lmgrep 0.1.17 → 0.2.0

package/dist/application/Lmgrep.js

@@ -0,0 +1,99 @@
+import { existsSync } from "node:fs";
+import { SearchCriteria } from "./search/SearchCriteria.js";
+/**
+ * The application façade: one object exposing every operation the CLI, the MCP
+ * server and the Pi extension need.
+ *
+ * It holds no logic of its own — each method delegates to the service that
+ * owns that behaviour. Its job is to be the single place where a caller
+ * acquires a working, wired-up lmgrep, so no entry point has to know the
+ * assembly order.
+ */
+export class Lmgrep {
+    services;
+    constructor(services) {
+        this.services = services;
+    }
+    get cwd() {
+        return this.services.cwd;
+    }
+    get config() {
+        return this.services.config;
+    }
+    get location() {
+        return this.services.location;
+    }
+    build(options = {}) {
+        return this.services.builder.build(options);
+    }
+    search(query, options = {}) {
+        return this.services.searcher.search(query, new SearchCriteria(options));
+    }
+    ask(question, onTrace) {
+        return this.services.researcher.research(question, onTrace);
+    }
+    status() {
+        return this.services.statusReporter.status();
+    }
+    /** Start watching; returns a function that stops it. */
+    watch() {
+        return this.services.watcher.start();
+    }
+    /** The current branch's file manifest, for change previews. */
+    currentManifest() {
+        return this.services.manifest.current();
+    }
+    /**
+     * Everything the index needs done to it that is not indexing: drop dead
+     * branches' manifests, remove duplicate and superseded rows, compact the
+     * fragments, and train the vector index if the table has grown into
+     * wanting one.
+     *
+     * These were three separate commands. They are one call because they are
+     * one intent and because the order between them is not optional: `dedupe`
+     * keeps any chunk version some manifest still references, so sweeping dead
+     * branches has to happen first or their leftover rows shield their own
+     * orphaned chunks from collection.
+     */
+    async tidy() {
+        await this.services.sweeper.sweep(this.services.location.root);
+        const deduped = await this.services.maintenance.dedupe();
+        const optimized = await this.services.maintenance.compact();
+        return { deduped, optimized };
+    }
+    get maintenance() {
+        return this.services.maintenance;
+    }
+    /** Every index this machine knows about. */
+    get registry() {
+        return this.services.registry;
+    }
+    /** This project's indexes under other embedding models. */
+    get alternatives() {
+        return this.services.alternatives;
+    }
+    get projectId() {
+        return this.services.projectId();
+    }
+    /**
+     * Whether this working directory has a usable index.
+     *
+     * An indexed ancestor counts: running from a subdirectory of an indexed
+     * tree is ordinary, and reporting it as unindexed would disable search for
+     * no reason. A manually targeted database is flat, so only its own path
+     * matters.
+     */
+    isIndexed() {
+        if (this.services.location.manual) {
+            return existsSync(this.services.location.path);
+        }
+        if (this.services.locator.findIndexedAncestor(this.services.cwd)) {
+            return true;
+        }
+        return existsSync(this.services.location.path);
+    }
+    async close() {
+        this.services.tables.close();
+    }
+}
+//# sourceMappingURL=Lmgrep.js.map

package/dist/application/Lmgrep.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"Lmgrep.js","sourceRoot":"","sources":["../../src/application/Lmgrep.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAkCrC,OAAO,EAAE,cAAc,EAAsB,MAAM,4BAA4B,CAAC;AAkChF;;;;;;;;GAQG;AACH,MAAM,OAAO,MAAM;IACW;IAA7B,YAA6B,QAAwB;QAAxB,aAAQ,GAAR,QAAQ,CAAgB;IAAG,CAAC;IAEzD,IAAI,GAAG;QACN,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;IAC1B,CAAC;IAED,IAAI,MAAM;QACT,OAAO,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;IAC7B,CAAC;IAED,IAAI,QAAQ;QACX,OAAO,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;IAC/B,CAAC;IAED,KAAK,CAAC,UAA6B,EAAE;QACpC,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC7C,CAAC;IAED,MAAM,CAAC,KAAa,EAAE,UAAyB,EAAE;QAChD,OAAO,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,IAAI,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC;IAC1E,CAAC;IAED,GAAG,CACF,QAAgB,EAChB,OAAqC;QAErC,OAAO,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC7D,CAAC;IAED,MAAM;QACL,OAAO,IAAI,CAAC,QAAQ,CAAC,cAAc,CAAC,MAAM,EAAE,CAAC;IAC9C,CAAC;IAED,wDAAwD;IACxD,KAAK;QACJ,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;IACtC,CAAC;IAED,+DAA+D;IAC/D,eAAe;QACd,OAAO,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC;IACzC,CAAC;IAED;;;;;;;;;;;OAWG;IACH,KAAK,CAAC,IAAI;QACT,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAC/D,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,MAAM,EAAE,CAAC;QACzD,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,OAAO,EAAE,CAAC;QAC5D,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;IAC/B,CAAC;IAED,IAAI,WAAW;QACd,OAAO,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC;IAClC,CAAC;IAED,4CAA4C;IAC5C,IAAI,QAAQ;QACX,OAAO,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;IAC/B,CAAC;IAED,2DAA2D;IAC3D,IAAI,YAAY;QACf,OAAO,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC;IACnC,CAAC;IAED,IAAI,SAAS;QACZ,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,EAAE,CAAC;IAClC,CAAC;IAED;;;;;;;OAOG;IACH,SAAS;QACR,IAAI,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC;YACnC,OAAO,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAChD,CAAC;QACD,IAAI,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,mBAAmB,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;YAClE,OAAO,IAAI,CAAC;QACb,CAAC;QACD,OAAO,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAChD,CAAC;IAED,KAAK,CAAC,KAAK;QACV,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;IAC9B,CAAC;CACD"}

