lmgrep 0.1.18 → 0.2.0

package/dist/infrastructure/fs/ProjectMetadataStore.d.ts

@@ -0,0 +1,42 @@
+import type { IndexMetadataPort } from "../../domain/ports/IndexMetadataPort.js";
+import type { IndexMetadata } from "../../domain/project/IndexMetadata.js";
+/**
+ * The `lmgrep.json` sidecar written beside each database.
+ *
+ * It records the model and dimensions the index was built with, which is what
+ * lets a later search refuse to run against incompatible embeddings instead of
+ * silently returning nonsense.
+ */
+export declare class ProjectMetadataStore implements IndexMetadataPort {
+    private static readonly FILE;
+    /** What an lmgrep database looks like from the outside. */
+    private static readonly MARKERS;
+    read(databasePath: string): IndexMetadata | undefined;
+    /**
+     * Write the sidecar, preserving the model and dimensions already recorded.
+     *
+     * The first index establishes the baseline; later writes must not overwrite
+     * it, or a search with a different model would stop being detectable as a
+     * mismatch.
+     */
+    write(databasePath: string, metadata: Omit<IndexMetadata, "indexedAt">): void;
+    /**
+     * Whether a directory holds an lmgrep database, or is an empty directory
+     * safe to treat as one.
+     *
+     * `--database <path>` lets a caller aim any directory at any command, so a
+     * destructive operation must confirm the target is ours before touching it
+     * — without this, `lmgrep prune --database .` would delete a working tree.
+     * An empty directory passes because that is what an interrupted index
+     * leaves behind, and removing it cannot lose data.
+     */
+    isDatabaseDirectory(databasePath: string): boolean;
+    /**
+     * Whether a directory actually holds an index.
+     *
+     * Same markers as {@link isDatabaseDirectory}, without its allowance for
+     * an empty directory — the two are kept on one marker set so that what
+     * counts as a database is defined in exactly one place.
+     */
+    holdsIndex(databasePath: string): boolean;
+}

package/dist/infrastructure/fs/ProjectMetadataStore.js

@@ -0,0 +1,86 @@
+import { mkdirSync, readdirSync, readFileSync, writeFileSync } from "node:fs";
+import { join } from "node:path";
+import { TableName } from "../lancedb/LanceTables.js";
+/**
+ * The `lmgrep.json` sidecar written beside each database.
+ *
+ * It records the model and dimensions the index was built with, which is what
+ * lets a later search refuse to run against incompatible embeddings instead of
+ * silently returning nonsense.
+ */
+export class ProjectMetadataStore {
+    static FILE = "lmgrep.json";
+    /** What an lmgrep database looks like from the outside. */
+    static MARKERS = new Set([
+        ProjectMetadataStore.FILE,
+        `${TableName.Chunks}.lance`,
+        `${TableName.Files}.lance`,
+        `${TableName.LegacyVocab}.lance`,
+    ]);
+    read(databasePath) {
+        try {
+            return JSON.parse(readFileSync(join(databasePath, ProjectMetadataStore.FILE), "utf-8"));
+        }
+        catch {
+            return undefined;
+        }
+    }
+    /**
+     * Write the sidecar, preserving the model and dimensions already recorded.
+     *
+     * The first index establishes the baseline; later writes must not overwrite
+     * it, or a search with a different model would stop being detectable as a
+     * mismatch.
+     */
+    write(databasePath, metadata) {
+        mkdirSync(databasePath, { recursive: true });
+        const existing = this.read(databasePath);
+        const merged = {
+            root: metadata.root,
+            remote: metadata.remote,
+            branch: metadata.branch,
+            indexedAt: new Date().toISOString(),
+            model: existing?.model ?? metadata.model,
+            dimensions: existing?.dimensions ?? metadata.dimensions,
+        };
+        writeFileSync(join(databasePath, ProjectMetadataStore.FILE), JSON.stringify(merged, null, 2));
+    }
+    /**
+     * Whether a directory holds an lmgrep database, or is an empty directory
+     * safe to treat as one.
+     *
+     * `--database <path>` lets a caller aim any directory at any command, so a
+     * destructive operation must confirm the target is ours before touching it
+     * — without this, `lmgrep prune --database .` would delete a working tree.
+     * An empty directory passes because that is what an interrupted index
+     * leaves behind, and removing it cannot lose data.
+     */
+    isDatabaseDirectory(databasePath) {
+        let entries;
+        try {
+            entries = readdirSync(databasePath);
+        }
+        catch {
+            return false;
+        }
+        if (entries.length === 0)
+            return true;
+        return entries.some((e) => ProjectMetadataStore.MARKERS.has(e));
+    }
+    /**
+     * Whether a directory actually holds an index.
+     *
+     * Same markers as {@link isDatabaseDirectory}, without its allowance for
+     * an empty directory — the two are kept on one marker set so that what
+     * counts as a database is defined in exactly one place.
+     */
+    holdsIndex(databasePath) {
+        try {
+            return readdirSync(databasePath).some((e) => ProjectMetadataStore.MARKERS.has(e));
+        }
+        catch {
+            return false;
+        }
+    }
+}
+//# sourceMappingURL=ProjectMetadataStore.js.map

