continue;
+                default:
+                    if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) {
+                        _ = 0;
+                        continue;
+                    }
+                    if (op[0] === 3 && (!t || op[1] > t[0] && op[1] < t[3])) {
+                        _.label = op[1];
+                        break;
+                    }
+                    if (op[0] === 6 && _.label < t[1]) {
+                        _.label = t[1];
+                        t = op;
+                        break;
+                    }
+                    if (t && _.label < t[2]) {
+                        _.label = t[2];
+                        _.ops.push(op);
+                        break;
+                    }
+                    if (t[2]) _.ops.pop();
+                    _.trys.pop();
+                    continue;
+            }
+            op = body.call(thisArg, _);
+        } catch (e) {
+            op = [
+                6,
+                e
+            ];
+            y = 0;
+        } finally{
+            f = t = 0;
+        }
+        if (op[0] & 5) throw op[1];
+        return {
+            value: op[0] ? op[1] : void 0,
+            done: true
+        };
+    }
+}
+var BATCH_CAP = 96;
+function cohereProvider(model, url, keyEnv) {
+    return _async_to_generator(function() {
+        var base, headers, key, dims;
+        function post(texts, inputType) {
+            return _async_to_generator(function() {
+                var res, body, vectors;
+                return _ts_generator(this, function(_state) {
+                    switch(_state.label){
+                        case 0:
+                            return [
+                                4,
+                                (0, _httpts.fetchWithRetry)("".concat(base, "/v2/embed"), {
+                                    method: 'POST',
+                                    headers: headers,
+                                    body: JSON.stringify({
+                                        model: model,
+                                        input_type: inputType,
+                                        texts: texts,
+                                        embedding_types: [
+                                            'float'
+                                        ]
+                                    })
+                                })
+                            ];
+                        case 1:
+                            res = _state.sent();
+                            if (!res.ok) throw new _errorsts.SenseError('EMBED_MODEL', "".concat(base, "/v2/embed -> HTTP ").concat(res.status));
+                            return [
+                                4,
+                                res.json()
+                            ];
+                        case 2:
+                            body = _state.sent();
+                            // Cohere returns one vector per text in order, with no index field to re-sort by; the
+                            // count is what a caller mapping positionally still has to be able to trust.
+                            vectors = body.embeddings.float;
+                            if (vectors.length !== texts.length) throw new _errorsts.SenseError('EMBED_MODEL', "".concat(base, "/v2/embed returned ").concat(vectors.length, " embeddings for ").concat(texts.length, " inputs"));
+                            return [
+                                2,
+                                vectors.map(function(v) {
+                                    return Float32Array.from(v);
+                                })
+                            ];
+                    }
+                });
+            })();
+        }
+        return _ts_generator(this, function(_state) {
+            switch(_state.label){
+                case 0:
+                    base = (url !== null && url !== void 0 ? url : 'https://api.cohere.com').replace(/\/+$/, '');
+                    headers = {
+                        'content-type': 'application/json'
+                    };
+                    key = keyEnv ? process.env[keyEnv] : undefined;
+                    if (key) headers.authorization = "Bearer ".concat(key);
+                    return [
+                        4,
+                        post([
+                            'dimension probe'
+                        ], 'search_query')
+                    ];
+                case 1:
+                    dims = _state.sent()[0].length;
+                    return [
+                        2,
+                        {
+                            id: "cohere:".concat(base, ":").concat(model),
+                            dims: dims,
+                            batchCap: BATCH_CAP,
+                            embedDocuments: function embedDocuments(texts) {
+                                return post(texts, 'search_document');
+                            },
+                            embedQuery: function embedQuery(text) {
+                                return _async_to_generator(function() {
+                                    return _ts_generator(this, function(_state) {
+                                        switch(_state.label){
+                                            case 0:
+                                                return [
+                                                    4,
+                                                    post([
+                                                        text
+                                                    ], 'search_query')
+                                                ];
+                                            case 1:
+                                                return [
+                                                    2,
+                                                    _state.sent()[0]
+                                                ];
+                                        }
+                                    });
+                                })();
+                            }
+                        }
+                    ];
+            }
+        });
+    })();
+}
+/* CJS INTEROP */ if (exports.__esModule && exports.default) { try { Object.defineProperty(exports.default, '__esModule', { value: true }); for (var key in exports) { exports.default[key] = exports[key]; } } catch (_) {}; module.exports = exports.default; }

