lmgrep 0.1.18 → 0.2.0

package/dist/infrastructure/lancedb/ChunkRepository.js

@@ -0,0 +1,166 @@
+import { Chunk } from "../../domain/corpus/Chunk.js";
+import { CodeLocation } from "../../domain/corpus/CodeLocation.js";
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import { FileVersion } from "../../domain/corpus/FileVersion.js";
+import { Hit } from "../../domain/retrieval/Hit.js";
+import { HitList } from "../../domain/retrieval/HitList.js";
+import { MissingIndexError } from "../../domain/retrieval/MissingIndexError.js";
+import { LanceTables, TableName } from "./LanceTables.js";
+import { SEARCH_COLUMNS, VectorIndexPolicy } from "./VectorIndexPolicy.js";
+/**
+ * Embedded chunks in LanceDB.
+ *
+ * Branch scoping lives here because it is a property of how chunks are stored:
+ * they are shared across branches by content hash, so a query must intersect
+ * them with the manifest of file versions the branch actually references.
+ */
+export class ChunkRepository {
+    tables;
+    manifest;
+    branch;
+    constructor(tables, manifest, branch) {
+        this.tables = tables;
+        this.manifest = manifest;
+        this.branch = branch;
+    }
+    async add(chunks) {
+        if (chunks.length === 0)
+            return;
+        const records = chunks.map(({ chunk, vector }) => this.toRow(chunk, vector));
+        const { table, seeded } = await this.tables.tableOrCreate(TableName.Chunks, records);
+        if (!seeded)
+            await table.add(records);
+    }
+    async search(query) {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table) {
+            throw new MissingIndexError();
+        }
+        const versions = query.scopeToBranch
+            ? await this.manifest.branchVersions()
+            : undefined;
+        // Over-fetch: branch/version filtering and dedup both discard rows, so
+        // pull extra to still return `limit` distinct results.
+        const fetchLimit = versions ? query.limit * 3 : query.limit * 2;
+        let builder = table
+            .query()
+            .nearestTo(query.vector.toArray())
+            .limit(fetchLimit)
+            .refineFactor(VectorIndexPolicy.REFINE_FACTOR)
+            .select([...SEARCH_COLUMNS]);
+        const predicate = this.buildPredicate(query);
+        if (predicate)
+            builder = builder.where(predicate);
+        const rows = await builder.toArray();
+        let hits = HitList.of(rows.map((r) => this.toHit(r)));
+        if (versions) {
+            hits = hits.filtered((h) => h.fileVersion.matches(versions.versionOf(h.location.filePath)));
+        }
+        return hits.deduplicated().takeAtMost(query.limit);
+    }
+    /**
+     * Delete chunks for files no longer referenced by ANY branch. A path still
+     * present in another branch's manifest keeps its chunks — they are shared
+     * by content, so deleting them would blank that branch's results.
+     */
+    async deleteByFiles(filePaths) {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table || filePaths.length === 0)
+            return;
+        const entries = await this.manifest.allEntries();
+        const referencedElsewhere = new Set(entries
+            .filter((e) => !e.branch.equals(this.branch))
+            .map((e) => e.filePath));
+        const toDelete = filePaths.filter((p) => !referencedElsewhere.has(p));
+        if (toDelete.length > 0) {
+            await LanceTables.deleteIn(table, "filePath", toDelete);
+        }
+    }
+    async existingHashes(hashes) {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table || hashes.length === 0)
+            return new Set();
+        const unique = [...new Set(hashes.map((h) => h.toString()))];
+        return LanceTables.selectIn(table, "hash", unique);
+    }
+    async count() {
+        const table = await this.tables.table(TableName.Chunks);
+        return table ? table.countRows() : 0;
+    }
+    async hashesByFile() {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table)
+            return new Map();
+        const rows = await table.query().select(["filePath", "hash"]).toArray();
+        const map = new Map();
+        for (const row of rows) {
+            const fp = row.filePath;
+            const existing = map.get(fp) ?? [];
+            existing.push(row.hash);
+            map.set(fp, existing);
+        }
+        return map;
+    }
+    async allHashes() {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table)
+            return new Set();
+        const rows = await table.query().select(["hash"]).toArray();
+        return new Set(rows.map((r) => r.hash));
+    }
+    async *streamTexts(batchSize = 1000) {
+        const table = await this.tables.table(TableName.Chunks);
+        if (!table)
+            return;
+        const rows = await table.query().select(["name", "content"]).toArray();
+        for (let i = 0; i < rows.length; i += batchSize) {
+            yield rows.slice(i, i + batchSize).map((r) => ({
+                name: r.name ?? "",
+                content: r.content ?? "",
+            }));
+        }
+    }
+    buildPredicate(query) {
+        const conditions = [];
+        if (query.filePrefix) {
+            conditions.push(`filePath LIKE '${LanceTables.quote(query.filePrefix)}%'`);
+        }
+        if (query.types && query.types.length > 0) {
+            const escaped = query.types.map((t) => `'${LanceTables.quote(t)}'`);
+            conditions.push(`type IN (${escaped.join(", ")})`);
+        }
+        return conditions.length > 0 ? conditions.join(" AND ") : undefined;
+    }
+    toRow(chunk, vector) {
+        return {
+            id: chunk.id,
+            filePath: chunk.location.filePath,
+            startLine: chunk.location.startLine,
+            endLine: chunk.location.endLine,
+            type: chunk.type,
+            name: chunk.name,
+            content: chunk.content,
+            context: chunk.context,
+            hash: chunk.hash.toString(),
+            fileHash: chunk.fileVersion.toStored(),
+            vector: vector.toArray(),
+        };
+    }
+    toHit(row) {
+        const distance = row._distance;
+        return new Hit(row.id, new CodeLocation(row.filePath, row.startLine, row.endLine), row.type, row.name, row.content, row.context, distance != null ? 1 - distance : 0, FileVersion.fromStored(row.fileHash));
+    }
+    /** Rebuild a domain Chunk from a stored row — used by import and dedup. */
+    static rowToChunk(row) {
+        return new Chunk({
+            location: new CodeLocation(row.filePath, row.startLine, row.endLine),
+            type: row.type,
+            name: row.name,
+            content: row.content,
+            context: row.context,
+            hash: ContentHash.fromStored(row.hash),
+            fileVersion: FileVersion.fromStored(row.fileHash),
+        });
+    }
+}
+//# sourceMappingURL=ChunkRepository.js.map