package/dist/application/LmgrepFactory.d.ts

@@ -0,0 +1,40 @@
+import type { LmgrepConfig } from "../domain/config/LmgrepConfig.js";
+import type { ChunkerPort } from "../domain/ports/ChunkerPort.js";
+import type { EmbedderPort } from "../domain/ports/EmbedderPort.js";
+import type { LoggerPort } from "../domain/ports/LoggerPort.js";
+import { Lmgrep } from "./Lmgrep.js";
+export interface LmgrepOptions {
+    cwd: string;
+    /**
+     * Target a specific database instead of the git-aware default. A bare name
+     * creates an independent index; a path points at a database directory.
+     */
+    database?: string;
+    config?: Partial<LmgrepConfig>;
+    embedder?: EmbedderPort;
+    chunker?: ChunkerPort;
+    logger?: LoggerPort;
+    /**
+     * Where configuration problems go. Separate from the logger because these
+     * are reported before anything is wired, and because the MCP server must
+     * silence them — stdout is its transport.
+     */
+    onWarning?: (message: string) => void;
+}
+/**
+ * Assembles a working {@link Lmgrep}.
+ *
+ * This is the only place that knows the full object graph. Everything else
+ * receives its collaborators, which is what keeps the layers testable and the
+ * dependency direction one-way.
+ */
+export declare class LmgrepFactory {
+    open(options: LmgrepOptions): Promise<Lmgrep>;
+    /**
+     * Opens another project's database read-only for cross-project search.
+     *
+     * Foreign indexes get their own connection and are closed after the query,
+     * so a `--across` over many projects does not leave native runtimes open.
+     */
+    private foreignOpener;
+}