package/dist/infrastructure/fs/ProjectMetadataStore.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ProjectMetadataStore.js","sourceRoot":"","sources":["../../../src/infrastructure/fs/ProjectMetadataStore.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,WAAW,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AAC9E,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAGjC,OAAO,EAAE,SAAS,EAAE,MAAM,2BAA2B,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,OAAO,oBAAoB;IACxB,MAAM,CAAU,IAAI,GAAG,aAAa,CAAC;IAC7C,2DAA2D;IACnD,MAAM,CAAU,OAAO,GAAG,IAAI,GAAG,CAAS;QACjD,oBAAoB,CAAC,IAAI;QACzB,GAAG,SAAS,CAAC,MAAM,QAAQ;QAC3B,GAAG,SAAS,CAAC,KAAK,QAAQ;QAC1B,GAAG,SAAS,CAAC,WAAW,QAAQ;KAChC,CAAC,CAAC;IAEH,IAAI,CAAC,YAAoB;QACxB,IAAI,CAAC;YACJ,OAAO,IAAI,CAAC,KAAK,CAChB,YAAY,CAAC,IAAI,CAAC,YAAY,EAAE,oBAAoB,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,CACnD,CAAC;QACpB,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,SAAS,CAAC;QAClB,CAAC;IACF,CAAC;IAED;;;;;;OAMG;IACH,KAAK,CACJ,YAAoB,EACpB,QAA0C;QAE1C,SAAS,CAAC,YAAY,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAC7C,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACzC,MAAM,MAAM,GAAkB;YAC7B,IAAI,EAAE,QAAQ,CAAC,IAAI;YACnB,MAAM,EAAE,QAAQ,CAAC,MAAM;YACvB,MAAM,EAAE,QAAQ,CAAC,MAAM;YACvB,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACnC,KAAK,EAAE,QAAQ,EAAE,KAAK,IAAI,QAAQ,CAAC,KAAK;YACxC,UAAU,EAAE,QAAQ,EAAE,UAAU,IAAI,QAAQ,CAAC,UAAU;SACvD,CAAC;QACF,aAAa,CACZ,IAAI,CAAC,YAAY,EAAE,oBAAoB,CAAC,IAAI,CAAC,EAC7C,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAC/B,CAAC;IACH,CAAC;IAED;;;;;;;;;OASG;IACH,mBAAmB,CAAC,YAAoB;QACvC,IAAI,OAAiB,CAAC;QACtB,IAAI,CAAC;YACJ,OAAO,GAAG,WAAW,CAAC,YAAY,CAAC,CAAC;QACrC,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,IAAI,CAAC;QACtC,OAAO,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,oBAAoB,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IACjE,CAAC;IAED;;;;;;OAMG;IACH,UAAU,CAAC,YAAoB;QAC9B,IAAI,CAAC;YACJ,OAAO,WAAW,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAC3C,oBAAoB,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CACnC,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;IACF,CAAC"}

package/dist/infrastructure/fs/ProjectRegistry.d.ts

@@ -0,0 +1,25 @@
+import type { ProjectRegistryPort, RegisteredIndex } from "../../domain/ports/ProjectRegistryPort.js";
+import type { StateDirectoryPort } from "../../domain/ports/StateDirectoryPort.js";
+/**
+ * ProjectRegistryPort as one small JSON file per database.
+ *
+ * One file per entry rather than one file listing all of them, because several
+ * lmgrep processes index different projects at the same time — a shared list
+ * would need a lock and would still lose writes if one crashed mid-rewrite.
+ * Separate files make every write independent and every partial write
+ * survivable: a corrupt entry is skipped, not fatal to the listing.
+ */
+export declare class ProjectRegistry implements ProjectRegistryPort {
+    private readonly state;
+    private static readonly SUFFIX;
+    private static readonly KEY_LENGTH;
+    constructor(state: StateDirectoryPort);
+    record(entry: Omit<RegisteredIndex, "indexedAt">): void;
+    list(): RegisteredIndex[];
+    dangling(): RegisteredIndex[];
+    forget(databasePath: string): void;
+    private all;
+    private exists;
+    /** Keyed by database path, so re-indexing refreshes rather than duplicates. */
+    private pathFor;
+}