package/dist/infrastructure/lancedb/ChunkRepository.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ChunkRepository.js","sourceRoot":"","sources":["../../../src/infrastructure/lancedb/ChunkRepository.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,8BAA8B,CAAC;AACrD,OAAO,EAAE,YAAY,EAAE,MAAM,qCAAqC,CAAC;AACnE,OAAO,EAAE,WAAW,EAAE,MAAM,oCAAoC,CAAC;AACjE,OAAO,EAAE,WAAW,EAAE,MAAM,oCAAoC,CAAC;AAUjE,OAAO,EAAE,GAAG,EAAE,MAAM,+BAA+B,CAAC;AACpD,OAAO,EAAE,OAAO,EAAE,MAAM,mCAAmC,CAAC;AAC5D,OAAO,EAAE,iBAAiB,EAAE,MAAM,6CAA6C,CAAC;AAChF,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,cAAc,EAAE,iBAAiB,EAAE,MAAM,wBAAwB,CAAC;AAiB3E;;;;;;GAMG;AACH,MAAM,OAAO,eAAe;IAET;IACA;IACA;IAHlB,YACkB,MAAmB,EACnB,QAAoC,EACpC,MAAc;QAFd,WAAM,GAAN,MAAM,CAAa;QACnB,aAAQ,GAAR,QAAQ,CAA4B;QACpC,WAAM,GAAN,MAAM,CAAQ;IAC7B,CAAC;IAEJ,KAAK,CAAC,GAAG,CAAC,MAAuB;QAChC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAChC,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE,EAAE,CAChD,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,CACzB,CAAC;QACF,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,aAAa,CACxD,SAAS,CAAC,MAAM,EAChB,OAAO,CACP,CAAC;QACF,IAAI,CAAC,MAAM;YAAE,MAAM,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IACvC,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,KAAiB;QAC7B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK,EAAE,CAAC;YACZ,MAAM,IAAI,iBAAiB,EAAE,CAAC;QAC/B,CAAC;QAED,MAAM,QAAQ,GAAG,KAAK,CAAC,aAAa;YACnC,CAAC,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,cAAc,EAAE;YACtC,CAAC,CAAC,SAAS,CAAC;QAEb,uEAAuE;QACvE,uDAAuD;QACvD,MAAM,UAAU,GAAG,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC;QAEhE,IAAI,OAAO,GAAG,KAAK;aACjB,KAAK,EAAE;aACP,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;aACjC,KAAK,CAAC,UAAU,CAAC;aACjB,YAAY,CAAC,iBAAiB,CAAC,aAAa,CAAC;aAC7C,MAAM,CAAC,CAAC,GAAG,cAAc,CAAC,CAAC,CAAC;QAE9B,MAAM,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,KAAK,CAAC,CAAC;QAC7C,IAAI,SAAS;YAAE,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QAElD,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;QACrC,IAAI,IAAI,GAAG,OAAO,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAEtD,IAAI,QAAQ,EAAE,CAAC;YACd,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,EAAE,CAC1B,CAAC,CAAC,WAAW,CAAC,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAC9D,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC,YAAY,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IACpD,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,aAAa,CAAC,SAAmB;QACtC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAE7C,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,UAAU,EAAE,CAAC;QACjD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAClC,OAAO;aACL,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;aAC5C,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CACxB,CAAC;QAEF,MAAM,QAAQ,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,mBAAmB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACtE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzB,MAAM,WAAW,CAAC,QAAQ,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;QACzD,CAAC;IACF,CAAC;IAED,KAAK,CAAC,cAAc,CAAC,MAAqB;QACzC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,IAAI,GAAG,EAAE,CAAC;QACpD,MAAM,MAAM,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC;QAC7D,OAAO,WAAW,CAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;IACpD,CAAC;IAED,KAAK,CAAC,KAAK;QACV,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACtC,CAAC;IAED,KAAK,CAAC,YAAY;QACjB,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,GAAG,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;QACxE,MAAM,GAAG,GAAG,IAAI,GAAG,EAAoB,CAAC;QACxC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACxB,MAAM,EAAE,GAAG,GAAG,CAAC,QAAkB,CAAC;YAClC,MAAM,QAAQ,GAAG,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC;YACnC,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,IAAc,CAAC,CAAC;YAClC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;QACvB,CAAC;QACD,OAAO,GAAG,CAAC;IACZ,CAAC;IAED,KAAK,CAAC,SAAS;QACd,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,GAAG,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;QAC5D,OAAO,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAc,CAAC,CAAC,CAAC;IACnD,CAAC;IAED,KAAK,CAAC,CAAC,WAAW,CAAC,SAAS,GAAG,IAAI;QAClC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACxD,IAAI,CAAC,KAAK;YAAE,OAAO;QACnB,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;QACvE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;YACjD,MAAM,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;gBAC9C,IAAI,EAAG,CAAC,CAAC,IAAe,IAAI,EAAE;gBAC9B,OAAO,EAAG,CAAC,CAAC,OAAkB,IAAI,EAAE;aACpC,CAAC,CAAC,CAAC;QACL,CAAC;IACF,CAAC;IAEO,cAAc,CAAC,KAAiB;QACvC,MAAM,UAAU,GAAa,EAAE,CAAC;QAChC,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YACtB,UAAU,CAAC,IAAI,CACd,kBAAkB,WAAW,CAAC,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,IAAI,CACzD,CAAC;QACH,CAAC;QACD,IAAI,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACpE,UAAU,CAAC,IAAI,CAAC,YAAY,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACpD,CAAC;QACD,OAAO,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACrE,CAAC;IAEO,KAAK,CAAC,KAAY,EAAE,MAAc;QACzC,OAAO;YACN,EAAE,EAAE,KAAK,CAAC,EAAE;YACZ,QAAQ,EAAE,KAAK,CAAC,QAAQ,CAAC,QAAQ;YACjC,SAAS,EAAE,KAAK,CAAC,QAAQ,CAAC,SAAS;YACnC,OAAO,EAAE,KAAK,CAAC,QAAQ,CAAC,OAAO;YAC/B,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,EAAE;YAC3B,QAAQ,EAAE,KAAK,CAAC,WAAW,CAAC,QAAQ,EAAE;YACtC,MAAM,EAAE,MAAM,CAAC,OAAO,EAAE;SACxB,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,GAA4B;QACzC,MAAM,QAAQ,GAAG,GAAG,CAAC,SAAsC,CAAC;QAC5D,OAAO,IAAI,GAAG,CACb,GAAG,CAAC,EAAY,EAChB,IAAI,YAAY,CACf,GAAG,CAAC,QAAkB,EACtB,GAAG,CAAC,SAAmB,EACvB,GAAG,CAAC,OAAiB,CACrB,EACD,GAAG,CAAC,IAAc,EAClB,GAAG,CAAC,IAAc,EAClB,GAAG,CAAC,OAAiB,EACrB,GAAG,CAAC,OAAiB,EACrB,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,EACnC,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC,QAA8B,CAAC,CAC1D,CAAC;IACH,CAAC;IAED,2EAA2E;IAC3E,MAAM,CAAC,UAAU,CAAC,GAA4B;QAC7C,OAAO,IAAI,KAAK,CAAC;YAChB,QAAQ,EAAE,IAAI,YAAY,CACzB,GAAG,CAAC,QAAkB,EACtB,GAAG,CAAC,SAAmB,EACvB,GAAG,CAAC,OAAiB,CACrB;YACD,IAAI,EAAE,GAAG,CAAC,IAAc;YACxB,IAAI,EAAE,GAAG,CAAC,IAAc;YACxB,OAAO,EAAE,GAAG,CAAC,OAAiB;YAC9B,OAAO,EAAE,GAAG,CAAC,OAAiB;YAC9B,IAAI,EAAE,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC,IAAc,CAAC;YAChD,WAAW,EAAE,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC,QAA8B,CAAC;SACvE,CAAC,CAAC;IACJ,CAAC;CACD"}

