fix(embeddings): fall back to text-bearing file nodes

gitnexus/src/core/embeddings/embedding-pipeline.ts

@@ -180,7 +180,50 @@ const queryEmbeddableNodes = async (
     }
   }
 
-  return allNodes;
+  return allNodes.length > 0 ? allNodes : queryFallbackFileNodes(executeQuery);
+};
+
+/**
+ * Static and documentation repositories may contain no code symbols while
+ * still persisting useful text on File nodes. Keep File embeddings as a
+ * zero-symbol fallback so code repositories retain symbol-first selection.
+ */
+const queryFallbackFileNodes = async (
+  executeQuery: (cypher: string) => Promise<any[]>,
+): Promise<EmbeddableNode[]> => {
+  try {
+    const rows = await executeQuery(`
+      MATCH (n:File)
+      RETURN n.id AS id, n.name AS name, 'File' AS label,
+             n.filePath AS filePath, n.content AS content
+    `);
+
+    return rows
+      .map((row) => {
+        const content = row.content ?? row[4] ?? '';
+        return {
+          id: row.id ?? row[0],
+          name: row.name ?? row[1],
+          label: row.label ?? row[2] ?? 'File',
+          filePath: row.filePath ?? row[3],
+          content,
+          startLine: 1,
+          endLine: Math.max(1, content.split('\n').length),
+        };
+      })
+      .filter(
+        (node) =>
+          node.id &&
+          node.filePath &&
+          node.content.trim() &&
+          node.content !== '[Binary file - content not stored]',
+      );
+  } catch (error) {
+    if (isDev) {
+      logger.warn({ error }, 'Fallback File-node embedding query failed:');
+    }
+    return [];
+  }
 };
 
 /**

gitnexus/test/unit/embedding-pipeline.test.ts

@@ -289,6 +289,78 @@ describe('runEmbeddingPipeline incremental filter', () => {
     progressUpdates.push({ ...p });
   };
 
+  it('falls back to text-bearing File nodes when a repo has no code symbols', async () => {
+    mockEmbedderSetup();
+
+    const fileNode = makeNode({
+      id: 'File:README.md',
+      name: 'README.md',
+      label: 'File',
+      filePath: 'README.md',
+      content: '# Static Site\n\nDeployment and recovery notes.',
+      startLine: 1,
+      endLine: 3,
+    });
+    const emptyFile = makeNode({
+      id: 'File:empty.txt',
+      name: 'empty.txt',
+      label: 'File',
+      filePath: 'empty.txt',
+      content: '   ',
+    });
+    const binaryFile = makeNode({
+      id: 'File:logo.png',
+      name: 'logo.png',
+      label: 'File',
+      filePath: 'logo.png',
+      content: '[Binary file - content not stored]',
+    });
+    const executeQuery = mockExecuteQuery([fileNode, emptyFile, binaryFile]);
+    const executeWithReusedStatement = mockExecuteWithReusedStatement();
+
+    const { runEmbeddingPipeline } =
+      await import('../../src/core/embeddings/embedding-pipeline.js');
+
+    const result = await runEmbeddingPipeline(executeQuery, executeWithReusedStatement, onProgress);
+
+    expect(queryCalls.some((cypher) => cypher.includes('MATCH (n:File)'))).toBe(true);
+    const insertedNodeIds = stmtCalls
+      .filter((call) => call.cypher.includes('CREATE'))
+      .flatMap((call) => call.params.map((param) => param.nodeId));
+    expect(insertedNodeIds).toContain(fileNode.id);
+    expect(insertedNodeIds).not.toContain(emptyFile.id);
+    expect(insertedNodeIds).not.toContain(binaryFile.id);
+    expect(result.nodesProcessed).toBe(1);
+  });
+
+  it('retains symbol-first selection when code symbols exist', async () => {
+    mockEmbedderSetup();
+
+    const functionNode = makeNode();
+    const fileNode = makeNode({
+      id: 'File:src/main.ts',
+      name: 'main.ts',
+      label: 'File',
+      filePath: 'src/main.ts',
+      content: 'function foo() { return 1; }',
+    });
+    const executeQuery = mockExecuteQuery([functionNode, fileNode]);
+    const executeWithReusedStatement = mockExecuteWithReusedStatement();
+
+    const { runEmbeddingPipeline } =
+      await import('../../src/core/embeddings/embedding-pipeline.js');
+
+    const result = await runEmbeddingPipeline(executeQuery, executeWithReusedStatement, onProgress);
+
+    expect(queryCalls.some((cypher) => cypher.includes('MATCH (n:File)'))).toBe(false);
+    const insertedNodeIds = stmtCalls
+      .filter((call) => call.cypher.includes('CREATE'))
+      .flatMap((call) => call.params.map((param) => param.nodeId));
+    expect(insertedNodeIds).toContain(functionNode.id);
+    expect(insertedNodeIds).not.toContain(fileNode.id);
+    expect(result.nodesProcessed).toBe(1);
+  });
+
   it('skips unchanged nodes when hash matches', async () => {
     mockEmbedderSetup();