bluera-knowledge 0.17.2 → 0.18.1

package/dist/index.js

@@ -3,34 +3,33 @@ import {
   ZilAdapter,
   runMCPServer,
   spawnBackgroundWorker
-} from "./chunk-YMDXPECI.js";
+} from "./chunk-EZXJ3W5X.js";
 import {
   IntelligentCrawler,
   getCrawlStrategy
-} from "./chunk-PFHK5Q22.js";
+} from "./chunk-VUGQ7HAR.js";
 import {
   ASTParser,
   AdapterRegistry,
   ChunkingService,
   JobService,
+  ProjectRootService,
   StoreDefinitionService,
   classifyWebContentType,
-  createDocumentId,
   createServices,
-  createStoreId,
   destroyServices,
   err,
   extractRepoName,
   isFileStoreDefinition,
+  isGitUrl,
   isRepoStoreDefinition,
   isWebStoreDefinition,
   ok
-} from "./chunk-WMALVLFW.js";
-import "./chunk-HRQD3MPH.js";
-
-// src/index.ts
-import { homedir as homedir2 } from "os";
-import { join as join5 } from "path";
+} from "./chunk-RDDGZIDL.js";
+import {
+  createDocumentId
+} from "./chunk-CLIMKLTW.js";
+import "./chunk-HXBIIMYL.js";
 
 // src/cli/commands/crawl.ts
 import { createHash } from "crypto";
@@ -61,7 +60,11 @@ function createCrawlCommand(getOptions) {
           );
         }
       }
-      const services = await createServices(globalOpts.config, globalOpts.dataDir);
+      const services = await createServices(
+        globalOpts.config,
+        globalOpts.dataDir,
+        globalOpts.projectRoot
+      );
       let store;
       let storeCreated = false;
       const existingStore = await services.store.getByIdOrName(storeIdOrName);
@@ -92,7 +95,12 @@ function createCrawlCommand(getOptions) {
       } else {
         store = existingStore;
       }
-      const maxPages = cmdOptions.maxPages !== void 0 ? parseInt(cmdOptions.maxPages, 10) : 50;
+      const maxPages = parseInt(cmdOptions.maxPages, 10);
+      if (Number.isNaN(maxPages)) {
+        throw new Error(
+          `Invalid value for --max-pages: "${cmdOptions.maxPages}" is not a valid integer`
+        );
+      }
       const isInteractive = process.stdout.isTTY && globalOpts.quiet !== true && globalOpts.format !== "json";
       let spinner;
       if (isInteractive) {
@@ -101,7 +109,8 @@ function createCrawlCommand(getOptions) {
       } else if (globalOpts.quiet !== true && globalOpts.format !== "json") {
         console.log(`Crawling ${url}`);
       }
-      const crawler = new IntelligentCrawler();
+      const appConfig = await services.config.load();
+      const crawler = new IntelligentCrawler(appConfig.crawl);
       const webChunker = ChunkingService.forContentType("web");
       let pagesIndexed = 0;
       let chunksCreated = 0;
@@ -122,6 +131,7 @@ function createCrawlCommand(getOptions) {
         }
       });
       try {
+        services.lance.setDimensions(await services.embeddings.ensureDimensions());
         await services.lance.initialize(store.id);
         const docs = [];
         for await (const result of crawler.crawl(url, {
@@ -150,7 +160,7 @@ function createCrawlCommand(getOptions) {
                 title: result.title,
                 extracted: result.extracted !== void 0,
                 depth: result.depth,
-                indexedAt: /* @__PURE__ */ new Date(),
+                indexedAt: (/* @__PURE__ */ new Date()).toISOString(),
                 fileType,
                 chunkIndex: chunk.chunkIndex,
                 totalChunks: chunk.totalChunks,
@@ -165,6 +175,7 @@ function createCrawlCommand(getOptions) {
           if (spinner) {
             spinner.text = "Indexing documents...";
           }
+          await services.lance.clearAllDocuments(store.id);
           await services.lance.addDocuments(store.id, docs);
           await services.lance.createFtsIndex(store.id);
         }
@@ -203,7 +214,7 @@ function createCrawlCommand(getOptions) {
         await destroyServices(services);
       }
       if (exitCode !== 0) {
-        process.exit(exitCode);