package/dist/infrastructure/lancedb/DatabaseImporter.d.ts

@@ -0,0 +1,22 @@
+import type { Branch } from "../../domain/project/Branch.js";
+import { type LanceTables } from "./LanceTables.js";
+export interface ImportResult {
+    chunks: number;
+    files: number;
+}
+/**
+ * Copies chunks and manifest rows out of another lmgrep database.
+ *
+ * Rows move verbatim, embeddings included — the point of importing is to avoid
+ * re-embedding, so re-deriving anything here would defeat it. The source must
+ * therefore have been built with a compatible model; the caller reports the
+ * source's model so the user can check.
+ */
+export declare class DatabaseImporter {
+    private readonly tables;
+    private readonly branch;
+    constructor(tables: LanceTables, branch: Branch);
+    importFrom(sourcePath: string): Promise<ImportResult>;
+    private copyChunks;
+    private copyFiles;
+}

package/dist/infrastructure/lancedb/DatabaseImporter.js

@@ -0,0 +1,74 @@
+import { connect } from "@lancedb/lancedb";
+import { TableName } from "./LanceTables.js";
+/**
+ * Copies chunks and manifest rows out of another lmgrep database.
+ *
+ * Rows move verbatim, embeddings included — the point of importing is to avoid
+ * re-embedding, so re-deriving anything here would defeat it. The source must
+ * therefore have been built with a compatible model; the caller reports the
+ * source's model so the user can check.
+ */
+export class DatabaseImporter {
+    tables;
+    branch;
+    constructor(tables, branch) {
+        this.tables = tables;
+        this.branch = branch;
+    }
+    async importFrom(sourcePath) {
+        const source = await connect(sourcePath);
+        const available = await source.tableNames();
+        return {
+            chunks: available.includes(TableName.Chunks)
+                ? await this.copyChunks(source)
+                : 0,
+            files: available.includes(TableName.Files)
+                ? await this.copyFiles(source)
+                : 0,
+        };
+    }
+    async copyChunks(source) {
+        const rows = await (await source.openTable(TableName.Chunks))
+            .query()
+            .toArray();
+        if (rows.length === 0)
+            return 0;
+        // Arrow hands back typed arrays; LanceDB needs plain values to infer a
+        // schema when the destination table does not exist yet.
+        const records = rows.map((r) => ({
+            id: r.id,
+            filePath: r.filePath,
+            startLine: r.startLine,
+            endLine: r.endLine,
+            type: r.type,
+            name: r.name,
+            content: r.content,
+            context: r.context,
+            hash: r.hash,
+            fileHash: r.fileHash ?? "",
+            vector: Array.from(r.vector),
+        }));
+        const { table, seeded } = await this.tables.tableOrCreate(TableName.Chunks, records);
+        if (!seeded)
+            await table.add(records);
+        return records.length;
+    }
+    async copyFiles(source) {
+        const rows = await (await source.openTable(TableName.Files))
+            .query()
+            .toArray();
+        if (rows.length === 0)
+            return 0;
+        // Legacy databases predate the branch column; adopt ours for those.
+        const records = rows.map((r) => ({
+            filePath: r.filePath,
+            fileHash: r.fileHash,
+            branch: r.branch ?? this.branch.toString(),
+        }));
+        const { table, seeded } = await this.tables.tableOrCreate(TableName.Files, records);
+        if (!seeded)
+            await table.add(records);
+        return records.length;
+    }
+}
+//# sourceMappingURL=DatabaseImporter.js.map