package/dist/cjs/embed/cohere.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/embed/cohere.ts"],"sourcesContent":["import { SenseError } from '../errors.ts';\nimport { fetchWithRetry } from './http.ts';\nimport type { EmbedProvider } from './types.ts';\n\nconst BATCH_CAP = 96;\n\n// Cohere's native /v2/embed: input_type distinguishes doc vs query embeddings, which the\n// OpenAI-compatible shape cannot express.\nexport async function cohereProvider(model: string, url: string | undefined, keyEnv: string | undefined): Promise<EmbedProvider> {\n  const base = (url ?? 'https://api.cohere.com').replace(/\\/+$/, '');\n  const headers: Record<string, string> = { 'content-type': 'application/json' };\n  const key = keyEnv ? process.env[keyEnv] : undefined;\n  if (key) headers.authorization = `Bearer ${key}`;\n\n  async function post(texts: string[], inputType: 'search_document' | 'search_query'): Promise<Float32Array[]> {\n    const res = await fetchWithRetry(`${base}/v2/embed`, { method: 'POST', headers, body: JSON.stringify({ model, input_type: inputType, texts, embedding_types: ['float'] }) });\n    if (!res.ok) throw new SenseError('EMBED_MODEL', `${base}/v2/embed -> HTTP ${res.status}`);\n    const body = (await res.json()) as { embeddings: { float: number[][] } };\n    // Cohere returns one vector per text in order, with no index field to re-sort by; the\n    // count is what a caller mapping positionally still has to be able to trust.\n    const vectors = body.embeddings.float;\n    if (vectors.length !== texts.length) throw new SenseError('EMBED_MODEL', `${base}/v2/embed returned ${vectors.length} embeddings for ${texts.length} inputs`);\n    return vectors.map((v) => Float32Array.from(v));\n  }\n\n  const dims = (await post(['dimension probe'], 'search_query'))[0].length;\n  return { id: `cohere:${base}:${model}`, dims, batchCap: BATCH_CAP, embedDocuments: (texts) => post(texts, 'search_document'), embedQuery: async (text) => (await post([text], 'search_query'))[0] };\n}\n"],"names":["cohereProvider","BATCH_CAP","model","url","keyEnv","base","headers","key","dims","post","texts","inputType","res","body","vectors","fetchWithRetry","method","JSON","stringify","input_type","embedding_types","ok","SenseError","status","json","embeddings","float","length","map","v","Float32Array","from","replace","process","env","undefined","authorization","id","batchCap","embedDocuments","embedQuery","text"],"mappings":";;;;+BAQsBA;;;eAAAA;;;wBARK;sBACI;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAG/B,IAAMC,YAAY;AAIX,SAAeD,eAAeE,KAAa,EAAEC,GAAuB,EAAEC,MAA0B;;YAC/FC,MACAC,SACAC,KAcAC;QAXN,SAAeC,KAAKC,KAAe,EAAEC,SAA6C;;oBAC1EC,KAEAC,MAGAC;;;;4BALM;;gCAAMC,IAAAA,sBAAc,EAAC,AAAC,GAAO,OAALV,MAAK,cAAY;oCAAEW,QAAQ;oCAAQV,SAAAA;oCAASO,MAAMI,KAAKC,SAAS,CAAC;wCAAEhB,OAAAA;wCAAOiB,YAAYR;wCAAWD,OAAAA;wCAAOU,eAAe;4CAAG;;oCAAS;gCAAG;;;4BAApKR,MAAM;4BACZ,IAAI,CAACA,IAAIS,EAAE,EAAE,MAAM,IAAIC,oBAAU,CAAC,eAAe,AAAC,GAA2BV,OAAzBP,MAAK,sBAA+B,OAAXO,IAAIW,MAAM;4BACzE;;gCAAMX,IAAIY,IAAI;;;4BAAtBX,OAAQ;4BACd,sFAAsF;4BACtF,6EAA6E;4BACvEC,UAAUD,KAAKY,UAAU,CAACC,KAAK;4BACrC,IAAIZ,QAAQa,MAAM,KAAKjB,MAAMiB,MAAM,EAAE,MAAM,IAAIL,oBAAU,CAAC,eAAe,AAAC,GAA4BR,OAA1BT,MAAK,uBAAsDK,OAAjCI,QAAQa,MAAM,EAAC,oBAA+B,OAAbjB,MAAMiB,MAAM,EAAC;4BACpJ;;gCAAOb,QAAQc,GAAG,CAAC,SAACC;2CAAMC,aAAaC,IAAI,CAACF;;;;;YAC9C;;;;;oBAdMxB,OAAO,CAACF,gBAAAA,iBAAAA,MAAO,0BAA0B6B,OAAO,CAAC,QAAQ;oBACzD1B,UAAkC;wBAAE,gBAAgB;oBAAmB;oBACvEC,MAAMH,SAAS6B,QAAQC,GAAG,CAAC9B,OAAO,GAAG+B;oBAC3C,IAAI5B,KAAKD,QAAQ8B,aAAa,GAAG,AAAC,UAAa,OAAJ7B;oBAa7B;;wBAAME;4BAAM;2BAAoB;;;oBAAxCD,OAAO,AAAC,aAAgD,CAAC,EAAE,CAACmB,MAAM;oBACxE;;wBAAO;4BAAEU,IAAI,AAAC,UAAiBnC,OAARG,MAAK,KAAS,OAANH;4BAASM,MAAAA;4BAAM8B,UAAUrC;4BAAWsC,gBAAgB,SAAhBA,eAAiB7B;uCAAUD,KAAKC,OAAO;;4BAAoB8B,YAAY,SAAZA,WAAmBC;;;;;gDAAU;;oDAAMhC;wDAAMgC;uDAAO;;;;;oDAAnB,aAAmC,CAAC,EAAE;;;;;;wBAAC;;;;IACpM"}