package/dist/application/LmgrepFactory.js

@@ -0,0 +1,155 @@
+import { ModelIdentity } from "../domain/project/ModelIdentity.js";
+import { ProjectLocator } from "../domain/project/ProjectLocator.js";
+import { AiSdkChatModel } from "../infrastructure/ai/AiSdkChatModel.js";
+import { AiSdkEmbedder } from "../infrastructure/ai/AiSdkEmbedder.js";
+import { LocalModelReloader } from "../infrastructure/ai/LocalModelReloader.js";
+import { ConfigLoader } from "../infrastructure/fs/ConfigLoader.js";
+import { DatabaseLocks } from "../infrastructure/fs/DatabaseLocks.js";
+import { ConsoleLogger } from "../infrastructure/fs/Loggers.js";
+import { ProjectIndexes } from "../infrastructure/fs/ProjectIndexes.js";
+import { ProjectMetadataStore } from "../infrastructure/fs/ProjectMetadataStore.js";
+import { ProjectRegistry } from "../infrastructure/fs/ProjectRegistry.js";
+import { StateDirectory } from "../infrastructure/fs/StateDirectory.js";
+import { Workspace } from "../infrastructure/fs/Workspace.js";
+import { GitClient } from "../infrastructure/git/GitClient.js";
+import { ChunkRepository } from "../infrastructure/lancedb/ChunkRepository.js";
+import { FileManifestRepository } from "../infrastructure/lancedb/FileManifestRepository.js";
+import { IndexMaintenance } from "../infrastructure/lancedb/IndexMaintenance.js";
+import { LanceTables } from "../infrastructure/lancedb/LanceTables.js";
+import { TreeSitterChunker } from "../infrastructure/treesitter/TreeSitterChunker.js";
+import { BranchBootstrapper } from "./indexing/BranchBootstrapper.js";
+import { BranchManifestSweeper } from "./indexing/BranchManifestSweeper.js";
+import { IndexBuilder } from "./indexing/IndexBuilder.js";
+import { Lmgrep } from "./Lmgrep.js";
+import { IndexAlternatives } from "./operations/IndexAlternatives.js";
+import { StatusService } from "./operations/StatusService.js";
+import { WatchService } from "./operations/WatchService.js";
+import { ResearchAgent } from "./research/ResearchAgent.js";
+import { SearchService } from "./search/SearchService.js";
+import { SearchTargetResolver, } from "./search/SearchTargetResolver.js";
+/**
+ * Assembles a working {@link Lmgrep}.
+ *
+ * This is the only place that knows the full object graph. Everything else
+ * receives its collaborators, which is what keeps the layers testable and the
+ * dependency direction one-way.
+ */
+export class LmgrepFactory {
+    async open(options) {
+        const { cwd } = options;
+        const logger = options.logger ?? new ConsoleLogger();
+        const loader = new ConfigLoader();
+        const config = {
+            ...loader.load(cwd),
+            ...options.config,
+        };
+        for (const warning of loader.warnings)
+            options.onWarning?.(warning);
+        const state = new StateDirectory();
+        const git = new GitClient();
+        // The model partitions the databases, so it has to be known before a
+        // location can be resolved — which is why config loads first.
+        const locator = new ProjectLocator(git, state, ModelIdentity.of(config.model), config.dimensions);
+        const location = locator.resolveDatabase(cwd, options.database);
+        const metadata = new ProjectMetadataStore();
+        const registry = new ProjectRegistry(state);
+        const alternatives = new IndexAlternatives(new ProjectIndexes(metadata), locator, location, config, cwd);
+        const locks = new DatabaseLocks(state.locksDirectory(), location.path, location.root);
+        const tables = new LanceTables(location.path, location.branch);
+        const manifest = new FileManifestRepository(tables, location.branch);
+        const chunks = new ChunkRepository(tables, manifest, location.branch);
+        const maintenance = new IndexMaintenance(tables, manifest);
+        const embedder = options.embedder ?? new AiSdkEmbedder(config);
+        const chunker = options.chunker ?? new TreeSitterChunker();
+        const workspace = new Workspace();
+        const reloader = new LocalModelReloader(config);
+        const sweeper = new BranchManifestSweeper(manifest, git, logger);
+        const builder = new IndexBuilder({
+            workspace,
+            chunker,
+            embedder,
+            chunks,
+            manifest,
+            maintenance,
+            bootstrapper: new BranchBootstrapper(manifest, git, logger),
+            sweeper,
+            logger,
+            config,
+            location,
+            locks,
+            recordMetadata: (dimensions) => {
+                const project = locator.resolveProject(cwd);
+                metadata.write(location.path, {
+                    root: project.root,
+                    remote: project.remote,
+                    branch: location.branch.toString(),
+                    model: config.model,
+                    dimensions,
+                });
+            },
+            // The sidecar says what this database is; the registry says that
+            // it exists. Nothing walks the filesystem looking for databases
+            // now that they live inside repositories, so an index that is
+            // never registered is one `lmgrep projects` and cross-project
+            // search cannot see.
+            registerIndex: () => {
+                const project = locator.resolveProject(cwd);
+                registry.record({
+                    root: locator.projectRootFor(cwd),
+                    name: location.manual ? options.database : undefined,
+                    remote: project.remote,
+                    databasePath: location.path,
+                    model: config.model,
+                    dimensions: metadata.read(location.path)?.dimensions,
+                });
+            },
+            reloadModel: () => reloader.reload(),
+            isLocalProvider: reloader.isLocal,
+        });
+        const searcher = new SearchService(embedder, new SearchTargetResolver(cwd, chunks, location, locator, this.foreignOpener(locator)), config, logger, () => metadata.read(location.path), alternatives);
+        const services = {
+            cwd,
+            config,
+            location,
+            locator,
+            logger,
+            tables,
+            chunks,
+            manifest,
+            maintenance,
+            metadata,
+            registry,
+            alternatives,
+            embedder,
+            chunker,
+            builder,
+            sweeper,
+            searcher,
+            statusReporter: new StatusService(chunks, maintenance, embedder, locator, metadata, location, config, cwd, loader.sources, alternatives),
+            watcher: new WatchService(builder, workspace, config, cwd, logger, locks),
+            researcher: new ResearchAgent(searcher, new AiSdkChatModel(config), config),
+            projectId: () => locator.resolveProject(cwd).id,
+        };
+        return new Lmgrep(services);
+    }
+    /**
+     * Opens another project's database read-only for cross-project search.
+     *
+     * Foreign indexes get their own connection and are closed after the query,
+     * so a `--across` over many projects does not leave native runtimes open.
+     */
+    foreignOpener(locator) {
+        return {
+            async open(projectPath) {
+                const project = locator.resolveProject(projectPath);
+                const tables = new LanceTables(locator.databasePathFor(projectPath), project.branch);
+                const manifest = new FileManifestRepository(tables, project.branch);
+                return {
+                    chunks: new ChunkRepository(tables, manifest, project.branch),
+                    close: async () => tables.close(),
+                };
+            },
+        };
+    }
+}
+//# sourceMappingURL=LmgrepFactory.js.map