package/dist/infrastructure/lancedb/DatabaseImporter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"DatabaseImporter.js","sourceRoot":"","sources":["../../../src/infrastructure/lancedb/DatabaseImporter.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAE3C,OAAO,EAAoB,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAO/D;;;;;;;GAOG;AACH,MAAM,OAAO,gBAAgB;IAEV;IACA;IAFlB,YACkB,MAAmB,EACnB,MAAc;QADd,WAAM,GAAN,MAAM,CAAa;QACnB,WAAM,GAAN,MAAM,CAAQ;IAC7B,CAAC;IAEJ,KAAK,CAAC,UAAU,CAAC,UAAkB;QAClC,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,UAAU,CAAC,CAAC;QACzC,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,UAAU,EAAE,CAAC;QAE5C,OAAO;YACN,MAAM,EAAE,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC;gBAC3C,CAAC,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;gBAC/B,CAAC,CAAC,CAAC;YACJ,KAAK,EAAE,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC,KAAK,CAAC;gBACzC,CAAC,CAAC,MAAM,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC;gBAC9B,CAAC,CAAC,CAAC;SACJ,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,UAAU,CACvB,MAA2C;QAE3C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;aAC3D,KAAK,EAAE;aACP,OAAO,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAEhC,uEAAuE;QACvE,wDAAwD;QACxD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAChC,EAAE,EAAE,CAAC,CAAC,EAAE;YACR,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,SAAS,EAAE,CAAC,CAAC,SAAS;YACtB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,QAAQ,EAAG,CAAC,CAAC,QAAmB,IAAI,EAAE;YACtC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,MAA0B,CAAC;SAChD,CAAC,CAAC,CAAC;QAEJ,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,aAAa,CACxD,SAAS,CAAC,MAAM,EAChB,OAAO,CACP,CAAC;QACF,IAAI,CAAC,MAAM;YAAE,MAAM,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACtC,OAAO,OAAO,CAAC,MAAM,CAAC;IACvB,CAAC;IAEO,KAAK,CAAC,SAAS,CACtB,MAA2C;QAE3C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;aAC1D,KAAK,EAAE;aACP,OAAO,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAEhC,oEAAoE;QACpE,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAChC,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,MAAM,EAAE,CAAC,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE;SAC1C,CAAC,CAAC,CAAC;QAEJ,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,aAAa,CACxD,SAAS,CAAC,KAAK,EACf,OAAO,CACP,CAAC;QACF,IAAI,CAAC,MAAM;YAAE,MAAM,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACtC,OAAO,OAAO,CAAC,MAAM,CAAC;IACvB,CAAC;CACD"}

package/dist/infrastructure/lancedb/FileManifestRepository.d.ts

@@ -0,0 +1,34 @@
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import { FileManifest, type SourceFile } from "../../domain/corpus/SourceFile.js";
+import type { FileManifestRepositoryPort, ManifestEntry } from "../../domain/ports/FileManifestRepositoryPort.js";
+import { Branch } from "../../domain/project/Branch.js";
+import { LanceTables } from "./LanceTables.js";
+/**
+ * The branch-scoped file manifest in LanceDB.
+ *
+ * The current branch's manifest is cached because every search consults it to
+ * scope results, and re-reading it per query would dominate search latency.
+ * Anything that writes the manifest must {@link invalidate}.
+ */
+export declare class FileManifestRepository implements FileManifestRepositoryPort {
+    private readonly tables;
+    private readonly branch;
+    private cached;
+    constructor(tables: LanceTables, branch: Branch);
+    current(): Promise<FileManifest>;
+    branchVersions(): Promise<FileManifest | undefined>;
+    invalidate(): void;
+    upsert(entries: SourceFile[]): Promise<void>;
+    deleteFiles(filePaths: string[]): Promise<void>;
+    /**
+     * Which of these file versions any branch has already indexed. A hit means
+     * the content is embedded, so the file can be registered on this branch
+     * without re-chunking it.
+     */
+    knownHashes(hashes: ContentHash[]): Promise<Set<string>>;
+    allEntries(): Promise<ManifestEntry[]>;
+    storedBranches(): Promise<string[]>;
+    deleteBranch(branch: string): Promise<void>;
+    /** Seed this branch's manifest from another's — used on a new branch. */
+    copyFromBranch(sourceBranch: string): Promise<number>;
+}