package/dist/infrastructure/fs/ProjectRegistry.js

@@ -0,0 +1,83 @@
+import { createHash } from "node:crypto";
+import { mkdirSync, readdirSync, readFileSync, statSync, unlinkSync, writeFileSync, } from "node:fs";
+import { join } from "node:path";
+/**
+ * ProjectRegistryPort as one small JSON file per database.
+ *
+ * One file per entry rather than one file listing all of them, because several
+ * lmgrep processes index different projects at the same time — a shared list
+ * would need a lock and would still lose writes if one crashed mid-rewrite.
+ * Separate files make every write independent and every partial write
+ * survivable: a corrupt entry is skipped, not fatal to the listing.
+ */
+export class ProjectRegistry {
+    state;
+    static SUFFIX = ".json";
+    static KEY_LENGTH = 16;
+    constructor(state) {
+        this.state = state;
+    }
+    record(entry) {
+        const directory = this.state.registryDirectory();
+        mkdirSync(directory, { recursive: true });
+        const full = {
+            ...entry,
+            indexedAt: new Date().toISOString(),
+        };
+        writeFileSync(this.pathFor(entry.databasePath), JSON.stringify(full, null, 2));
+    }
+    list() {
+        return this.all().filter((e) => this.exists(e.databasePath));
+    }
+    dangling() {
+        return this.all().filter((e) => !this.exists(e.databasePath));
+    }
+    forget(databasePath) {
+        try {
+            unlinkSync(this.pathFor(databasePath));
+        }
+        catch { }
+    }
+    all() {
+        let names;
+        try {
+            names = readdirSync(this.state.registryDirectory());
+        }
+        catch {
+            return [];
+        }
+        const out = [];
+        for (const name of names) {
+            if (!name.endsWith(ProjectRegistry.SUFFIX))
+                continue;
+            try {
+                const parsed = JSON.parse(readFileSync(join(this.state.registryDirectory(), name), "utf-8"));
+                if (typeof parsed.databasePath === "string" && parsed.databasePath) {
+                    out.push(parsed);
+                }
+            }
+            catch {
+                // A truncated or hand-edited entry is a lost pointer, never a
+                // reason to fail the listing that would have shown the rest.
+            }
+        }
+        return out;
+    }
+    exists(databasePath) {
+        try {
+            return statSync(databasePath).isDirectory();
+        }
+        catch {
+            return false;
+        }
+    }
+    /** Keyed by database path, so re-indexing refreshes rather than duplicates. */
+    pathFor(databasePath) {
+        const key = createHash("sha256")
+            .update(databasePath)
+            .digest("hex")
+            .slice(0, ProjectRegistry.KEY_LENGTH);
+        return join(this.state.registryDirectory(), `${key}${ProjectRegistry.SUFFIX}`);
+    }
+}
+//# sourceMappingURL=ProjectRegistry.js.map

package/dist/infrastructure/fs/ProjectRegistry.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ProjectRegistry.js","sourceRoot":"","sources":["../../../src/infrastructure/fs/ProjectRegistry.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EACN,SAAS,EACT,WAAW,EACX,YAAY,EACZ,QAAQ,EACR,UAAU,EACV,aAAa,GACb,MAAM,SAAS,CAAC;AACjB,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAOjC;;;;;;;;GAQG;AACH,MAAM,OAAO,eAAe;IAIE;IAHrB,MAAM,CAAU,MAAM,GAAG,OAAO,CAAC;IACjC,MAAM,CAAU,UAAU,GAAG,EAAE,CAAC;IAExC,YAA6B,KAAyB;QAAzB,UAAK,GAAL,KAAK,CAAoB;IAAG,CAAC;IAE1D,MAAM,CAAC,KAAyC;QAC/C,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,iBAAiB,EAAE,CAAC;QACjD,SAAS,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1C,MAAM,IAAI,GAAoB;YAC7B,GAAG,KAAK;YACR,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;SACnC,CAAC;QACF,aAAa,CACZ,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,YAAY,CAAC,EAChC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAC7B,CAAC;IACH,CAAC;IAED,IAAI;QACH,OAAO,IAAI,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC;IAC9D,CAAC;IAED,QAAQ;QACP,OAAO,IAAI,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC;IAC/D,CAAC;IAED,MAAM,CAAC,YAAoB;QAC1B,IAAI,CAAC;YACJ,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC;QACxC,CAAC;QAAC,MAAM,CAAC,CAAA,CAAC;IACX,CAAC;IAEO,GAAG;QACV,IAAI,KAAe,CAAC;QACpB,IAAI,CAAC;YACJ,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,iBAAiB,EAAE,CAAC,CAAC;QACrD,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,EAAE,CAAC;QACX,CAAC;QAED,MAAM,GAAG,GAAsB,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,eAAe,CAAC,MAAM,CAAC;gBAAE,SAAS;YACrD,IAAI,CAAC;gBACJ,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CACxB,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,iBAAiB,EAAE,EAAE,IAAI,CAAC,EAAE,OAAO,CAAC,CACrC,CAAC;gBAC9B,IAAI,OAAO,MAAM,CAAC,YAAY,KAAK,QAAQ,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;oBACpE,GAAG,CAAC,IAAI,CAAC,MAAyB,CAAC,CAAC;gBACrC,CAAC;YACF,CAAC;YAAC,MAAM,CAAC;gBACR,8DAA8D;gBAC9D,6DAA6D;YAC9D,CAAC;QACF,CAAC;QACD,OAAO,GAAG,CAAC;IACZ,CAAC;IAEO,MAAM,CAAC,YAAoB;QAClC,IAAI,CAAC;YACJ,OAAO,QAAQ,CAAC,YAAY,CAAC,CAAC,WAAW,EAAE,CAAC;QAC7C,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;IACF,CAAC;IAED,+EAA+E;IACvE,OAAO,CAAC,YAAoB;QACnC,MAAM,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC;aAC9B,MAAM,CAAC,YAAY,CAAC;aACpB,MAAM,CAAC,KAAK,CAAC;aACb,KAAK,CAAC,CAAC,EAAE,eAAe,CAAC,UAAU,CAAC,CAAC;QACvC,OAAO,IAAI,CACV,IAAI,CAAC,KAAK,CAAC,iBAAiB,EAAE,EAC9B,GAAG,GAAG,GAAG,eAAe,CAAC,MAAM,EAAE,CACjC,CAAC;IACH,CAAC"}

