snow-ai 0.6.52 → 0.6.53

package/bundle/cli.mjs

@@ -12,6 +12,18 @@ const require = Object.assign((moduleName) => {
 const __filename = _fileURLToPath(import.meta.url);
 const __dirname = _fileURLToPath(new URL('.', import.meta.url));
 
+// Pre-load @microsoft/signalr runtime dependencies into require.cache
+// SignalR uses dynamic require() which esbuild cannot bundle statically
+// We load them here so they're available when SignalR tries to require() them
+const __signalr_deps = {
+  'abort-controller': require('abort-controller'),
+  'eventsource': require('eventsource'),
+  'fetch-cookie': require('fetch-cookie'),
+  'node-fetch': require('node-fetch'),
+  'tough-cookie': require('tough-cookie'),
+  'ws': require('ws')
+};
+
 // Polyfill for @microsoft/signalr dynamic require
 // SignalR uses: const requireFunc = typeof __webpack_require__ === "function" ? __non_webpack_require__ : require;
 // Keep __non_webpack_require__ aligned with our wrapped require for both branches.
@@ -64512,7 +64524,7 @@ var init_streamableHttp = __esm({
   }
 });
 
-// node_modules/eventsource/dist/index.js
+// node_modules/@modelcontextprotocol/sdk/node_modules/eventsource/dist/index.js
 function syntaxError(message) {
   const DomException = globalThis.DOMException;
   return typeof DomException == "function" ? new DomException(message, "SyntaxError") : new SyntaxError(message);
@@ -64536,7 +64548,7 @@ function getBaseURL() {
 }
 var ErrorEvent2, __typeError2, __accessCheck2, __privateGet2, __privateAdd2, __privateSet2, __privateMethod2, _readyState, _url, _redirectUrl, _withCredentials, _fetch, _reconnectInterval, _reconnectTimer, _lastEventId, _controller, _parser, _onError, _onMessage, _onOpen, _EventSource_instances, connect_fn, _onFetchResponse, _onFetchError, getRequestOptions_fn, _onEvent, _onRetryChange, failConnection_fn, scheduleReconnect_fn, _reconnect, EventSource2;
 var init_dist5 = __esm({
-  "node_modules/eventsource/dist/index.js"() {
+  "node_modules/@modelcontextprotocol/sdk/node_modules/eventsource/dist/index.js"() {
     init_dist4();
     ErrorEvent2 = class extends Event {
       /**
@@ -579305,6 +579317,7 @@ var init_codebaseIndexAgent = __esm({
     init_embedding();
     init_codebaseConfig();
     init_retryUtils();
+    init_office_parser_utils();
     CodebaseIndexAgent = class _CodebaseIndexAgent {
       constructor(projectRoot) {
         Object.defineProperty(this, "db", {
@@ -579580,8 +579593,8 @@ var init_codebaseIndexAgent = __esm({
             persistent: true
           });
           this.fileWatcher.on("add", (filePath) => {
-            const ext = path55.extname(filePath);
-            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext)) {
+            const ext = path55.extname(filePath).toLowerCase();
+            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext) && !_CodebaseIndexAgent.OFFICE_EXTENSIONS.has(ext)) {
               return;
             }
             const relativePath = path55.relative(this.projectRoot, filePath);
@@ -579589,8 +579602,8 @@ var init_codebaseIndexAgent = __esm({
             this.debounceFileChange(filePath, relativePath);
           });
           this.fileWatcher.on("change", (filePath) => {
-            const ext = path55.extname(filePath);
-            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext)) {
+            const ext = path55.extname(filePath).toLowerCase();
+            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext) && !_CodebaseIndexAgent.OFFICE_EXTENSIONS.has(ext)) {
               return;
             }
             const relativePath = path55.relative(this.projectRoot, filePath);
@@ -579598,8 +579611,8 @@ var init_codebaseIndexAgent = __esm({
             this.debounceFileChange(filePath, relativePath);
           });
           this.fileWatcher.on("unlink", (filePath) => {
-            const ext = path55.extname(filePath);
-            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext)) {
+            const ext = path55.extname(filePath).toLowerCase();
+            if (!_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext) && !_CodebaseIndexAgent.OFFICE_EXTENSIONS.has(ext)) {
               return;
             }
             const relativePath = path55.relative(this.projectRoot, filePath);
@@ -579734,7 +579747,7 @@ var init_codebaseIndexAgent = __esm({
               scanDir(fullPath);
             } else if (entry.isFile()) {
               const ext = path55.extname(entry.name);
-              if (_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext)) {
+              if (_CodebaseIndexAgent.CODE_EXTENSIONS.has(ext) || _CodebaseIndexAgent.OFFICE_EXTENSIONS.has(ext)) {
                 files.push(fullPath);
               }
             }
@@ -579775,14 +579788,28 @@ var init_codebaseIndexAgent = __esm({
             currentFile: relativePath,
             status: "indexing"
           });
-          const content = fs49.readFileSync(filePath, "utf-8");
