@o-lang/semantic-doc-search 1.0.9 → 1.0.11

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@o-lang/semantic-doc-search",
-  "version": "1.0.9",
+  "version": "1.0.11",
   "description": "O-lang Semantic Document Search Resolver with hybrid search, embeddings, rerank, and streaming.",
   "main": "src/index.js",
   "type": "commonjs",
@@ -23,6 +23,7 @@
     "node-stream-zip": "*",
     "openai": "^4.3.1",
     "pdf-parse": "^1.1.1",
+    "pg": "^8.16.3",
     "pinecone-client": "^1.0.0",
     "readline": "^1.3.0",
     "redis": "^5.2.0"

package/src/index.js

@@ -257,12 +257,50 @@ async function loadAllDocuments(context) {
   return documents;
 }
 
-// ✅ PGVECTOR SEARCH FUNCTION (Using imported adapter)
+// 🔥 AUTO-MIGRATION HELPER FUNCTIONS
+async function checkPgVectorHasData(pgVectorAdapter) {
+  try {
+    const result = await pgVectorAdapter.pool.query('SELECT COUNT(*) FROM doc_embeddings');
+    return parseInt(result.rows[0].count) > 0;
+  } catch (error) {
+    // Table doesn't exist or other error - treat as empty
+    return false;
+  }
+}
+
+async function migrateDocumentsToPgVector(docRoot, pgVectorAdapter, embedder) {
+  const baseDir = safeResolve(process.cwd(), docRoot);
+  if (!fs.existsSync(baseDir)) {
+    console.log('📁 No docs directory found, skipping migration');
+    return;
+  }
+
+  const files = fs.readdirSync(baseDir).filter(f => f.endsWith(".txt") || f.endsWith(".md"));
+  console.log(`🔄 Migrating ${files.length} documents to pgvector...`);
+  
+  for (const file of files) {
+    try {
+      const content = fs.readFileSync(path.join(baseDir, file), "utf8");
+      const vector = await embedder.embed(content);
+      
+      await pgVectorAdapter.upsert({
+        id: file,
+        vector: vector,
+        content: content,
+        source: `file:${file}`
+      });
+      console.log(`✅ Migrated ${file}`);
+    } catch (error) {
+      console.warn(`⚠️ Failed to migrate ${file}: ${error.message}`);
+    }
+  }
+}
+
+// ✅ PGVECTOR SEARCH FUNCTION WITH AUTO-MIGRATION
 async function performPgVectorSearch(query, context = {}) {
   const options = context.options || {};
   const topK = options.topK || 5;
   
-  // Check for POSTGRES_URL in context or environment
   const postgresUrl = context.POSTGRES_URL || process.env.POSTGRES_URL;
   if (!postgresUrl) {
     return { 
@@ -282,6 +320,16 @@ async function performPgVectorSearch(query, context = {}) {
   });
 
   try {
+    // 🔥 AUTO-MIGRATION LOGIC
+    if (context.migrate_on_demand && context.doc_root) {
+      const hasData = await checkPgVectorHasData(pgVectorAdapter);
+      if (!hasData) {
+        console.log('🔄 Auto-migrating documents to pgvector (first run)...');
+        await migrateDocumentsToPgVector(context.doc_root, pgVectorAdapter, embedder);
+        console.log('✅ Migration completed');
+      }
+    }
+
     const queryVector = await embedder.embed(query);
     const docs = await pgVectorAdapter.query(queryVector, topK);
 
@@ -465,7 +513,7 @@ async function performHybridDocQA(query, context = {}) {
 async function performDocQA(query, context = {}) {
   // 🔍 AUTO-DETECT MODE BASED ON CONTEXT
   
-  // Mode 1: pgvector mode (if PostgreSQL URL provided in context or env)
+  // Mode 1: pgvector mode (if PostgreSQL URL provided)
   const postgresUrl = context.POSTGRES_URL || process.env.POSTGRES_URL;
   if (postgresUrl) {
     console.log('🔍 Using pgvector search mode');