package/dist/infrastructure/fs/StateDirectory.d.ts

@@ -0,0 +1,28 @@
+import type { StateDirectoryPort } from "../../domain/ports/StateDirectoryPort.js";
+/**
+ * StateDirectoryPort over the real filesystem, rooted at the XDG state
+ * location.
+ *
+ * The subdirectories are new; earlier versions wrote databases and their lock
+ * files as immediate children of the root, interleaved. {@link RESERVED} is
+ * what keeps the two schemes apart, so a legacy scan never mistakes the
+ * registry for a database.
+ *
+ * Existence probes swallow their errors: an unreadable or missing path is
+ * simply "not a database", never a failure.
+ */
+export declare class StateDirectory implements StateDirectoryPort {
+    private static readonly LOCKS;
+    private static readonly REGISTRY;
+    private static readonly DATABASES;
+    /** Subdirectories of the state root that are not legacy databases. */
+    private static readonly RESERVED;
+    private readonly base;
+    constructor(base?: string);
+    root(): string;
+    isDirectory(path: string): boolean;
+    locksDirectory(): string;
+    registryDirectory(): string;
+    databasesDirectory(): string;
+    legacyDatabaseDirectories(): string[];
+}

package/dist/infrastructure/fs/StateDirectory.js

@@ -0,0 +1,71 @@
+import { readdirSync, statSync } from "node:fs";
+import { homedir } from "node:os";
+import { join } from "node:path";
+/**
+ * StateDirectoryPort over the real filesystem, rooted at the XDG state
+ * location.
+ *
+ * The subdirectories are new; earlier versions wrote databases and their lock
+ * files as immediate children of the root, interleaved. {@link RESERVED} is
+ * what keeps the two schemes apart, so a legacy scan never mistakes the
+ * registry for a database.
+ *
+ * Existence probes swallow their errors: an unreadable or missing path is
+ * simply "not a database", never a failure.
+ */
+export class StateDirectory {
+    static LOCKS = "locks";
+    static REGISTRY = "registry";
+    static DATABASES = "db";
+    /** Subdirectories of the state root that are not legacy databases. */
+    static RESERVED = new Set([
+        StateDirectory.LOCKS,
+        StateDirectory.REGISTRY,
+        StateDirectory.DATABASES,
+    ]);
+    base;
+    constructor(base) {
+        // LMGREP_STATE_DIR relocates everything machine-global in one step,
+        // which is what makes an isolated run possible without also moving
+        // HOME — moving HOME breaks provider resolution, since AI SDK packages
+        // are found through the user's global node_modules.
+        this.base =
+            base ??
+                process.env.LMGREP_STATE_DIR ??
+                join(homedir(), ".local", "state", "lmgrep");
+    }
+    root() {
+        return this.base;
+    }
+    isDirectory(path) {
+        try {
+            return statSync(path).isDirectory();
+        }
+        catch {
+            return false;
+        }
+    }
+    locksDirectory() {
+        return join(this.base, StateDirectory.LOCKS);
+    }
+    registryDirectory() {
+        return join(this.base, StateDirectory.REGISTRY);
+    }
+    databasesDirectory() {
+        return join(this.base, StateDirectory.DATABASES);
+    }
+    legacyDatabaseDirectories() {
+        let entries;
+        try {
+            entries = readdirSync(this.base);
+        }
+        catch {
+            return [];
+        }
+        return entries
+            .filter((name) => !StateDirectory.RESERVED.has(name))
+            .map((name) => join(this.base, name))
+            .filter((p) => this.isDirectory(p));
+    }
+}
+//# sourceMappingURL=StateDirectory.js.map