+        process.exitCode = exitCode;
       }
     }
   );
@@ -213,9 +224,13 @@ function createCrawlCommand(getOptions) {
 import { Command as Command2 } from "commander";
 import ora2 from "ora";
 function createIndexCommand(getOptions) {
-  const index = new Command2("index").description("Scan store files, chunk text, generate embeddings, save to LanceDB").argument("<store>", "Store ID or name").option("--force", "Re-index all files even if unchanged").action(async (storeIdOrName, _options) => {
+  const index = new Command2("index").description("Scan store files, chunk text, generate embeddings, save to LanceDB").argument("<store>", "Store ID or name").option("--force", "Re-index all files even if unchanged").action(async (storeIdOrName, options) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
     let exitCode = 0;
     try {
       indexLogic: {
@@ -232,19 +247,21 @@ function createIndexCommand(getOptions) {
         } else if (globalOpts.quiet !== true && globalOpts.format !== "json") {
           console.log(`Indexing store: ${store.name}`);
         }
+        services.lance.setDimensions(await services.embeddings.ensureDimensions());
         await services.lance.initialize(store.id);
-        const result = await services.index.indexStore(store, (event) => {
+        const progressCallback = (event) => {
           if (event.type === "progress") {
             if (spinner) {
               spinner.text = `Indexing: ${String(event.current)}/${String(event.total)} files - ${event.message}`;
             }
           }
-        });
+        };
+        const result = options.force === true ? await services.index.indexStore(store, progressCallback) : await services.index.indexStoreIncremental(store, progressCallback);
         if (result.success) {
           if (globalOpts.format === "json") {
             console.log(JSON.stringify(result.data, null, 2));
           } else {
-            const message = `Indexed ${String(result.data.documentsIndexed)} documents, ${String(result.data.chunksCreated)} chunks in ${String(result.data.timeMs)}ms`;
+            const message = `Indexed ${String(result.data.filesIndexed)} documents, ${String(result.data.chunksCreated)} chunks in ${String(result.data.timeMs)}ms`;
             if (spinner !== void 0) {
               spinner.succeed(message);
             } else if (globalOpts.quiet !== true) {
@@ -266,7 +283,7 @@ function createIndexCommand(getOptions) {
       await destroyServices(services);
     }
     if (exitCode !== 0) {
-      process.exit(exitCode);
+      process.exitCode = exitCode;
     }
   });
   index.command("watch <store>").description("Watch store directory; re-index when files change").option(
@@ -275,20 +292,35 @@ function createIndexCommand(getOptions) {
     "1000"
   ).action(async (storeIdOrName, options) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
     const store = await services.store.getByIdOrName(storeIdOrName);
     if (store === void 0 || store.type !== "file" && store.type !== "repo") {
       console.error(`Error: File/repo store not found: ${storeIdOrName}`);
-      process.exit(3);
+      await destroyServices(services);
+      process.exitCode = 3;
+      return;
+    }
+    const appConfig = await services.config.load();
+    const { WatchService } = await import("./watch.service-VDSUQ72Z.js");
+    const watchService = new WatchService(services.index, services.lance, services.embeddings, {
+      ignorePatterns: appConfig.indexing.ignorePatterns
+    });
+    const debounceMs = parseInt(options.debounce, 10);
+    if (Number.isNaN(debounceMs)) {
+      throw new Error(
+        `Invalid value for --debounce: "${options.debounce}" is not a valid integer`
+      );
     }
-    const { WatchService } = await import("./watch.service-OPLKIDFQ.js");
-    const watchService = new WatchService(services.index, services.lance);