package/dist/application/LmgrepFactory.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"LmgrepFactory.js","sourceRoot":"","sources":["../../src/application/LmgrepFactory.ts"],"names":[],"mappings":"AAIA,OAAO,EAAE,aAAa,EAAE,MAAM,oCAAoC,CAAC;AACnE,OAAO,EAAE,cAAc,EAAE,MAAM,qCAAqC,CAAC;AACrE,OAAO,EAAE,cAAc,EAAE,MAAM,wCAAwC,CAAC;AACxE,OAAO,EAAE,aAAa,EAAE,MAAM,uCAAuC,CAAC;AACtE,OAAO,EAAE,kBAAkB,EAAE,MAAM,4CAA4C,CAAC;AAChF,OAAO,EAAE,YAAY,EAAE,MAAM,sCAAsC,CAAC;AACpE,OAAO,EAAE,aAAa,EAAE,MAAM,uCAAuC,CAAC;AACtE,OAAO,EAAE,aAAa,EAAE,MAAM,iCAAiC,CAAC;AAChE,OAAO,EAAE,cAAc,EAAE,MAAM,wCAAwC,CAAC;AACxE,OAAO,EAAE,oBAAoB,EAAE,MAAM,8CAA8C,CAAC;AACpF,OAAO,EAAE,eAAe,EAAE,MAAM,yCAAyC,CAAC;AAC1E,OAAO,EAAE,cAAc,EAAE,MAAM,wCAAwC,CAAC;AACxE,OAAO,EAAE,SAAS,EAAE,MAAM,mCAAmC,CAAC;AAC9D,OAAO,EAAE,SAAS,EAAE,MAAM,oCAAoC,CAAC;AAC/D,OAAO,EAAE,eAAe,EAAE,MAAM,8CAA8C,CAAC;AAC/E,OAAO,EAAE,sBAAsB,EAAE,MAAM,qDAAqD,CAAC;AAC7F,OAAO,EAAE,gBAAgB,EAAE,MAAM,+CAA+C,CAAC;AACjF,OAAO,EAAE,WAAW,EAAE,MAAM,0CAA0C,CAAC;AACvE,OAAO,EAAE,iBAAiB,EAAE,MAAM,mDAAmD,CAAC;AACtF,OAAO,EAAE,kBAAkB,EAAE,MAAM,kCAAkC,CAAC;AACtE,OAAO,EAAE,qBAAqB,EAAE,MAAM,qCAAqC,CAAC;AAC5E,OAAO,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAC1D,OAAO,EAAE,MAAM,EAAuB,MAAM,aAAa,CAAC;AAC1D,OAAO,EAAE,iBAAiB,EAAE,MAAM,mCAAmC,CAAC;AACtE,OAAO,EAAE,aAAa,EAAE,MAAM,+BAA+B,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,8BAA8B,CAAC;AAC5D,OAAO,EAAE,aAAa,EAAE,MAAM,6BAA6B,CAAC;AAC5D,OAAO,EAAE,aAAa,EAAE,MAAM,2BAA2B,CAAC;AAC1D,OAAO,EAEN,oBAAoB,GACpB,MAAM,kCAAkC,CAAC;AAqB1C;;;;;;GAMG;AACH,MAAM,OAAO,aAAa;IACzB,KAAK,CAAC,IAAI,CAAC,OAAsB;QAChC,MAAM,EAAE,GAAG,EAAE,GAAG,OAAO,CAAC;QACxB,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;QAErD,MAAM,MAAM,GAAG,IAAI,YAAY,EAAE,CAAC;QAClC,MAAM,MAAM,GAAiB;YAC5B,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC;YACnB,GAAG,OAAO,CAAC,MAAM;SACjB,CAAC;QACF,KAAK,MAAM,OAAO,IAAI,MAAM,CAAC,QAAQ;YAAE,OAAO,CAAC,SAAS,EAAE,CAAC,OAAO,CAAC,CAAC;QAEpE,MAAM,KAAK,GAAG,IAAI,cAAc,EAAE,CAAC;QACnC,MAAM,GAAG,GAAG,IAAI,SAAS,EAAE,CAAC;QAC5B,qEAAqE;QACrE,8DAA8D;QAC9D,MAAM,OAAO,GAAG,IAAI,cAAc,CACjC,GAAG,EACH,KAAK,EACL,aAAa,CAAC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,EAC9B,MAAM,CAAC,UAAU,CACjB,CAAC;QACF,MAAM,QAAQ,GAAG,OAAO,CAAC,eAAe,CAAC,GAAG,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC;QAChE,MAAM,QAAQ,GAAG,IAAI,oBAAoB,EAAE,CAAC;QAC5C,MAAM,QAAQ,GAAG,IAAI,eAAe,CAAC,KAAK,CAAC,CAAC;QAC5C,MAAM,YAAY,GAAG,IAAI,iBAAiB,CACzC,IAAI,cAAc,CAAC,QAAQ,CAAC,EAC5B,OAAO,EACP,QAAQ,EACR,MAAM,EACN,GAAG,CACH,CAAC;QACF,MAAM,KAAK,GAAG,IAAI,aAAa,CAC9B,KAAK,CAAC,cAAc,EAAE,EACtB,QAAQ,CAAC,IAAI,EACb,QAAQ,CAAC,IAAI,CACb,CAAC;QAEF,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;QAC/D,MAAM,QAAQ,GAAG,IAAI,sBAAsB,CAAC,MAAM,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;QACrE,MAAM,MAAM,GAAG,IAAI,eAAe,CAAC,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;QACtE,MAAM,WAAW,GAAG,IAAI,gBAAgB,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;QAE3D,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