package/dist/infrastructure/fs/StateDirectory.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"StateDirectory.js","sourceRoot":"","sources":["../../../src/infrastructure/fs/StateDirectory.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAChD,OAAO,EAAE,OAAO,EAAE,MAAM,SAAS,CAAC;AAClC,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAGjC;;;;;;;;;;;GAWG;AACH,MAAM,OAAO,cAAc;IAClB,MAAM,CAAU,KAAK,GAAG,OAAO,CAAC;IAChC,MAAM,CAAU,QAAQ,GAAG,UAAU,CAAC;IACtC,MAAM,CAAU,SAAS,GAAG,IAAI,CAAC;IACzC,sEAAsE;IAC9D,MAAM,CAAU,QAAQ,GAAG,IAAI,GAAG,CAAS;QAClD,cAAc,CAAC,KAAK;QACpB,cAAc,CAAC,QAAQ;QACvB,cAAc,CAAC,SAAS;KACxB,CAAC,CAAC;IAEc,IAAI,CAAS;IAE9B,YAAY,IAAa;QACxB,oEAAoE;QACpE,mEAAmE;QACnE,uEAAuE;QACvE,oDAAoD;QACpD,IAAI,CAAC,IAAI;YACR,IAAI;gBACJ,OAAO,CAAC,GAAG,CAAC,gBAAgB;gBAC5B,IAAI,CAAC,OAAO,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC;IAC/C,CAAC;IAED,IAAI;QACH,OAAO,IAAI,CAAC,IAAI,CAAC;IAClB,CAAC;IAED,WAAW,CAAC,IAAY;QACvB,IAAI,CAAC;YACJ,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,CAAC;QACrC,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;IACF,CAAC;IAED,cAAc;QACb,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,KAAK,CAAC,CAAC;IAC9C,CAAC;IAED,iBAAiB;QAChB,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,QAAQ,CAAC,CAAC;IACjD,CAAC;IAED,kBAAkB;QACjB,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,SAAS,CAAC,CAAC;IAClD,CAAC;IAED,yBAAyB;QACxB,IAAI,OAAiB,CAAC;QACtB,IAAI,CAAC;YACJ,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClC,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,EAAE,CAAC;QACX,CAAC;QACD,OAAO,OAAO;aACZ,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,cAAc,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;aACpD,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;aACpC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;IACtC,CAAC"}

package/dist/infrastructure/fs/Workspace.d.ts

@@ -0,0 +1,25 @@
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import type { ChangeSet, ExtensionRules, WatchHandle, WorkspacePort } from "../../domain/ports/WorkspacePort.js";
+/** The working tree on the real filesystem. */
+export declare class Workspace implements WorkspacePort {
+    listFiles(cwd: string, extraIgnore?: string[], extensions?: ExtensionRules): string[];
+    hashOf(cwd: string, filePath: string): ContentHash | undefined;
+    modifiedSince(files: string[], cwd: string, cutoffMs: number): string[];
+    /**
+     * Compare on-disk hashes against what the manifest recorded.
+     *
+     * Unreadable files are skipped rather than reported as changed — a
+     * transient read error should not evict a file from the index.
+     */
+    detectChanges(files: string[], manifest: {
+        versionOf(path: string): ContentHash | undefined;
+    }, cwd: string, force?: boolean): ChangeSet;
+    /**
+     * Watch for changes, coalescing a burst into one callback.
+     *
+     * fs.watch's recursive mode misses events on Linux (new subdirectories,
+     * editor atomic saves); callers pair this with a periodic reconcile rather
+     * than trusting it alone.
+     */
+    watch(cwd: string, extraIgnore: string[] | undefined, onChanges: (changedFiles: string[]) => void, debounceMs: number, extensions?: ExtensionRules): WatchHandle;
+}

package/dist/infrastructure/fs/Workspace.js

@@ -0,0 +1,90 @@
+import { readFileSync, statSync, watch } from "node:fs";
+import { join } from "node:path";
+import { globSync } from "glob";
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import { SourceFile } from "../../domain/corpus/SourceFile.js";
+import { IndexableFileRules } from "./IndexableFileRules.js";
+/** The working tree on the real filesystem. */
+export class Workspace {
+    listFiles(cwd, extraIgnore, extensions) {
+        const rules = new IndexableFileRules(cwd, extraIgnore, extensions);
+        const all = globSync("**/*", { cwd, nodir: true, dot: false });
+        rules.loadNestedIgnores(all);
+        return all.filter((f) => rules.admits(f));
+    }
+    hashOf(cwd, filePath) {
+        try {
+            return ContentHash.of(readFileSync(join(cwd, filePath)));
+        }
+        catch {
+            return undefined;
+        }
+    }
+    modifiedSince(files, cwd, cutoffMs) {
+        return files.filter((f) => {
+            try {
+                return statSync(join(cwd, f)).mtimeMs >= cutoffMs;
+            }
+            catch {
+                return false;
+            }
+        });
+    }
+    /**
+     * Compare on-disk hashes against what the manifest recorded.
+     *
+     * Unreadable files are skipped rather than reported as changed — a
+     * transient read error should not evict a file from the index.
+     */
+    detectChanges(files, manifest, cwd, force = false) {
+        const changed = [];
+        const current = new Map();
+        for (const file of files) {
+            const hash = this.hashOf(cwd, file);
+            if (!hash)
+                continue;
+            current.set(file, hash);
+            const stored = manifest.versionOf(file);
+            if (force || stored === undefined || !stored.equals(hash)) {
+                changed.push(new SourceFile(file, hash));
+            }
+        }
+        return { changed, current };
+    }
+    /**
+     * Watch for changes, coalescing a burst into one callback.
+     *
+     * fs.watch's recursive mode misses events on Linux (new subdirectories,
+     * editor atomic saves); callers pair this with a periodic reconcile rather
+     * than trusting it alone.
+     */
+    watch(cwd, extraIgnore, onChanges, debounceMs, extensions) {
+        const rules = new IndexableFileRules(cwd, extraIgnore, extensions);
+        let timer;
+        let pending = new Set();
+        const watcher = watch(cwd, { recursive: true }, (_event, filename) => {
+            if (!filename)
+                return;
+            if (!rules.hasIndexableExtension(filename))
+                return;
+            if (rules.isIgnored(filename))
+                return;
+            pending.add(filename);
+            if (timer)
+                clearTimeout(timer);
+            timer = setTimeout(() => {
+                const files = [...pending];
+                pending = new Set();
+                onChanges(files);
+            }, debounceMs);
+        });
+        return {
+            close() {
+                if (timer)
+                    clearTimeout(timer);
+                watcher.close();
+            },
+        };
+    }
+}
+//# sourceMappingURL=Workspace.js.map