     if (globalOpts.quiet !== true) {
       console.log(`Watching ${store.name} for changes...`);
     }
     await watchService.watch(
       store,
-      parseInt(options.debounce ?? "1000", 10),
+      debounceMs,
       () => {
         if (globalOpts.quiet !== true) {
           console.log(`Re-indexed ${store.name}`);
@@ -301,8 +333,11 @@ function createIndexCommand(getOptions) {
     process.on("SIGINT", () => {
       void (async () => {
         await watchService.unwatchAll();
+        await destroyServices(services);
         process.exit(0);
-      })().catch(() => {
+      })().catch((err2) => {
+        console.error("Shutdown error:", err2);
+        process.exit(1);
       });
     });
   });
@@ -316,7 +351,8 @@ function createMCPCommand(getOptions) {
     const opts = getOptions();
     await runMCPServer({
       dataDir: opts.dataDir,
-      config: opts.config
+      config: opts.config,
+      projectRoot: opts.projectRoot
     });
   });
   return mcp;
@@ -509,6 +545,7 @@ var DependencyUsageAnalyzer = class {
           if (![
             "node_modules",
             ".git",
+            ".bluera",
             "dist",
             "build",
             "coverage",
@@ -815,16 +852,19 @@ async function handleAddRepo(args, options = {}) {
     });
     if (!result.success) {
       console.error(`Error: ${result.error.message}`);
-      process.exit(1);
+      process.exitCode = 1;
+      return;
     }
     console.log(`Created store: ${storeName} (${result.data.id})`);
     if ("path" in result.data) {
       console.log(`Location: ${result.data.path}`);
     }
     console.log("\nIndexing...");
+    services.lance.setDimensions(await services.embeddings.ensureDimensions());
+    await services.lance.initialize(result.data.id);
     const indexResult = await services.index.indexStore(result.data);
     if (indexResult.success) {
-      console.log(`Indexed ${String(indexResult.data.documentsIndexed)} files`);
+      console.log(`Indexed ${String(indexResult.data.filesIndexed)} files`);
     } else {
       console.error(`Indexing failed: ${indexResult.error.message}`);
     }
@@ -849,16 +889,19 @@ async function handleAddFolder(args, options = {}) {
     });
     if (!result.success) {
       console.error(`Error: ${result.error.message}`);
-      process.exit(1);
+      process.exitCode = 1;
+      return;
     }
     console.log(`Created store: ${storeName} (${result.data.id})`);
     if ("path" in result.data) {
       console.log(`Location: ${result.data.path}`);
     }
     console.log("\nIndexing...");
+    services.lance.setDimensions(await services.embeddings.ensureDimensions());
+    await services.lance.initialize(result.data.id);
     const indexResult = await services.index.indexStore(result.data);
     if (indexResult.success) {
-      console.log(`Indexed ${String(indexResult.data.documentsIndexed)} files`);
+      console.log(`Indexed ${String(indexResult.data.filesIndexed)} files`);
     } else {
       console.error(`Indexing failed: ${indexResult.error.message}`);
     }
@@ -913,7 +956,8 @@ async function handleSuggest(options = {}) {
     spinner.stop();
     if (!result.success) {
       console.error(`Error: ${result.error.message}`);
-      process.exit(1);
+      process.exitCode = 1;
+      return;
     }
     const { usages, totalFilesScanned, skippedFiles } = result.data;
     console.log(
@@ -1031,14 +1075,18 @@ function createSearchCommand(getOptions) {
   ).option("-n, --limit <count>", "Maximum results to return (default: 10)", "10").option("-t, --threshold <score>", "Minimum score 0-1; omit low-relevance results").option(
     "--min-relevance <score>",
     "Minimum raw cosine similarity 0-1; returns empty if no results meet threshold"
-  ).option("--include-content", "Show full document content, not just preview snippet").option(