-          const fileHash = crypto6.createHash("sha256").update(content).digest("hex");
+          const ext = path55.extname(filePath).toLowerCase();
+          const isOfficeFile = _CodebaseIndexAgent.OFFICE_EXTENSIONS.has(ext);
+          let content;
+          let fileHash;
+          if (isOfficeFile) {
+            const docContent = await readOfficeDocument(filePath);
+            if (!docContent) {
+              logger.warn(`Failed to parse Office document: ${relativePath}`);
+              return;
+            }
+            content = docContent.text;
+            fileHash = crypto6.createHash("sha256").update(content).digest("hex");
+          } else {
+            content = fs49.readFileSync(filePath, "utf-8");
+            fileHash = crypto6.createHash("sha256").update(content).digest("hex");
+          }
           if (this.db.hasFileHash(fileHash)) {
             logger.debug(`File unchanged, skipping: ${relativePath}`);
             return;
           }
           this.db.deleteChunksByFile(relativePath);
-          const chunks = this.splitIntoChunks(content, relativePath);
+          const chunks = isOfficeFile ? this.splitDocumentIntoChunks(content, relativePath) : this.splitIntoChunks(content, relativePath);
           if (chunks.length === 0) {
             logger.debug(`No chunks generated for: ${relativePath}`);
             return;
@@ -579897,6 +579924,73 @@ var init_codebaseIndexAgent = __esm({
         }
         return chunks;
       }
+      /**
+       * Split document content into chunks based on semantic boundaries
+       * Documents (PDF, Word, etc.) need different chunking than code files
+       * - Uses paragraph boundaries instead of fixed line counts
+       * - Respects heading structures
+       * - Maintains semantic coherence
+       */
+      splitDocumentIntoChunks(content, filePath) {
+        const chunks = [];
+        const MAX_CHUNK_CHARS = 3e3;
+        const MIN_CHUNK_CHARS = 200;
+        const paragraphs = content.split(/\n{2,}/).map((p) => p.trim()).filter((p) => p.length > 0);
+        if (paragraphs.length === 0) {
+          return chunks;
+        }
+        let currentChunk = [];
+        let currentCharCount = 0;
+        let startParagraph = 0;
+        for (let i = 0; i < paragraphs.length; i++) {
+          const paragraph3 = paragraphs[i];
+          const paraLength = paragraph3.length;
+          if (currentCharCount + paraLength > MAX_CHUNK_CHARS && currentChunk.length > 0) {
+            const chunkContent = currentChunk.join("\n\n");
+            if (chunkContent.length >= MIN_CHUNK_CHARS) {
+              chunks.push({
+                filePath,
+                content: chunkContent,
+                startLine: startParagraph + 1,
+                // Use paragraph index (1-based)
+                endLine: i,
+                // End paragraph index
+                embedding: [],
+                fileHash: "",
+                createdAt: 0,
+                updatedAt: 0
+              });
+            }
+            const overlapStart = Math.max(0, currentChunk.length - 1);
+            currentChunk = currentChunk.slice(overlapStart);
+            currentCharCount = currentChunk.reduce((sum, p) => sum + p.length, 0);
+            startParagraph = i - currentChunk.length;
+          }
+          currentChunk.push(paragraph3);
+          currentCharCount += paraLength;
+        }
+        if (currentChunk.length > 0) {
+          const chunkContent = currentChunk.join("\n\n");
+          if (chunkContent.length >= MIN_CHUNK_CHARS) {
+            chunks.push({
+              filePath,
+              content: chunkContent,
+              startLine: startParagraph + 1,
+              endLine: paragraphs.length,
+              embedding: [],
+              fileHash: "",
+              createdAt: 0,
+              updatedAt: 0
+            });
+          } else if (chunks.length > 0) {
+            const lastChunk = chunks[chunks.length - 1];
+            lastChunk.content += "\n\n" + chunkContent;
+            lastChunk.endLine = paragraphs.length;
+          }
+        }
+        logger.debug(`Document split into ${chunks.length} semantic chunks for: ${filePath}`);
+        return chunks;
+      }
       /**
        * Notify progress to callback
        */
@@ -579935,6 +580029,7 @@ var init_codebaseIndexAgent = __esm({
         ".sh",
         ".bash",
         ".sql",
+        ".txt",
         ".graphql",
         ".proto",
         ".json",
@@ -579950,6 +580045,20 @@ var init_codebaseIndexAgent = __esm({
         ".svelte"
       ])
     });
+    Object.defineProperty(CodebaseIndexAgent, "OFFICE_EXTENSIONS", {
+      enumerable: true,
+      configurable: true,
+      writable: true,
+      value: /* @__PURE__ */ new Set([
+        ".pdf",
+        ".docx",
+        ".doc",
+        ".xlsx",
+        ".xls",
+        ".pptx",
+        ".ppt"
+      ])
+    });
   }
 });
 