package/dist/infrastructure/fs/Workspace.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"Workspace.js","sourceRoot":"","sources":["../../../src/infrastructure/fs/Workspace.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,SAAS,CAAC;AACxD,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AACjC,OAAO,EAAE,QAAQ,EAAE,MAAM,MAAM,CAAC;AAChC,OAAO,EAAE,WAAW,EAAE,MAAM,oCAAoC,CAAC;AACjE,OAAO,EAAE,UAAU,EAAE,MAAM,mCAAmC,CAAC;AAO/D,OAAO,EAAE,kBAAkB,EAAE,MAAM,yBAAyB,CAAC;AAE7D,+CAA+C;AAC/C,MAAM,OAAO,SAAS;IACrB,SAAS,CACR,GAAW,EACX,WAAsB,EACtB,UAA2B;QAE3B,MAAM,KAAK,GAAG,IAAI,kBAAkB,CAAC,GAAG,EAAE,WAAW,EAAE,UAAU,CAAC,CAAC;QACnE,MAAM,GAAG,GAAG,QAAQ,CAAC,MAAM,EAAE,EAAE,GAAG,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAC;QAC/D,KAAK,CAAC,iBAAiB,CAAC,GAAG,CAAC,CAAC;QAC7B,OAAO,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAC3C,CAAC;IAED,MAAM,CAAC,GAAW,EAAE,QAAgB;QACnC,IAAI,CAAC;YACJ,OAAO,WAAW,CAAC,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC;QAC1D,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,SAAS,CAAC;QAClB,CAAC;IACF,CAAC;IAED,aAAa,CAAC,KAAe,EAAE,GAAW,EAAE,QAAgB;QAC3D,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YACzB,IAAI,CAAC;gBACJ,OAAO,QAAQ,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,IAAI,QAAQ,CAAC;YACnD,CAAC;YAAC,MAAM,CAAC;gBACR,OAAO,KAAK,CAAC;YACd,CAAC;QACF,CAAC,CAAC,CAAC;IACJ,CAAC;IAED;;;;;OAKG;IACH,aAAa,CACZ,KAAe,EACf,QAA8D,EAC9D,GAAW,EACX,KAAK,GAAG,KAAK;QAEb,MAAM,OAAO,GAAiB,EAAE,CAAC;QACjC,MAAM,OAAO,GAAG,IAAI,GAAG,EAAuB,CAAC;QAE/C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;YACpC,IAAI,CAAC,IAAI;gBAAE,SAAS;YACpB,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;YACxB,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,KAAK,IAAI,MAAM,KAAK,SAAS,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC3D,OAAO,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;YAC1C,CAAC;QACF,CAAC;QAED,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC;IAC7B,CAAC;IAED;;;;;;OAMG;IACH,KAAK,CACJ,GAAW,EACX,WAAiC,EACjC,SAA2C,EAC3C,UAAkB,EAClB,UAA2B;QAE3B,MAAM,KAAK,GAAG,IAAI,kBAAkB,CAAC,GAAG,EAAE,WAAW,EAAE,UAAU,CAAC,CAAC;QACnE,IAAI,KAAgD,CAAC;QACrD,IAAI,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;QAEhC,MAAM,OAAO,GAAG,KAAK,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,EAAE,CAAC,MAAM,EAAE,QAAQ,EAAE,EAAE;YACpE,IAAI,CAAC,QAAQ;gBAAE,OAAO;YACtB,IAAI,CAAC,KAAK,CAAC,qBAAqB,CAAC,QAAQ,CAAC;gBAAE,OAAO;YACnD,IAAI,KAAK,CAAC,SAAS,CAAC,QAAQ,CAAC;gBAAE,OAAO;YAEtC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;YACtB,IAAI,KAAK;gBAAE,YAAY,CAAC,KAAK,CAAC,CAAC;YAC/B,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE;gBACvB,MAAM,KAAK,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC;gBAC3B,OAAO,GAAG,IAAI,GAAG,EAAE,CAAC;gBACpB,SAAS,CAAC,KAAK,CAAC,CAAC;YAClB,CAAC,EAAE,UAAU,CAAC,CAAC;QAChB,CAAC,CAAC,CAAC;QAEH,OAAO;YACN,KAAK;gBACJ,IAAI,KAAK;oBAAE,YAAY,CAAC,KAAK,CAAC,CAAC;gBAC/B,OAAO,CAAC,KAAK,EAAE,CAAC;YACjB,CAAC;SACD,CAAC;IACH,CAAC;CACD"}