+  ).option(
     "--detail <level>",
     "Context detail: minimal, contextual, full (default: minimal)",
     "minimal"
   ).action(
     async (query, options) => {
       const globalOpts = getOptions();
-      const services = await createServices(globalOpts.config, globalOpts.dataDir);
+      const services = await createServices(
+        globalOpts.config,
+        globalOpts.dataDir,
+        globalOpts.projectRoot
+      );
       let exitCode = 0;
       try {
         let storeIds = (await services.store.list()).map((s) => s.id);
@@ -1063,18 +1111,48 @@ function createSearchCommand(getOptions) {
             exitCode = 1;
             break searchLogic;
           }
+          services.lance.setDimensions(await services.embeddings.ensureDimensions());
           for (const storeId of storeIds) {
             await services.lance.initialize(storeId);
           }
+          const limit = parseInt(options.limit, 10);
+          if (Number.isNaN(limit)) {
+            throw new Error(
+              `Invalid value for --limit: "${options.limit}" is not a valid integer`
+            );
+          }
+          let threshold;
+          if (options.threshold !== void 0) {
+            threshold = parseFloat(options.threshold);
+            if (Number.isNaN(threshold)) {
+              throw new Error(
+                `Invalid value for --threshold: "${options.threshold}" is not a valid number`
+              );
+            }
+            if (threshold < 0 || threshold > 1) {
+              throw new Error(`Invalid value for --threshold: must be between 0 and 1`);
+            }
+          }
+          let minRelevance;
+          if (options.minRelevance !== void 0) {
+            minRelevance = parseFloat(options.minRelevance);
+            if (Number.isNaN(minRelevance)) {
+              throw new Error(
+                `Invalid value for --min-relevance: "${options.minRelevance}" is not a valid number`
+              );
+            }
+            if (minRelevance < 0 || minRelevance > 1) {
+              throw new Error(`Invalid value for --min-relevance: must be between 0 and 1`);
+            }
+          }
           const results = await services.search.search({
             query,
             stores: storeIds,
-            mode: options.mode ?? "hybrid",
-            limit: parseInt(options.limit ?? "10", 10),
-            threshold: options.threshold !== void 0 ? parseFloat(options.threshold) : void 0,
-            minRelevance: options.minRelevance !== void 0 ? parseFloat(options.minRelevance) : void 0,
-            includeContent: options.includeContent,
-            detail: options.detail ?? "minimal"
+            mode: options.mode,
+            limit,
+            threshold,
+            minRelevance,
+            detail: options.detail
           });
           if (globalOpts.format === "json") {
             console.log(JSON.stringify(results, null, 2));
@@ -1086,7 +1164,7 @@ function createSearchCommand(getOptions) {
           } else {
             console.log(`
 Search: "${query}"`);
-            let statusLine = `Mode: ${results.mode} | Detail: ${String(options.detail)} | Stores: ${String(results.stores.length)} | Results: ${String(results.totalResults)} | Time: ${String(results.timeMs)}ms`;
+            let statusLine = `Mode: ${results.mode} | Detail: ${options.detail} | Stores: ${String(results.stores.length)} | Results: ${String(results.totalResults)} | Time: ${String(results.timeMs)}ms`;
             if (results.confidence !== void 0) {
               statusLine += ` | Confidence: ${results.confidence}`;
             }
@@ -1168,6 +1246,8 @@ import { serve } from "@hono/node-server";
 import { Command as Command6 } from "commander";
 
 // src/server/app.ts
+import { rm } from "fs/promises";
+import { join as join2 } from "path";
 import { Hono } from "hono";