,IAAI,aAAa,CAAC,MAAM,CAAC,CAAC;QAC/D,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,IAAI,iBAAiB,EAAE,CAAC;QAC3D,MAAM,SAAS,GAAG,IAAI,SAAS,EAAE,CAAC;QAClC,MAAM,QAAQ,GAAG,IAAI,kBAAkB,CAAC,MAAM,CAAC,CAAC;QAEhD,MAAM,OAAO,GAAG,IAAI,qBAAqB,CAAC,QAAQ,EAAE,GAAG,EAAE,MAAM,CAAC,CAAC;QAEjE,MAAM,OAAO,GAAG,IAAI,YAAY,CAAC;YAChC,SAAS;YACT,OAAO;YACP,QAAQ;YACR,MAAM;YACN,QAAQ;YACR,WAAW;YACX,YAAY,EAAE,IAAI,kBAAkB,CAAC,QAAQ,EAAE,GAAG,EAAE,MAAM,CAAC;YAC3D,OAAO;YACP,MAAM;YACN,MAAM;YACN,QAAQ;YACR,KAAK;YACL,cAAc,EAAE,CAAC,UAAU,EAAE,EAAE;gBAC9B,MAAM,OAAO,GAAG,OAAO,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;gBAC5C,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE;oBAC7B,IAAI,EAAE,OAAO,CAAC,IAAI;oBAClB,MAAM,EAAE,OAAO,CAAC,MAAM;oBACtB,MAAM,EAAE,QAAQ,CAAC,MAAM,CAAC,QAAQ,EAAE;oBAClC,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,UAAU;iBACV,CAAC,CAAC;YACJ,CAAC;YACD,iEAAiE;YACjE,gEAAgE;YAChE,8DAA8D;YAC9D,8DAA8D;YAC9D,qBAAqB;YACrB,aAAa,EAAE,GAAG,EAAE;gBACnB,MAAM,OAAO,GAAG,OAAO,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;gBAC5C,QAAQ,CAAC,MAAM,CAAC;oBACf,IAAI,EAAE,OAAO,CAAC,cAAc,CAAC,GAAG,CAAC;oBACjC,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;oBACpD,MAAM,EAAE,OAAO,CAAC,MAAM;oBACtB,YAAY,EAAE,QAAQ,CAAC,IAAI;oBAC3B,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,UAAU,EAAE,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,UAAU;iBACpD,CAAC,CAAC;YACJ,CAAC;YACD,WAAW,EAAE,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,EAAE;YACpC,eAAe,EAAE,QAAQ,CAAC,OAAO;SACjC,CAAC,CAAC;QAEH,MAAM,QAAQ,GAAG,IAAI,aAAa,CACjC,QAAQ,EACR,IAAI,oBAAoB,CACvB,GAAG,EACH,MAAM,EACN,QAAQ,EACR,OAAO,EACP,IAAI,CAAC,aAAa,CAAC,OAAO,CAAC,CAC3B,EACD,MAAM,EACN,MAAM,EACN,GAAG,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAClC,YAAY,CACZ,CAAC;QAEF,MAAM,QAAQ,GAAmB;YAChC,GAAG;YACH,MAAM;YACN,QAAQ;YACR,OAAO;YACP,MAAM;YACN,MAAM;YACN,MAAM;YACN,QAAQ;YACR,WAAW;YACX,QAAQ;YACR,QAAQ;YACR,YAAY;YACZ,QAAQ;YACR,OAAO;YACP,OAAO;YACP,OAAO;YACP,QAAQ;YACR,cAAc,EAAE,IAAI,aAAa,CAChC,MAAM,EACN,WAAW,EACX,QAAQ,EACR,OAAO,EACP,QAAQ,EACR,QAAQ,EACR,MAAM,EACN,GAAG,EACH,MAAM,CAAC,OAAO,EACd,YAAY,CACZ;YACD,OAAO,EAAE,IAAI,YAAY,CAAC,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,MAAM,EAAE,KAAK,CAAC;YACzE,UAAU,EAAE,IAAI,aAAa,CAC5B,QAAQ,EACR,IAAI,cAAc,CAAC,MAAM,CAAC,EAC1B,MAAM,CACN;YACD,SAAS,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,EAAE;SAC/C,CAAC;QAEF,OAAO,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC7B,CAAC;IAED;;;;;OAKG;IACK,aAAa,CAAC,OAAuB;QAC5C,OAAO;YACN,KAAK,CAAC,IAAI,CAAC,WAAmB;gBAC7B,MAAM,OAAO,GAAG,OAAO,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC;gBACpD,MAAM,MAAM,GAAG,IAAI,WAAW,CAC7B,OAAO,CAAC,eAAe,CAAC,WAAW,CAAC,EACpC,OAAO,CAAC,MAAM,CACd,CAAC;gBACF,MAAM,QAAQ,GAAG,IAAI,sBAAsB,CAAC,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;gBACpE,OAAO;oBACN,MAAM,EAAE,IAAI,eAAe,CAAC,MAAM,EAAE,QAAQ,EAAE,OAAO,CAAC,MAAM,CAAC;oBAC7D,KAAK,EAAE,KAAK,IAAI,EAAE,CAAC,MAAM,CAAC,KAAK,EAAE;iBACjC,CAAC;YACH,CAAC;SACD,CAAC;IACH,CAAC;CACD"}