package/dist/cjs/embed/http.d.cts

@@ -0,0 +1,2 @@
+export declare function fetchWithRetry(url: string, init: RequestInit, baseDelayMs?: number): Promise<Response>;
+export declare function probeReachable(url: string, timeoutMs?: number): Promise<boolean>;

package/dist/cjs/embed/http.d.ts

@@ -0,0 +1,2 @@
+export declare function fetchWithRetry(url: string, init: RequestInit, baseDelayMs?: number): Promise<Response>;
+export declare function probeReachable(url: string, timeoutMs?: number): Promise<boolean>;

package/dist/cjs/embed/http.js

@@ -0,0 +1,258 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", {
+    value: true
+});
+function _export(target, all) {
+    for(var name in all)Object.defineProperty(target, name, {
+        enumerable: true,
+        get: Object.getOwnPropertyDescriptor(all, name).get
+    });
+}
+_export(exports, {
+    get fetchWithRetry () {
+        return fetchWithRetry;
+    },
+    get probeReachable () {
+        return probeReachable;
+    }
+});
+function asyncGeneratorStep(gen, resolve, reject, _next, _throw, key, arg) {
+    try {
+        var info = gen[key](arg);
+        var value = info.value;
+    } catch (error) {
+        reject(error);
+        return;
+    }
+    if (info.done) resolve(value);
+    else Promise.resolve(value).then(_next, _throw);
+}
+function _async_to_generator(fn) {
+    return function() {
+        var self = this, args = arguments;
+        return new Promise(function(resolve, reject) {
+            var gen = fn.apply(self, args);
+            function _next(value) {
+                asyncGeneratorStep(gen, resolve, reject, _next, _throw, "next", value);
+            }
+            function _throw(err) {
+                asyncGeneratorStep(gen, resolve, reject, _next, _throw, "throw", err);
+            }
+            _next(undefined);
+        });
+    };
+}
+function _ts_generator(thisArg, body) {
+    var f, y, t, _ = {
+        label: 0,
+        sent: function() {
+            if (t[0] & 1) throw t[1];
+            return t[1];
+        },
+        trys: [],
+        ops: []
+    }, g = Object.create((typeof Iterator === "function" ? Iterator : Object).prototype), d = Object.defineProperty;
+    return d(g, "next", {
+        value: verb(0)
+    }), d(g, "throw", {
+        value: verb(1)
+    }), d(g, "return", {
+        value: verb(2)
+    }), typeof Symbol === "function" && d(g, Symbol.iterator, {
+        value: function() {
+            return this;
+        }
+    }), g;
+    function verb(n) {
+        return function(v) {
+            return step([
+                n,
+                v
+            ]);
+        };
+    }
+    function step(op) {
+        if (f) throw new TypeError("Generator is already executing.");
+        while(g && (g = 0, op[0] && (_ = 0)), _)try {
+            if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
+            if (y = 0, t) op = [
+                op[0] & 2,
+                t.value
+            ];
+            switch(op[0]){
+                case 0:
+                case 1:
+                    t = op;
+                    break;
+                case 4:
+                    _.label++;
+                    return {
+                        value: op[1],
+                        done: false
+                    };
+                case 5:
+                    _.label++;
+                    y = op[1];
+                    op = [
+                        0
+                    ];
+                    continue;
+                case 7:
+                    op = _.ops.pop();
+                    _.trys.pop();
+                    continue;
+                default:
+                    if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) {
+                        _ = 0;
+                        continue;
+                    }
+                    if (op[0] === 3 && (!t || op[1] > t[0] && op[1] < t[3])) {
+                        _.label = op[1];
+                        break;
+                    }
+                    if (op[0] === 6 && _.label < t[1]) {
+                        _.label = t[1];
+                        t = op;
+                        break;
+                    }
+                    if (t && _.label < t[2]) {
+                        _.label = t[2];
+                        _.ops.push(op);
+                        break;
+                    }
+                    if (t[2]) _.ops.pop();
+                    _.trys.pop();
+                    continue;