@@ -599829,7 +599938,7 @@ var require_package3 = __commonJS({
   "package.json"(exports2, module2) {
     module2.exports = {
       name: "snow-ai",
-      version: "0.6.52",
+      version: "0.6.53",
       description: "Agentic coding in your terminal",
       license: "MIT",
       bin: {
@@ -599876,13 +599985,18 @@ var require_package3 = __commonJS({
         "scripts"
       ],
       dependencies: {
+        "@microsoft/signalr": "^10.0.0",
+        "abort-controller": "^3.0.0",
+        eventsource: "^2.0.2",
+        "fetch-cookie": "^3.0.1",
+        "node-fetch": "^2.7.0",
         ssh2: "^1.17.0",
+        "tough-cookie": "^4.1.3",
         ws: "^8.14.2"
       },
       devDependencies: {
         "@agentclientprotocol/sdk": "^0.14.1",
         "@inkjs/ui": "^2.0.0",
-        "@microsoft/signalr": "^10.0.0",
         "@modelcontextprotocol/sdk": "^1.17.3",
         "@sindresorhus/tsconfig": "^3.0.1",
         "@types/diff": "^7.0.2",

package/bundle/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "snow-ai",
-	"version": "0.6.52",
+	"version": "0.6.53",
 	"description": "Agentic coding in your terminal",
 	"license": "MIT",
 	"bin": {
@@ -47,13 +47,18 @@
 		"scripts"
 	],
 	"dependencies": {
+		"@microsoft/signalr": "^10.0.0",
+		"abort-controller": "^3.0.0",
+		"eventsource": "^2.0.2",
+		"fetch-cookie": "^3.0.1",
+		"node-fetch": "^2.7.0",
 		"ssh2": "^1.17.0",
+		"tough-cookie": "^4.1.3",
 		"ws": "^8.14.2"
 	},
 	"devDependencies": {
 		"@agentclientprotocol/sdk": "^0.14.1",
 		"@inkjs/ui": "^2.0.0",
-		"@microsoft/signalr": "^10.0.0",
 		"@modelcontextprotocol/sdk": "^1.17.3",
 		"@sindresorhus/tsconfig": "^3.0.1",
 		"@types/diff": "^7.0.2",

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "snow-ai",
-	"version": "0.6.52",
+	"version": "0.6.53",
 	"description": "Agentic coding in your terminal",
 	"license": "MIT",
 	"bin": {
@@ -47,13 +47,18 @@
 		"scripts"
 	],
 	"dependencies": {
+		"@microsoft/signalr": "^10.0.0",
+		"abort-controller": "^3.0.0",
+		"eventsource": "^2.0.2",
+		"fetch-cookie": "^3.0.1",
+		"node-fetch": "^2.7.0",
 		"ssh2": "^1.17.0",
+		"tough-cookie": "^4.1.3",
 		"ws": "^8.14.2"
 	},
 	"devDependencies": {
 		"@agentclientprotocol/sdk": "^0.14.1",
 		"@inkjs/ui": "^2.0.0",
-		"@microsoft/signalr": "^10.0.0",
 		"@modelcontextprotocol/sdk": "^1.17.3",
 		"@sindresorhus/tsconfig": "^3.0.1",
 		"@types/diff": "^7.0.2",