package/dist/application/indexing/BranchBootstrapper.d.ts

@@ -0,0 +1,21 @@
+import type { FileManifestRepositoryPort } from "../../domain/ports/FileManifestRepositoryPort.js";
+import type { GitPort } from "../../domain/ports/GitPort.js";
+import type { LoggerPort } from "../../domain/ports/LoggerPort.js";
+/**
+ * Seeds a new branch's manifest from its closest relative.
+ *
+ * Checking out a branch would otherwise present as "nothing indexed" and
+ * trigger a full re-embed of the whole tree, even though almost every file is
+ * byte-identical to the branch it came from. Copying the nearest branch's
+ * manifest turns that into a diff of the handful of files that actually differ.
+ */
+export declare class BranchBootstrapper {
+    private readonly manifest;
+    private readonly git;
+    private readonly logger;
+    constructor(manifest: FileManifestRepositoryPort, git: GitPort, logger: LoggerPort);
+    /** No-op unless this branch's manifest is empty. */
+    bootstrap(repoRoot: string): Promise<void>;
+    /** The stored branch fewest commits behind HEAD. */
+    private closestBranch;
+}

package/dist/application/indexing/BranchBootstrapper.js

@@ -0,0 +1,52 @@
+/**
+ * Seeds a new branch's manifest from its closest relative.
+ *
+ * Checking out a branch would otherwise present as "nothing indexed" and
+ * trigger a full re-embed of the whole tree, even though almost every file is
+ * byte-identical to the branch it came from. Copying the nearest branch's
+ * manifest turns that into a diff of the handful of files that actually differ.
+ */
+export class BranchBootstrapper {
+    manifest;
+    git;
+    logger;
+    constructor(manifest, git, logger) {
+        this.manifest = manifest;
+        this.git = git;
+        this.logger = logger;
+    }
+    /** No-op unless this branch's manifest is empty. */
+    async bootstrap(repoRoot) {
+        if (!(await this.manifest.current()).isEmpty)
+            return;
+        const stored = await this.manifest.storedBranches();
+        if (stored.length === 0)
+            return;
+        const source = this.closestBranch(repoRoot, stored);
+        if (!source)
+            return;
+        const copied = await this.manifest.copyFromBranch(source);
+        if (copied > 0) {
+            this.logger.info(`Bootstrapped from "${source}" manifest (${copied} files). Diffing changes...`);
+        }
+    }
+    /** The stored branch fewest commits behind HEAD. */
+    closestBranch(repoRoot, candidates) {
+        let best;
+        let bestDistance = Number.POSITIVE_INFINITY;
+        for (const candidate of candidates) {
+            const mergeBase = this.git.mergeBase(repoRoot, candidate);
+            if (!mergeBase)
+                continue;
+            const distance = this.git.commitDistance(repoRoot, mergeBase, "HEAD");
+            if (distance === undefined)
+                continue;
+            if (distance < bestDistance) {
+                bestDistance = distance;
+                best = candidate;
+            }
+        }
+        return best;
+    }
+}
+//# sourceMappingURL=BranchBootstrapper.js.map

package/dist/application/indexing/BranchBootstrapper.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"BranchBootstrapper.js","sourceRoot":"","sources":["../../../src/application/indexing/BranchBootstrapper.ts"],"names":[],"mappings":"AAIA;;;;;;;GAOG;AACH,MAAM,OAAO,kBAAkB;IAEZ;IACA;IACA;IAHlB,YACkB,QAAoC,EACpC,GAAY,EACZ,MAAkB;QAFlB,aAAQ,GAAR,QAAQ,CAA4B;QACpC,QAAG,GAAH,GAAG,CAAS;QACZ,WAAM,GAAN,MAAM,CAAY;IACjC,CAAC;IAEJ,oDAAoD;IACpD,KAAK,CAAC,SAAS,CAAC,QAAgB;QAC/B,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC,CAAC,OAAO;YAAE,OAAO;QAErD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,cAAc,EAAE,CAAC;QACpD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAEhC,MAAM,MAAM,GAAG,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;QACpD,IAAI,CAAC,MAAM;YAAE,OAAO;QAEpB,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC;QAC1D,IAAI,MAAM,GAAG,CAAC,EAAE,CAAC;YAChB,IAAI,CAAC,MAAM,CAAC,IAAI,CACf,sBAAsB,MAAM,eAAe,MAAM,6BAA6B,CAC9E,CAAC;QACH,CAAC;IACF,CAAC;IAED,oDAAoD;IAC5C,aAAa,CACpB,QAAgB,EAChB,UAAoB;QAEpB,IAAI,IAAwB,CAAC;QAC7B,IAAI,YAAY,GAAG,MAAM,CAAC,iBAAiB,CAAC;QAE5C,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;YACpC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;YAC1D,IAAI,CAAC,SAAS;gBAAE,SAAS;YACzB,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,CAAC,QAAQ,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC;YACtE,IAAI,QAAQ,KAAK,SAAS;gBAAE,SAAS;YACrC,IAAI,QAAQ,GAAG,YAAY,EAAE,CAAC;gBAC7B,YAAY,GAAG,QAAQ,CAAC;gBACxB,IAAI,GAAG,SAAS,CAAC;YAClB,CAAC;QACF,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC;CACD"}

package/dist/application/indexing/BranchManifestSweeper.d.ts

@@ -0,0 +1,17 @@
+import type { FileManifestRepositoryPort } from "../../domain/ports/FileManifestRepositoryPort.js";
+import type { GitPort } from "../../domain/ports/GitPort.js";
+import type { LoggerPort } from "../../domain/ports/LoggerPort.js";
+/**
+ * Drops manifests for branches git no longer has.
+ *
+ * Without this, every deleted feature branch leaves its file manifest behind
+ * forever, and those rows keep chunks alive that nothing can reach — the index
+ * grows monotonically across a project's lifetime.
+ */
+export declare class BranchManifestSweeper {
+    private readonly manifest;
+    private readonly git;
+    private readonly logger;
+    constructor(manifest: FileManifestRepositoryPort, git: GitPort, logger: LoggerPort);
+    sweep(repoRoot: string): Promise<void>;
+}