 import { cors } from "hono/cors";
 import { z } from "zod";
@@ -1229,6 +1309,13 @@ function createApp(services) {
   app.delete("/api/stores/:id", async (c) => {
     const store = await services.store.getByIdOrName(c.req.param("id"));
     if (!store) return c.json({ error: "Not found" }, 404);
+    await services.lance.deleteStore(store.id);
+    await services.codeGraph.deleteGraph(store.id);
+    await services.manifest.delete(store.id);
+    const resolvedDataDir = services.config.resolveDataDir();
+    if (store.type === "repo" && "url" in store && store.url !== void 0 && store.path.startsWith(join2(resolvedDataDir, "repos"))) {
+      await rm(store.path, { recursive: true, force: true });
+    }
     const result = await services.store.delete(store.id);
     if (result.success) return c.json({ deleted: true });
     return c.json({ error: result.error.message }, 400);
@@ -1240,10 +1327,22 @@ function createApp(services) {
       return c.json({ error: parseResult.error.issues[0]?.message ?? "Invalid request body" }, 400);
     }
     const storeIds = (await services.store.list()).map((s) => s.id);
+    services.lance.setDimensions(await services.embeddings.ensureDimensions());
     for (const id of storeIds) {
       await services.lance.initialize(id);
     }
-    const requestedStores = parseResult.data.stores !== void 0 ? parseResult.data.stores.map((s) => createStoreId(s)) : storeIds;
+    let requestedStores = storeIds;
+    if (parseResult.data.stores !== void 0) {
+      const resolvedStores = [];
+      for (const requested of parseResult.data.stores) {
+        const store = await services.store.getByIdOrName(requested);
+        if (store === void 0) {
+          return c.json({ error: `Store not found: ${requested}` }, 404);
+        }
+        resolvedStores.push(store.id);
+      }
+      requestedStores = resolvedStores;
+    }
     const query = {
       query: parseResult.data.query,
       detail: parseResult.data.detail ?? "minimal",
@@ -1256,6 +1355,7 @@ function createApp(services) {
   app.post("/api/stores/:id/index", async (c) => {
     const store = await services.store.getByIdOrName(c.req.param("id"));
     if (!store) return c.json({ error: "Not found" }, 404);
+    services.lance.setDimensions(await services.embeddings.ensureDimensions());
     await services.lance.initialize(store.id);
     const result = await services.index.indexStore(store);
     if (result.success) return c.json(result.data);
@@ -1266,16 +1366,28 @@ function createApp(services) {
 
 // src/cli/commands/serve.ts
 function createServeCommand(getOptions) {
-  return new Command6("serve").description("Start HTTP API server for programmatic search access").option("-p, --port <port>", "Port to listen on (default: 3847)", "3847").option(
+  return new Command6("serve").description("Start HTTP API server for programmatic search access").option("-p, --port <port>", "Port to listen on (reads from config if not specified)").option(
     "--host <host>",
-    "Bind address (default: 127.0.0.1, use 0.0.0.0 for all interfaces)",
-    "127.0.0.1"
+    "Bind address (reads from config if not specified, use 0.0.0.0 for all interfaces)"
   ).action(async (options) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
+    const appConfig = await services.config.load();
     const app = createApp(services);
-    const port = parseInt(options.port ?? "3847", 10);
-    const host = options.host ?? "127.0.0.1";
+    let port;
+    if (options.port !== void 0) {
+      port = parseInt(options.port, 10);
+      if (Number.isNaN(port)) {
+        throw new Error(`Invalid value for --port: "${options.port}" is not a valid integer`);
+      }
+    } else {
+      port = appConfig.server.port;
+    }
+    const host = options.host ?? appConfig.server.host;
     console.log(`Starting server on http://${host}:${String(port)}`);