package/dist/infrastructure/git/GitClient.d.ts

@@ -0,0 +1,34 @@
+import type { GitPort } from "../../domain/ports/GitPort.js";
+/**
+ * GitPort over the `git` binary.
+ *
+ * Every call is best-effort: a missing binary, a non-repo directory, or a
+ * detached/oprhaned state all surface as `undefined` rather than throwing,
+ * because every caller treats "not a git repo" as an ordinary case.
+ */
+export declare class GitClient implements GitPort {
+    private static readonly TIMEOUT_MS;
+    toplevel(cwd: string): string | undefined;
+    /**
+     * `--git-common-dir` reports a path *relative to the invocation directory*
+     * (`../../.git` from a subdirectory), so it is resolved against that
+     * directory rather than used as given. `--path-format=absolute` would do
+     * the same but only on git 2.31 and later.
+     */
+    commonDir(cwd: string): string | undefined;
+    currentBranch(repoRoot: string): string | undefined;
+    originUrl(repoRoot: string): string | undefined;
+    mergeBase(repoRoot: string, ref: string): string | undefined;
+    commitDistance(repoRoot: string, from: string, to: string): number | undefined;
+    localBranches(repoRoot: string): string[];
+    /**
+     * Run git with an argv array rather than a command string.
+     *
+     * This must not go through a shell. Git's own format specifiers contain
+     * parentheses (`--format=%(refname:short)`), and branch names may legally
+     * contain `(`, `)`, `&`, `;`, `$` and quotes — all of which a shell would
+     * interpret. Passing a string here silently broke branch listing entirely
+     * and made any branch name with a metacharacter unusable.
+     */
+    private run;
+}

package/dist/infrastructure/git/GitClient.js

@@ -0,0 +1,74 @@
+import { execFileSync } from "node:child_process";
+import { resolve } from "node:path";
+/**
+ * GitPort over the `git` binary.
+ *
+ * Every call is best-effort: a missing binary, a non-repo directory, or a
+ * detached/oprhaned state all surface as `undefined` rather than throwing,
+ * because every caller treats "not a git repo" as an ordinary case.
+ */
+export class GitClient {
+    static TIMEOUT_MS = 5000;
+    toplevel(cwd) {
+        return this.run(cwd, "rev-parse", "--show-toplevel");
+    }
+    /**
+     * `--git-common-dir` reports a path *relative to the invocation directory*
+     * (`../../.git` from a subdirectory), so it is resolved against that
+     * directory rather than used as given. `--path-format=absolute` would do
+     * the same but only on git 2.31 and later.
+     */
+    commonDir(cwd) {
+        const out = this.run(cwd, "rev-parse", "--git-common-dir");
+        return out === undefined ? undefined : resolve(cwd, out);
+    }
+    currentBranch(repoRoot) {
+        return this.run(repoRoot, "rev-parse", "--abbrev-ref", "HEAD");
+    }
+    originUrl(repoRoot) {
+        return this.run(repoRoot, "remote", "get-url", "origin");
+    }
+    mergeBase(repoRoot, ref) {
+        return this.run(repoRoot, "merge-base", "HEAD", ref);
+    }
+    commitDistance(repoRoot, from, to) {
+        const out = this.run(repoRoot, "rev-list", "--count", `${from}..${to}`);
+        if (out === undefined)
+            return undefined;
+        const n = Number.parseInt(out, 10);
+        return Number.isNaN(n) ? undefined : n;
+    }
+    localBranches(repoRoot) {
+        const out = this.run(repoRoot, "for-each-ref", "--format=%(refname:short)", "refs/heads");
+        if (!out)
+            return [];
+        return out
+            .split("\n")
+            .map((l) => l.trim())
+            .filter(Boolean);
+    }
+    /**
+     * Run git with an argv array rather than a command string.
+     *
+     * This must not go through a shell. Git's own format specifiers contain
+     * parentheses (`--format=%(refname:short)`), and branch names may legally
+     * contain `(`, `)`, `&`, `;`, `$` and quotes — all of which a shell would
+     * interpret. Passing a string here silently broke branch listing entirely
+     * and made any branch name with a metacharacter unusable.
+     */
+    run(cwd, ...args) {
+        try {
+            return execFileSync("git", args, {
+                cwd,
+                stdio: ["ignore", "pipe", "ignore"],
+                timeout: GitClient.TIMEOUT_MS,
+            })
+                .toString()
+                .trim();
+        }
+        catch {
+            return undefined;
+        }
+    }
+}
+//# sourceMappingURL=GitClient.js.map