package/dist/application/indexing/BranchManifestSweeper.js

@@ -0,0 +1,35 @@
+import { Branch } from "../../domain/project/Branch.js";
+/**
+ * Drops manifests for branches git no longer has.
+ *
+ * Without this, every deleted feature branch leaves its file manifest behind
+ * forever, and those rows keep chunks alive that nothing can reach — the index
+ * grows monotonically across a project's lifetime.
+ */
+export class BranchManifestSweeper {
+    manifest;
+    git;
+    logger;
+    constructor(manifest, git, logger) {
+        this.manifest = manifest;
+        this.git = git;
+        this.logger = logger;
+    }
+    async sweep(repoRoot) {
+        const branches = this.git.localBranches(repoRoot);
+        // No git, or an unreadable repo: nothing can be proven stale, so leave
+        // every manifest alone rather than guessing.
+        if (branches.length === 0)
+            return;
+        const live = new Set(branches);
+        // Non-git projects index under the default branch, which git never lists.
+        live.add(Branch.DEFAULT_NAME);
+        for (const stored of await this.manifest.storedBranches()) {
+            if (live.has(stored))
+                continue;
+            await this.manifest.deleteBranch(stored);
+            this.logger.info(`Swept stale manifest for deleted branch "${stored}"`);
+        }
+    }
+}
+//# sourceMappingURL=BranchManifestSweeper.js.map

package/dist/application/indexing/BranchManifestSweeper.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"BranchManifestSweeper.js","sourceRoot":"","sources":["../../../src/application/indexing/BranchManifestSweeper.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,MAAM,EAAE,MAAM,gCAAgC,CAAC;AAExD;;;;;;GAMG;AACH,MAAM,OAAO,qBAAqB;IAEf;IACA;IACA;IAHlB,YACkB,QAAoC,EACpC,GAAY,EACZ,MAAkB;QAFlB,aAAQ,GAAR,QAAQ,CAA4B;QACpC,QAAG,GAAH,GAAG,CAAS;QACZ,WAAM,GAAN,MAAM,CAAY;IACjC,CAAC;IAEJ,KAAK,CAAC,KAAK,CAAC,QAAgB;QAC3B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;QAClD,uEAAuE;QACvE,6CAA6C;QAC7C,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAElC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC/B,0EAA0E;QAC1E,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,YAAY,CAAC,CAAC;QAE9B,KAAK,MAAM,MAAM,IAAI,MAAM,IAAI,CAAC,QAAQ,CAAC,cAAc,EAAE,EAAE,CAAC;YAC3D,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,SAAS;YAC/B,MAAM,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YACzC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,4CAA4C,MAAM,GAAG,CAAC,CAAC;QACzE,CAAC;IACF,CAAC;CACD"}

package/dist/application/indexing/Duration.d.ts

@@ -0,0 +1,9 @@
+/** A human-written duration like "10m" or "2h", as milliseconds. */
+export declare class Duration {
+    readonly milliseconds: number;
+    private static readonly UNIT_MS;
+    private constructor();
+    static parse(value: string): Duration;
+    /** The wall-clock instant this duration ago. */
+    agoFrom(now: number): number;
+}

package/dist/application/indexing/Duration.js

@@ -0,0 +1,25 @@
+/** A human-written duration like "10m" or "2h", as milliseconds. */
+export class Duration {
+    milliseconds;
+    static UNIT_MS = {
+        s: 1000,
+        m: 60_000,
+        h: 3_600_000,
+        d: 86_400_000,
+    };
+    constructor(milliseconds) {
+        this.milliseconds = milliseconds;
+    }
+    static parse(value) {
+        const match = value.match(/^(\d+)\s*(s|m|h|d)$/);
+        if (!match) {
+            throw new Error(`Invalid duration "${value}". Use e.g. 10m, 2h, 1d`);
+        }
+        return new Duration(Number.parseInt(match[1], 10) * Duration.UNIT_MS[match[2]]);
+    }
+    /** The wall-clock instant this duration ago. */
+    agoFrom(now) {
+        return now - this.milliseconds;
+    }
+}
+//# sourceMappingURL=Duration.js.map

package/dist/application/indexing/Duration.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"Duration.js","sourceRoot":"","sources":["../../../src/application/indexing/Duration.ts"],"names":[],"mappings":"AAAA,oEAAoE;AACpE,MAAM,OAAO,QAAQ;IAQS;IAPrB,MAAM,CAAU,OAAO,GAA2B;QACzD,CAAC,EAAE,IAAI;QACP,CAAC,EAAE,MAAM;QACT,CAAC,EAAE,SAAS;QACZ,CAAC,EAAE,UAAU;KACb,CAAC;IAEF,YAA6B,YAAoB;QAApB,iBAAY,GAAZ,YAAY,CAAQ;IAAG,CAAC;IAErD,MAAM,CAAC,KAAK,CAAC,KAAa;QACzB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAC;QACjD,IAAI,CAAC,KAAK,EAAE,CAAC;YACZ,MAAM,IAAI,KAAK,CAAC,qBAAqB,KAAK,yBAAyB,CAAC,CAAC;QACtE,CAAC;QACD,OAAO,IAAI,QAAQ,CAClB,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAC1D,CAAC;IACH,CAAC;IAED,gDAAgD;IAChD,OAAO,CAAC,GAAW;QAClB,OAAO,GAAG,GAAG,IAAI,CAAC,YAAY,CAAC;IAChC,CAAC"}