package/dist/infrastructure/lancedb/FileManifestRepository.js

@@ -0,0 +1,137 @@
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import { FileManifest, } from "../../domain/corpus/SourceFile.js";
+import { Branch } from "../../domain/project/Branch.js";
+import { LanceTables, TableName } from "./LanceTables.js";
+/**
+ * The branch-scoped file manifest in LanceDB.
+ *
+ * The current branch's manifest is cached because every search consults it to
+ * scope results, and re-reading it per query would dominate search latency.
+ * Anything that writes the manifest must {@link invalidate}.
+ */
+export class FileManifestRepository {
+    tables;
+    branch;
+    cached;
+    constructor(tables, branch) {
+        this.tables = tables;
+        this.branch = branch;
+    }
+    async current() {
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return FileManifest.empty();
+        const rows = await table
+            .query()
+            .where(`branch = '${this.branch.toSqlLiteral()}'`)
+            .select(["filePath", "fileHash"])
+            .toArray();
+        return FileManifest.fromEntries(rows.map((r) => [
+            r.filePath,
+            ContentHash.fromStored(r.fileHash),
+        ]));
+    }
+    async branchVersions() {
+        if (this.cached)
+            return this.cached;
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return undefined;
+        this.cached = await this.current();
+        return this.cached;
+    }
+    invalidate() {
+        this.cached = undefined;
+    }
+    async upsert(entries) {
+        if (entries.length === 0)
+            return;
+        const records = entries.map((e) => ({
+            filePath: e.path,
+            fileHash: e.hash.toString(),
+            branch: this.branch.toString(),
+        }));
+        const { table, seeded } = await this.tables.tableOrCreate(TableName.Files, records);
+        if (seeded) {
+            this.invalidate();
+            return;
+        }
+        // Replace this branch's rows for these paths, leaving other branches'
+        // rows for the same paths untouched.
+        const paths = records.map((r) => r.filePath);
+        for (let i = 0; i < paths.length; i += LanceTables.FILTER_BATCH_SIZE) {
+            const batch = paths.slice(i, i + LanceTables.FILTER_BATCH_SIZE);
+            await table.delete(`branch = '${this.branch.toSqlLiteral()}' AND ${LanceTables.inFilter("filePath", batch)}`);
+        }
+        await table.add(records);
+        this.invalidate();
+    }
+    async deleteFiles(filePaths) {
+        const table = await this.tables.table(TableName.Files);
+        if (!table || filePaths.length === 0)
+            return;
+        for (let i = 0; i < filePaths.length; i += LanceTables.FILTER_BATCH_SIZE) {
+            const batch = filePaths.slice(i, i + LanceTables.FILTER_BATCH_SIZE);
+            await table.delete(`branch = '${this.branch.toSqlLiteral()}' AND ${LanceTables.inFilter("filePath", batch)}`);
+        }
+        this.invalidate();
+    }
+    /**
+     * Which of these file versions any branch has already indexed. A hit means
+     * the content is embedded, so the file can be registered on this branch
+     * without re-chunking it.
+     */
+    async knownHashes(hashes) {
+        const table = await this.tables.table(TableName.Files);
+        if (!table || hashes.length === 0)
+            return new Set();
+        return LanceTables.selectIn(table, "fileHash", hashes.map((h) => h.toString()));
+    }
+    async allEntries() {
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return [];
+        const rows = await table.query().toArray();
+        return rows.map((r) => ({
+            filePath: r.filePath,
+            fileHash: ContentHash.fromStored(r.fileHash),
+            branch: Branch.of(r.branch),
+        }));
+    }
+    async storedBranches() {
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return [];
+        const rows = await table.query().select(["branch"]).toArray();
+        return [...new Set(rows.map((r) => r.branch))];
+    }
+    async deleteBranch(branch) {
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return;
+        await table.delete(`branch = '${LanceTables.quote(branch)}'`);
+        this.invalidate();
+    }
+    /** Seed this branch's manifest from another's — used on a new branch. */
+    async copyFromBranch(sourceBranch) {
+        const table = await this.tables.table(TableName.Files);
+        if (!table)
+            return 0;
+        const rows = await table
+            .query()
+            .where(`branch = '${LanceTables.quote(sourceBranch)}'`)
+            .select(["filePath", "fileHash"])
+            .toArray();
+        if (rows.length === 0)
+            return 0;
+        const records = rows.map((r) => ({
+            filePath: r.filePath,
+            fileHash: r.fileHash,
+            branch: this.branch.toString(),
+        }));
+        await table.add(records);
+        this.invalidate();
+        return records.length;
+    }
+}
+//# sourceMappingURL=FileManifestRepository.js.map