     const server = serve({
       fetch: app.fetch,
@@ -1300,7 +1412,7 @@ import { spawnSync } from "child_process";
 import { existsSync as existsSync2 } from "fs";
 import { mkdir } from "fs/promises";
 import { homedir } from "os";
-import { join as join2 } from "path";
+import { join as join3 } from "path";
 import { Command as Command7 } from "commander";
 import ora4 from "ora";
 
@@ -1357,7 +1469,7 @@ var DEFAULT_REPOS = [
 ];
 
 // src/cli/commands/setup.ts
-var DEFAULT_REPOS_DIR = join2(homedir(), ".bluera", "bluera-knowledge", "repos");
+var DEFAULT_REPOS_DIR = join3(homedir(), ".bluera", "bluera-knowledge", "repos");
 function createSetupCommand(getOptions) {
   const setup = new Command7("setup").description(
     "Quick-start with pre-configured Claude/Anthropic documentation repos"
@@ -1382,26 +1494,31 @@ function createSetupCommand(getOptions) {
         }
         return;
       }
-      const services = await createServices(globalOpts.config, globalOpts.dataDir);
-      try {
-        let repos = DEFAULT_REPOS;
-        if (options.only !== void 0 && options.only !== "") {
-          const onlyNames = options.only.split(",").map((n) => n.trim().toLowerCase());
-          repos = DEFAULT_REPOS.filter(
-            (r) => onlyNames.some((n) => r.name.toLowerCase().includes(n))
-          );
-          if (repos.length === 0) {
-            console.error(`No repos matched: ${options.only}`);
-            console.log("Available repos:", DEFAULT_REPOS.map((r) => r.name).join(", "));
-            process.exit(1);
-          }
+      let repos = DEFAULT_REPOS;
+      if (options.only !== void 0 && options.only !== "") {
+        const onlyNames = options.only.split(",").map((n) => n.trim().toLowerCase());
+        repos = DEFAULT_REPOS.filter(
+          (r) => onlyNames.some((n) => r.name.toLowerCase().includes(n))
+        );
+        if (repos.length === 0) {
+          console.error(`No repos matched: ${options.only}`);
+          console.log("Available repos:", DEFAULT_REPOS.map((r) => r.name).join(", "));
+          process.exitCode = 1;
+          return;
         }
+      }
+      const services = await createServices(
+        globalOpts.config,
+        globalOpts.dataDir,
+        globalOpts.projectRoot
+      );
+      try {
         console.log(`
 Setting up ${String(repos.length)} repositories...
 `);
         await mkdir(options.reposDir, { recursive: true });
         for (const repo of repos) {
-          const repoPath = join2(options.reposDir, repo.name);
+          const repoPath = join3(options.reposDir, repo.name);
           const spinner = ora4(`Processing ${repo.name}`).start();
           try {
             if (options.skipClone !== true) {
@@ -1450,6 +1567,7 @@ Setting up ${String(repos.length)} repositories...
               spinner.text = `${repo.name}: Indexing...`;
               const store = await services.store.getByIdOrName(storeId);
               if (store) {
+                services.lance.setDimensions(await services.embeddings.ensureDimensions());
                 await services.lance.initialize(store.id);
                 const indexResult = await services.index.indexStore(store, (event) => {
                   if (event.type === "progress") {
@@ -1458,7 +1576,7 @@ Setting up ${String(repos.length)} repositories...
                 });
                 if (indexResult.success) {
                   spinner.succeed(
-                    `${repo.name}: ${String(indexResult.data.documentsIndexed)} docs, ${String(indexResult.data.chunksCreated)} chunks`
+                    `${repo.name}: ${String(indexResult.data.filesIndexed)} docs, ${String(indexResult.data.chunksCreated)} chunks`
                   );
                 } else {
                   throw new Error(
@@ -1485,8 +1603,8 @@ Setting up ${String(repos.length)} repositories...