package/dist/application/indexing/EmbeddingAbortError.d.ts

@@ -0,0 +1,11 @@
+/**
+ * Raised when too many consecutive batches fail — the provider is down rather
+ * than the input being bad. Successes are already persisted by the time this
+ * throws, so the caller's job is to report and exit, not to roll back.
+ */
+export declare class EmbeddingAbortError extends Error {
+    readonly succeeded: number;
+    readonly failedIndices: Set<number>;
+    readonly total: number;
+    constructor(succeeded: number, failedIndices: Set<number>, total: number);
+}

package/dist/application/indexing/EmbeddingAbortError.js

@@ -0,0 +1,20 @@
+/**
+ * Raised when too many consecutive batches fail — the provider is down rather
+ * than the input being bad. Successes are already persisted by the time this
+ * throws, so the caller's job is to report and exit, not to roll back.
+ */
+export class EmbeddingAbortError extends Error {
+    succeeded;
+    failedIndices;
+    total;
+    constructor(succeeded, failedIndices, total) {
+        super(`Embedding aborted after ${failedIndices.size} failures. ` +
+            `${succeeded}/${total} chunks embedded successfully. ` +
+            "Check your embedding provider and run `lmgrep index` to resume.");
+        this.succeeded = succeeded;
+        this.failedIndices = failedIndices;
+        this.total = total;
+        this.name = "EmbeddingAbortError";
+    }
+}
+//# sourceMappingURL=EmbeddingAbortError.js.map

package/dist/application/indexing/EmbeddingAbortError.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"EmbeddingAbortError.js","sourceRoot":"","sources":["../../../src/application/indexing/EmbeddingAbortError.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH,MAAM,OAAO,mBAAoB,SAAQ,KAAK;IAEnC;IACA;IACA;IAHV,YACU,SAAiB,EACjB,aAA0B,EAC1B,KAAa;QAEtB,KAAK,CACJ,2BAA2B,aAAa,CAAC,IAAI,aAAa;YACzD,GAAG,SAAS,IAAI,KAAK,iCAAiC;YACtD,iEAAiE,CAClE,CAAC;QARO,cAAS,GAAT,SAAS,CAAQ;QACjB,kBAAa,GAAb,aAAa,CAAa;QAC1B,UAAK,GAAL,KAAK,CAAQ;QAOtB,IAAI,CAAC,IAAI,GAAG,qBAAqB,CAAC;IACnC,CAAC;CACD"}

package/dist/application/indexing/EmbeddingPipeline.d.ts

@@ -0,0 +1,59 @@
+import type { LmgrepConfig } from "../../domain/config/LmgrepConfig.js";
+import type { Vector } from "../../domain/corpus/Vector.js";
+import type { EmbedderPort } from "../../domain/ports/EmbedderPort.js";
+/** One successfully embedded item, keyed by its position in the input. */
+export interface EmbeddedItem {
+    index: number;
+    vector: Vector;
+}
+export interface EmbeddingProgress {
+    onBatchStart?(batchNumber: number, totalBatches: number): void;
+    onBatchDone?(batchNumber: number, succeeded: number, failed: number): void;
+    onReload?(attempt: number, max: number): void;
+}
+/**
+ * Embeds a list of texts batch by batch, surviving a flaky provider.
+ *
+ * Three behaviours matter here and are easy to lose:
+ *
+ *  - **Incremental persistence.** Each batch is handed to `persist` before the
+ *    next starts, so a crash leaves completed work in the index and a resume
+ *    skips it. Nothing is buffered for the caller to write at the end.
+ *  - **Per-item retry.** A failed batch is retried one item at a time, so a
+ *    single oversized chunk cannot cost the other 99 in its batch.
+ *  - **Model reload.** A local provider that has fallen over is given a chance
+ *    to reload, and the failed span is replayed rather than abandoned.
+ */
+export declare class EmbeddingPipeline {
+    private readonly embedder;
+    private readonly config;
+    private readonly progress;
+    /** Reloads a local model; returns false when it cannot be done. */
+    private readonly reloadModel;
+    /** Whether the provider is local, and therefore reloadable. */
+    private readonly isLocal;
+    private static readonly BATCH_DELAY_MS;
+    private static readonly MAX_CONSECUTIVE_FAILURES;
+    private static readonly MAX_RELOADS;
+    private consecutiveFailures;
+    private reloads;
+    constructor(embedder: EmbedderPort, config: LmgrepConfig, progress?: EmbeddingProgress, 
+    /** Reloads a local model; returns false when it cannot be done. */
+    reloadModel?: () => Promise<boolean>, 
+    /** Whether the provider is local, and therefore reloadable. */
+    isLocal?: boolean);
+    /**
+     * Returns the positions that could not be embedded. Successes are reported
+     * through `persist` exactly once each, never returned in bulk.
+     */
+    run(texts: string[], persist: (items: EmbeddedItem[]) => Promise<void>): Promise<{
+        failedIndices: Set<number>;
+    }>;
+    /** Retry a failed batch item by item; returns how many still failed. */
+    private retryIndividually;
+    /**
+     * Give a local model a chance to come back, and return the loop position to
+     * rewind to. Undefined means give up.
+     */
+    private tryReload;
+}