package/dist/infrastructure/lancedb/FileManifestRepository.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"FileManifestRepository.js","sourceRoot":"","sources":["../../../src/infrastructure/lancedb/FileManifestRepository.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,oCAAoC,CAAC;AACjE,OAAO,EACN,YAAY,GAEZ,MAAM,mCAAmC,CAAC;AAK3C,OAAO,EAAE,MAAM,EAAE,MAAM,gCAAgC,CAAC;AACxD,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAQ1D;;;;;;GAMG;AACH,MAAM,OAAO,sBAAsB;IAIhB;IACA;IAJV,MAAM,CAA2B;IAEzC,YACkB,MAAmB,EACnB,MAAc;QADd,WAAM,GAAN,MAAM,CAAa;QACnB,WAAM,GAAN,MAAM,CAAQ;IAC7B,CAAC;IAEJ,KAAK,CAAC,OAAO;QACZ,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO,YAAY,CAAC,KAAK,EAAE,CAAC;QAExC,MAAM,IAAI,GAAG,MAAM,KAAK;aACtB,KAAK,EAAE;aACP,KAAK,CAAC,aAAa,IAAI,CAAC,MAAM,CAAC,YAAY,EAAE,GAAG,CAAC;aACjD,MAAM,CAAC,CAAC,UAAU,EAAE,UAAU,CAAC,CAAC;aAChC,OAAO,EAAE,CAAC;QAEZ,OAAO,YAAY,CAAC,WAAW,CAC9B,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC;YACf,CAAC,CAAC,QAAkB;YACpB,WAAW,CAAC,UAAU,CAAC,CAAC,CAAC,QAAkB,CAAC;SAC5C,CAAC,CACF,CAAC;IACH,CAAC;IAED,KAAK,CAAC,cAAc;QACnB,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO,IAAI,CAAC,MAAM,CAAC;QACpC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO,SAAS,CAAC;QAC7B,IAAI,CAAC,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;QACnC,OAAO,IAAI,CAAC,MAAM,CAAC;IACpB,CAAC;IAED,UAAU;QACT,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC;IACzB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,OAAqB;QACjC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAEjC,MAAM,OAAO,GAAc,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC9C,QAAQ,EAAE,CAAC,CAAC,IAAI;YAChB,QAAQ,EAAE,CAAC,CAAC,IAAI,CAAC,QAAQ,EAAE;YAC3B,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE;SAC9B,CAAC,CAAC,CAAC;QAEJ,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,aAAa,CACxD,SAAS,CAAC,KAAK,EACf,OAAO,CACP,CAAC;QACF,IAAI,MAAM,EAAE,CAAC;YACZ,IAAI,CAAC,UAAU,EAAE,CAAC;YAClB,OAAO;QACR,CAAC;QAED,sEAAsE;QACtE,qCAAqC;QACrC,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;QAC7C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,CAAC,iBAAiB,EAAE,CAAC;YACtE,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,iBAAiB,CAAC,CAAC;YAChE,MAAM,KAAK,CAAC,MAAM,CACjB,aAAa,IAAI,CAAC,MAAM,CAAC,YAAY,EAAE,SAAS,WAAW,CAAC,QAAQ,CAAC,UAAU,EAAE,KAAK,CAAC,EAAE,CACzF,CAAC;QACH,CAAC;QACD,MAAM,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACzB,IAAI,CAAC,UAAU,EAAE,CAAC;IACnB,CAAC;IAED,KAAK,CAAC,WAAW,CAAC,SAAmB;QACpC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAC7C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,CAAC,iBAAiB,EAAE,CAAC;YAC1E,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,iBAAiB,CAAC,CAAC;YACpE,MAAM,KAAK,CAAC,MAAM,CACjB,aAAa,IAAI,CAAC,MAAM,CAAC,YAAY,EAAE,SAAS,WAAW,CAAC,QAAQ,CAAC,UAAU,EAAE,KAAK,CAAC,EAAE,CACzF,CAAC;QACH,CAAC;QACD,IAAI,CAAC,UAAU,EAAE,CAAC;IACnB,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,WAAW,CAAC,MAAqB;QACtC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,IAAI,GAAG,EAAE,CAAC;QACpD,OAAO,WAAW,CAAC,QAAQ,CAC1B,KAAK,EACL,UAAU,EACV,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAC,CAC/B,CAAC;IACH,CAAC;IAED,KAAK,CAAC,UAAU;QACf,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,KAAK,EAAE,CAAC,OAAO,EAAE,CAAC;QAC3C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACvB,QAAQ,EAAE,CAAC,CAAC,QAAkB;YAC9B,QAAQ,EAAE,WAAW,CAAC,UAAU,CAAC,CAAC,CAAC,QAAkB,CAAC;YACtD,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,MAAgB,CAAC;SACrC,CAAC,CAAC,CAAC;IACL,CAAC;IAED,KAAK,CAAC,cAAc;QACnB,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;QAC9D,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAgB,CAAC,CAAC,CAAC,CAAC;IAC1D,CAAC;IAED,KAAK,CAAC,YAAY,CAAC,MAAc;QAChC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO;QACnB,MAAM,KAAK,CAAC,MAAM,CAAC,aAAa,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC9D,IAAI,CAAC,UAAU,EAAE,CAAC;IACnB,CAAC;IAED,yEAAyE;IACzE,KAAK,CAAC,cAAc,CAAC,YAAoB;QACxC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO,CAAC,CAAC;QAErB,MAAM,IAAI,GAAG,MAAM,KAAK;aACtB,KAAK,EAAE;aACP,KAAK,CAAC,aAAa,WAAW,CAAC,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC;aACtD,MAAM,CAAC,CAAC,UAAU,EAAE,UAAU,CAAC,CAAC;aAChC,OAAO,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAEhC,MAAM,OAAO,GAAc,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC3C,QAAQ,EAAE,CAAC,CAAC,QAAkB;YAC9B,QAAQ,EAAE,CAAC,CAAC,QAAkB;YAC9B,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE;SAC9B,CAAC,CAAC,CAAC;QACJ,MAAM,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACzB,IAAI,CAAC,UAAU,EAAE,CAAC;QAClB,OAAO,OAAO,CAAC,MAAM,CAAC;IACvB,CAAC;CACD"}