 }
 
 // src/cli/commands/store.ts
-import { rm } from "fs/promises";
-import { join as join3 } from "path";
+import { rm as rm2 } from "fs/promises";
+import { join as join4 } from "path";
 import { Command as Command8 } from "commander";
 function createStoreCommand(getOptions) {
   const store = new Command8("store").description(
@@ -1494,7 +1612,11 @@ function createStoreCommand(getOptions) {
   );
   store.command("list").description("Show all stores with their type (file/repo/web) and ID").option("-t, --type <type>", "Filter by type: file, repo, or web").action(async (options) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
     try {
       const stores = await services.store.list(options.type);
       if (globalOpts.format === "json") {
@@ -1524,10 +1646,14 @@ function createStoreCommand(getOptions) {
   ).requiredOption("-s, --source <path>", "Local path for file/repo stores, URL for web stores").option("-b, --branch <branch>", "Git branch to clone (repo stores only)").option("-d, --description <desc>", "Optional description for the store").option("--tags <tags>", "Comma-separated tags for filtering").action(
     async (name, options) => {
       const globalOpts = getOptions();
-      const services = await createServices(globalOpts.config, globalOpts.dataDir);
+      const services = await createServices(
+        globalOpts.config,
+        globalOpts.dataDir,
+        globalOpts.projectRoot
+      );
       let exitCode = 0;
       try {
-        const isUrl = options.source.startsWith("http://") || options.source.startsWith("https://");
+        const isUrl = isGitUrl(options.source);
         const result = await services.store.create({
           name,
           type: options.type,
@@ -1559,7 +1685,11 @@ Created store: ${result.data.name} (${result.data.id})
   );
   store.command("info <store>").description("Show store details: ID, type, path/URL, timestamps").action(async (storeIdOrName) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
     let exitCode = 0;
     storeInfo: try {
       const s = await services.store.getByIdOrName(storeIdOrName);
@@ -1591,7 +1721,11 @@ Store: ${s.name}`);
   });
   store.command("delete <store>").description("Remove store and its indexed documents from LanceDB").option("-f, --force", "Delete without confirmation prompt").option("-y, --yes", "Alias for --force").action(async (storeIdOrName, options) => {
     const globalOpts = getOptions();
-    const services = await createServices(globalOpts.config, globalOpts.dataDir);
+    const services = await createServices(
+      globalOpts.config,
+      globalOpts.dataDir,
+      globalOpts.projectRoot
+    );
     let exitCode = 0;
     storeDelete: try {
       const s = await services.store.getByIdOrName(storeIdOrName);
@@ -1625,10 +1759,11 @@ Store: ${s.name}`);
       }
       await services.lance.deleteStore(s.id);
       await services.codeGraph.deleteGraph(s.id);
+      await services.manifest.delete(s.id);
       if (s.type === "repo" && "url" in s && s.url !== void 0) {
         const dataDir = services.config.resolveDataDir();
-        const repoPath = join3(dataDir, "repos", s.id);
-        await rm(repoPath, { recursive: true, force: true });
+        const repoPath = join4(dataDir, "repos", s.id);
+        await rm2(repoPath, { recursive: true, force: true });
       }
       const result = await services.store.delete(s.id);
       if (result.success) {
@@ -1648,6 +1783,8 @@ Store: ${s.name}`);
 }
 
 // src/cli/commands/sync.ts
+import { rm as rm3 } from "fs/promises";
+import { join as join5 } from "path";
 import { Command as Command9 } from "commander";
 async function createStoreFromDefinition(def, defService, storeService) {
   try {
@@ -1693,6 +1830,9 @@ async function createStoreFromDefinition(def, defService, storeService) {
           type: "web",
           url: def.url,
           depth: def.depth,
+          maxPages: def.maxPages,
+          crawlInstructions: def.crawlInstructions,
+          extractInstructions: def.extractInstructions,
           description: def.description,
           tags: def.tags