package/dist/infrastructure/git/GitClient.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"GitClient.js","sourceRoot":"","sources":["../../../src/infrastructure/git/GitClient.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAGpC;;;;;;GAMG;AACH,MAAM,OAAO,SAAS;IACb,MAAM,CAAU,UAAU,GAAG,IAAI,CAAC;IAE1C,QAAQ,CAAC,GAAW;QACnB,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,WAAW,EAAE,iBAAiB,CAAC,CAAC;IACtD,CAAC;IAED;;;;;OAKG;IACH,SAAS,CAAC,GAAW;QACpB,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,WAAW,EAAE,kBAAkB,CAAC,CAAC;QAC3D,OAAO,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC;IAC1D,CAAC;IAED,aAAa,CAAC,QAAgB;QAC7B,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,WAAW,EAAE,cAAc,EAAE,MAAM,CAAC,CAAC;IAChE,CAAC;IAED,SAAS,CAAC,QAAgB;QACzB,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;IAC1D,CAAC;IAED,SAAS,CAAC,QAAgB,EAAE,GAAW;QACtC,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,YAAY,EAAE,MAAM,EAAE,GAAG,CAAC,CAAC;IACtD,CAAC;IAED,cAAc,CACb,QAAgB,EAChB,IAAY,EACZ,EAAU;QAEV,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,UAAU,EAAE,SAAS,EAAE,GAAG,IAAI,KAAK,EAAE,EAAE,CAAC,CAAC;QACxE,IAAI,GAAG,KAAK,SAAS;YAAE,OAAO,SAAS,CAAC;QACxC,MAAM,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACnC,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;IACxC,CAAC;IAED,aAAa,CAAC,QAAgB;QAC7B,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CACnB,QAAQ,EACR,cAAc,EACd,2BAA2B,EAC3B,YAAY,CACZ,CAAC;QACF,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,CAAC;QACpB,OAAO,GAAG;aACR,KAAK,CAAC,IAAI,CAAC;aACX,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;aACpB,MAAM,CAAC,OAAO,CAAC,CAAC;IACnB,CAAC;IAED;;;;;;;;OAQG;IACK,GAAG,CAAC,GAAW,EAAE,GAAG,IAAc;QACzC,IAAI,CAAC;YACJ,OAAO,YAAY,CAAC,KAAK,EAAE,IAAI,EAAE;gBAChC,GAAG;gBACH,KAAK,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC;gBACnC,OAAO,EAAE,SAAS,CAAC,UAAU;aAC7B,CAAC;iBACA,QAAQ,EAAE;iBACV,IAAI,EAAE,CAAC;QACV,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,SAAS,CAAC;QAClB,CAAC;IACF,CAAC"}

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

@@ -0,0 +1,38 @@
+import { Chunk } from "../../domain/corpus/Chunk.js";
+import { ContentHash } from "../../domain/corpus/ContentHash.js";
+import type { ChunkQuery, ChunkRepositoryPort, ChunkText, EmbeddedChunk } from "../../domain/ports/ChunkRepositoryPort.js";
+import type { FileManifestRepositoryPort } from "../../domain/ports/FileManifestRepositoryPort.js";
+import type { Branch } from "../../domain/project/Branch.js";
+import { HitList } from "../../domain/retrieval/HitList.js";
+import { LanceTables } from "./LanceTables.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 declare class ChunkRepository implements ChunkRepositoryPort {
+    private readonly tables;
+    private readonly manifest;
+    private readonly branch;
+    constructor(tables: LanceTables, manifest: FileManifestRepositoryPort, branch: Branch);
+    add(chunks: EmbeddedChunk[]): Promise<void>;
+    search(query: ChunkQuery): Promise<HitList>;
+    /**
+     * 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.
+     */
+    deleteByFiles(filePaths: string[]): Promise<void>;
+    existingHashes(hashes: ContentHash[]): Promise<Set<string>>;
+    count(): Promise<number>;
+    hashesByFile(): Promise<Map<string, string[]>>;
+    allHashes(): Promise<Set<string>>;
+    streamTexts(batchSize?: number): AsyncGenerator<ChunkText[]>;
+    private buildPredicate;
+    private toRow;
+    private toHit;
+    /** Rebuild a domain Chunk from a stored row — used by import and dedup. */
+    static rowToChunk(row: Record<string, unknown>): Chunk;
+}