package/dist/infrastructure/lancedb/IndexMaintenance.d.ts

@@ -0,0 +1,53 @@
+import type { FileManifestRepositoryPort } from "../../domain/ports/FileManifestRepositoryPort.js";
+import type { DedupeReport, IndexMaintenancePort, OptimizeOptions, OptimizeReport, VectorIndexState } from "../../domain/ports/IndexMaintenancePort.js";
+import { type LanceTables } from "./LanceTables.js";
+/**
+ * Keeps the LanceDB tables compact and their ANN indexes current.
+ *
+ * Without a vector index LanceDB answers a search by brute force: it decodes
+ * every stored embedding to score it. At high dimensions that walks the whole
+ * table through memory per query, so peak RSS tracks index size roughly 1:1.
+ * IVF-PQ turns that into a handful of probes.
+ *
+ * Compaction matters for the same reason — small appends leave hundreds of
+ * fragments, each carrying its own decode buffers.
+ */
+export declare class IndexMaintenance implements IndexMaintenancePort {
+    private readonly tables;
+    private readonly manifest;
+    /** Tables holding embeddings, and therefore wanting a vector index. */
+    private static readonly VECTOR_TABLES;
+    constructor(tables: LanceTables, manifest: FileManifestRepositoryPort);
+    /**
+     * Cheap and idempotent when there is nothing to do, so it is safe to call
+     * after every build.
+     */
+    optimize(options?: OptimizeOptions): Promise<OptimizeReport>;
+    /**
+     * What `lmgrep compact` runs: the same work as {@link optimize} but
+     * unconditional, plus compaction of the files table (which holds no
+     * embeddings, so it never needs a vector index).
+     */
+    compact(): Promise<OptimizeReport>;
+    /**
+     * Remove tables no longer part of the schema.
+     *
+     * Databases built before faceting was removed still carry a `vocab` table
+     * holding one embedding per corpus term — nothing reads it now, and at
+     * full embedding width that is real disk. Dropping it here means users
+     * reclaim the space by running the command they already run.
+     */
+    private dropLegacyTables;
+    private optimizeTable;
+    /**
+     * Remove rows that should not be there: exact id duplicates (from
+     * concurrent unlocked indexing) and chunks whose file version no branch
+     * references any more. Survivors are rewritten into a fresh table, which is
+     * also what reclaims the disk the dropped rows held.
+     */
+    dedupe(): Promise<DedupeReport>;
+    vectorIndexState(): Promise<VectorIndexState>;
+    reset(): Promise<void>;
+    /** Arrow rows carry typed arrays; LanceDB needs plain values to re-infer a schema. */
+    private toPlainRow;
+}