         },
@@ -1717,7 +1857,7 @@ function createSyncCommand(getOptions) {
   );
   sync.option("--dry-run", "Show what would happen without making changes").option("--prune", "Remove stores not in definitions").option("--reindex", "Re-index existing stores after sync").action(async (options) => {
     const globalOpts = getOptions();
-    const projectRoot = globalOpts.projectRoot ?? process.cwd();
+    const projectRoot = globalOpts.projectRoot ?? ProjectRootService.resolve();
     const defService = new StoreDefinitionService(projectRoot);
     const services = await createServices(globalOpts.config, globalOpts.dataDir, projectRoot);
     try {
@@ -1765,6 +1905,14 @@ function createSyncCommand(getOptions) {
           for (const orphanName of result.orphans) {
             const store = await services.store.getByName(orphanName);
             if (store !== void 0) {
+              await services.lance.deleteStore(store.id);
+              await services.codeGraph.deleteGraph(store.id);
+              await services.manifest.delete(store.id);
+              if (store.type === "repo" && "url" in store && store.url !== void 0) {
+                const dataDir = services.config.resolveDataDir();
+                const repoPath = join5(dataDir, "repos", store.id);
+                await rm3(repoPath, { recursive: true, force: true });
+              }
               const deleteResult = await services.store.delete(store.id, {
                 skipDefinitionSync: true
               });
@@ -1892,13 +2040,13 @@ function printHumanReadable(result, quiet) {
 
 // src/cli/program.ts
 import { readFileSync } from "fs";
-import { dirname, join as join4 } from "path";
+import { dirname, join as join6 } from "path";
 import { fileURLToPath } from "url";
 import { Command as Command10 } from "commander";
 function getVersion() {
   const __filename2 = fileURLToPath(import.meta.url);
   const __dirname2 = dirname(__filename2);
-  const content = readFileSync(join4(__dirname2, "../package.json"), "utf-8");
+  const content = readFileSync(join6(__dirname2, "../package.json"), "utf-8");
   const pkg = JSON.parse(content);
   return pkg.version;
 }
@@ -1906,7 +2054,7 @@ var version = getVersion();
 function createProgram() {
   const program2 = new Command10();
   program2.name("bluera-knowledge").description("CLI tool for managing knowledge stores with semantic search").version(version);
-  program2.option("-c, --config <path>", "Path to config file").option("-d, --data-dir <path>", "Data directory").option("-p, --project-root <path>", "Project root directory (for resolving relative paths)").option("-f, --format <format>", "Output format: json | table | plain", "table").option("-q, --quiet", "Suppress non-essential output").option("-v, --verbose", "Enable verbose logging");
+  program2.option("-c, --config <path>", "Path to config file").option("-d, --data-dir <path>", "Data directory").option("-p, --project-root <path>", "Project root directory (for resolving relative paths)").option("-f, --format <format>", "Output format: json | table | plain", "table").option("-q, --quiet", "Suppress non-essential output");
   return program2;
 }
 function getGlobalOptions(program2) {
@@ -1916,17 +2064,16 @@ function getGlobalOptions(program2) {
     dataDir: opts.dataDir,
     projectRoot: opts.projectRoot,
     format: opts.format,
-    quiet: opts.quiet,
-    verbose: opts.verbose
+    quiet: opts.quiet
   };
 }
 
 // src/index.ts
 var registry = AdapterRegistry.getInstance();
 registry.register(new ZilAdapter());
-var DEFAULT_DATA_DIR = join5(homedir2(), ".bluera", "bluera-knowledge", "data");
-var DEFAULT_CONFIG = join5(homedir2(), ".bluera", "bluera-knowledge", "config.json");
-var DEFAULT_REPOS_DIR2 = join5(homedir2(), ".bluera", "bluera-knowledge", "repos");
+var DEFAULT_DATA_DIR = ".bluera/bluera-knowledge/data";
+var DEFAULT_CONFIG = ".bluera/bluera-knowledge/config.json";
+var DEFAULT_REPOS_DIR2 = ".bluera/bluera-knowledge/data/repos/<store-id>/";
 function formatCommandHelp(cmd, indent = "") {
   const lines = [];
   const name = cmd.name();