+            }
+            op = body.call(thisArg, _);
+        } catch (e) {
+            op = [
+                6,
+                e
+            ];
+            y = 0;
+        } finally{
+            f = t = 0;
+        }
+        if (op[0] & 5) throw op[1];
+        return {
+            value: op[0] ? op[1] : void 0,
+            done: true
+        };
+    }
+}
+var MAX_ATTEMPTS = 3;
+function sleep(ms) {
+    return new Promise(function(resolve) {
+        return setTimeout(resolve, ms);
+    });
+}
+function fetchWithRetry(url, init) {
+    var baseDelayMs = arguments.length > 2 && arguments[2] !== void 0 ? arguments[2] : 200;
+    return _async_to_generator(function() {
+        var attempt, res, retryAfter, err;
+        return _ts_generator(this, function(_state) {
+            switch(_state.label){
+                case 0:
+                    attempt = 0;
+                    _state.label = 1;
+                case 1:
+                    if (!(attempt < MAX_ATTEMPTS)) return [
+                        3,
+                        8
+                    ];
+                    _state.label = 2;
+                case 2:
+                    _state.trys.push([
+                        2,
+                        5,
+                        ,
+                        7
+                    ]);
+                    return [
+                        4,
+                        fetch(url, init)
+                    ];
+                case 3:
+                    res = _state.sent();
+                    if (res.ok || res.status !== 429 && res.status < 500 || attempt === MAX_ATTEMPTS - 1) return [
+                        2,
+                        res
+                    ];
+                    retryAfter = Number(res.headers.get('retry-after'));
+                    // Capped: a server naming a huge Retry-After must not stall an embed run for it.
+                    return [
+                        4,
+                        sleep(retryAfter > 0 ? Math.min(retryAfter * 1000, 10000) : baseDelayMs * Math.pow(2, attempt))
+                    ];
+                case 4:
+                    _state.sent();
+                    return [
+                        3,
+                        7
+                    ];
+                case 5:
+                    err = _state.sent();
+                    if (attempt === MAX_ATTEMPTS - 1) throw err;
+                    return [
+                        4,
+                        sleep(baseDelayMs * Math.pow(2, attempt))
+                    ];
+                case 6:
+                    _state.sent();
+                    return [
+                        3,
+                        7
+                    ];
+                case 7:
+                    attempt++;
+                    return [
+                        3,
+                        1
+                    ];
+                case 8:
+                    throw new Error('unreachable');
+            }
+        });
+    })();
+}
+function probeReachable(url) {
+    var timeoutMs = arguments.length > 1 && arguments[1] !== void 0 ? arguments[1] : 1500;
+    return _async_to_generator(function() {
+        var unused;
+        return _ts_generator(this, function(_state) {
+            switch(_state.label){
+                case 0:
+                    _state.trys.push([
+                        0,
+                        2,
+                        ,
+                        3
+                    ]);
+                    return [
+                        4,
+                        fetch(url, {
+                            signal: AbortSignal.timeout(timeoutMs)
+                        })
+                    ];
+                case 1:
+                    _state.sent();
+                    return [
+                        2,
+                        true
+                    ];
+                case 2:
+                    unused = _state.sent();
+                    return [
+                        2,
+                        false
+                    ];
+                case 3:
+                    return [
+                        2
+                    ];
+            }
+        });
+    })();
+}
+/* CJS INTEROP */ if (exports.__esModule && exports.default) { try { Object.defineProperty(exports.default, '__esModule', { value: true }); for (var key in exports) { exports.default[key] = exports[key]; } } catch (_) {}; module.exports = exports.default; }

package/dist/cjs/embed/http.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/embed/http.ts"],"sourcesContent":["const MAX_ATTEMPTS = 3;\n\nfunction sleep(ms: number): Promise<void> {\n  return new Promise((resolve) => setTimeout(resolve, ms));\n}\n\n// Shared retry glue for the HTTP providers: 429/5xx/network failure retry with exponential\n// backoff honoring Retry-After; any other status returns immediately for the caller to handle.\nexport async function fetchWithRetry(url: string, init: RequestInit, baseDelayMs = 200): Promise<Response> {\n  for (let attempt = 0; attempt < MAX_ATTEMPTS; attempt++) {\n    try {\n      const res = await fetch(url, init);\n      if (res.ok || (res.status !== 429 && res.status < 500) || attempt === MAX_ATTEMPTS - 1) return res;\n      const retryAfter = Number(res.headers.get('retry-after'));\n      // Capped: a server naming a huge Retry-After must not stall an embed run for it.\n      await sleep(retryAfter > 0 ? Math.min(retryAfter * 1000, 10_000) : baseDelayMs * 2 ** attempt);\n    } catch (err) {\n      if (attempt === MAX_ATTEMPTS - 1) throw err;\n      await sleep(baseDelayMs * 2 ** attempt);\n    }\n  }\n  throw new Error('unreachable');\n}\n\n// Any HTTP response means the endpoint is up; only a network failure or the timeout counts as\n// unreachable, so an air-gapped machine never waits more than timeoutMs on `sense status`.\nexport async function probeReachable(url: string, timeoutMs = 1500): Promise<boolean> {\n  try {\n    await fetch(url, { signal: AbortSignal.timeout(timeoutMs) });\n    return true;\n  } catch {\n    return false;\n  }\n}\n"],"names":["fetchWithRetry","probeReachable","MAX_ATTEMPTS","sleep","ms","Promise","resolve","setTimeout","url","init","baseDelayMs","attempt","res","retryAfter","err","fetch","ok","status","Number","headers","get","Math","min","Error","timeoutMs","signal","AbortSignal","timeout"],"mappings":";;;;;;;;;;;QAQsBA;eAAAA;;QAkBAC;eAAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA1BtB,IAAMC,eAAe;AAErB,SAASC,MAAMC,EAAU;IACvB,OAAO,IAAIC,QAAQ,SAACC;eAAYC,WAAWD,SAASF;;AACtD;AAIO,SAAeJ,eAAeQ,GAAW,EAAEC,IAAiB;QAAEC,cAAAA,iEAAc;;YACxEC,SAECC,KAEAC,YAGCC;;;;oBAPFH,UAAU;;;yBAAGA,CAAAA,UAAUT,YAAW;;;;;;;;;;;;oBAE3B;;wBAAMa,MAAMP,KAAKC;;;oBAAvBG,MAAM;oBACZ,IAAIA,IAAII,EAAE,IAAKJ,IAAIK,MAAM,KAAK,OAAOL,IAAIK,MAAM,GAAG,OAAQN,YAAYT,eAAe,GAAG;;wBAAOU;;oBACzFC,aAAaK,OAAON,IAAIO,OAAO,CAACC,GAAG,CAAC;oBAC1C,iFAAiF;oBACjF;;wBAAMjB,MAAMU,aAAa,IAAIQ,KAAKC,GAAG,CAACT,aAAa,MAAM,SAAUH,uBAAc,GAAKC;;;oBAAtF;;;;;;oBACOG;oBACP,IAAIH,YAAYT,eAAe,GAAG,MAAMY;oBACxC;;wBAAMX,MAAMO,uBAAc,GAAKC;;;oBAA/B;;;;;;oBAT0CA;;;;;;oBAY9C,MAAM,IAAIY,MAAM;;;IAClB;;AAIO,SAAetB,eAAeO,GAAW;QAAEgB,YAAAA,iEAAY;;;;;;;;;;;;oBAE1D;;wBAAMT,MAAMP,KAAK;4BAAEiB,QAAQC,YAAYC,OAAO,CAACH;wBAAW;;;oBAA1D;oBACA;;wBAAO;;;;oBAEP;;wBAAO;;;;;;;;IAEX"}

package/dist/cjs/embed/identity.d.cts

@@ -0,0 +1,13 @@
+export declare const MODEL_FILES: string[];
+export declare const MODEL_FILENAMES: string;
+export declare function isDownloadable(model: string): boolean;
+export declare function refsMainPath(model: string): string;
+export declare function readRef(model: string): string | undefined;
+export declare function writeRef(model: string, sha: string): void;
+export declare function snapshotDir(model: string, sha: string): string;
+export declare function languagesPath(model: string): string;
+export declare function readLanguages(model: string): string[] | undefined;
+export declare function writeLanguages(model: string, languages: string[]): void;
+export declare function modelDir(model: string): string;
+export declare function hasModelFiles(model: string): boolean;
+export declare function modelIdentity(model: string): string | undefined;