@o-lang/semantic-doc-search 1.0.3 → 1.0.5

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@o-lang/semantic-doc-search",
-  "version": "1.0.3",
+  "version": "1.0.5",
   "description": "O-lang Semantic Document Search Resolver with hybrid search, embeddings, rerank, and streaming.",
   "main": "src/index.js",
   "type": "module",

package/src/index.js

@@ -1,67 +1,12 @@
 // doc-search.js
-const fs = require('fs');
-const path = require('path');
-
-// Load your utility modules - you'll need to convert these to CommonJS too
-// For now, we'll create simplified versions inline
-
-// ✅ Simplified LocalEmbedding (replace with your actual implementation)
-class LocalEmbedding {
-  embed(text) {
-    // Simple hash-based "embedding" for demo
-    // Replace with your actual embedding logic
-    const hash = text.split('').reduce((acc, char) => acc + char.charCodeAt(0), 0);
-    return Array(1536).fill(0).map((_, i) => Math.sin(hash + i));
-  }
-}
-
-// ✅ Simplified utility functions
-function chunkText(text, chunkSize, overlap) {
-  const chunks = [];
-  let start = 0;
-  while (start < text.length) {
-    chunks.push(text.slice(start, start + chunkSize));
-    start += chunkSize - overlap;
-  }
-  return chunks;
-}
-
-function extractKeywords(text) {
-  return text.toLowerCase().match(/\b\w{3,}\b/g) || [];
-}
-
-function cosine(vec1, vec2) {
-  if (!vec1 || !vec2) return 0;
-  const dot = vec1.reduce((sum, val, i) => sum + val * (vec2[i] || 0), 0);
-  const mag1 = Math.sqrt(vec1.reduce((sum, val) => sum + val * val, 0));
-  const mag2 = Math.sqrt(vec2.reduce((sum, val) => sum + val * val, 0));
-  return dot / (mag1 * mag2) || 0;
-}
-
-function highlightMatches(text, keywords) {
-  let result = text;
-  keywords.forEach(keyword => {
-    const regex = new RegExp(`(${keyword})`, 'gi');
-    result = result.replace(regex, '**$1**');
-  });
-  return result;
-}
-
-// LLM creation - simplified for now
-function createLLM(config) {
-  // This would integrate with your actual LLM router
-  // For now, return a mock that just returns the prompt
-  return {
-    async generate({ prompt }) {
-      return { text: prompt };
-    },
-    async stream({ prompt, onToken }) {
-      // Mock streaming
-      const words = prompt.split(' ');
-      words.forEach(word => onToken?.(word + ' '));
-    }
-  };
-}
+const fs = require("fs");
+const path = require("path");
+const { createLLM } = require("./llm/router.js");
+const { LocalEmbedding } = require("./embeddings/local.js");
+const { chunkText } = require("./utils/chunker.js");
+const { extractKeywords } = require("./utils/extractText.js");
+const { cosine } = require("./utils/similarity.js");
+const { highlightMatches } = require("./utils/highlight.js");
 
 const CACHE_PATH = path.join(process.cwd(), "embeddings.json");
 
@@ -88,7 +33,7 @@ function saveCache(cache) {
   } catch {}
 }
 
-// ✅ UNIVERSAL DATABASE ADAPTER
+// ✅ UNIVERSAL DATABASE ADAPTER (NEW - Keep your existing imports!)
 class DatabaseAdapter {
   constructor() {
     this.initialized = false;
@@ -97,6 +42,7 @@ class DatabaseAdapter {
   async initialize(context) {
     if (this.initialized) return;
     
+    // Initialize based on context configuration
     if (context.db_type === 'mongodb' || context.MONGO_URI) {
       await this.initMongo(context);
     } else if (context.db_type === 'sqlite' || context.db_path) {
@@ -149,6 +95,7 @@ class DatabaseAdapter {
       password: context.DB_PASSWORD,
       database: context.DB_NAME || 'olang'
     };
+    // Remove undefined/null values
     Object.keys(poolConfig).forEach(key => {
       if (poolConfig[key] === undefined || poolConfig[key] === null) {
         delete poolConfig[key];
@@ -211,6 +158,7 @@ class DatabaseAdapter {
       try {
         filter = JSON.parse(doc_filter);
       } catch {
+        // Text search fallback
         filter = { $text: { $search: doc_filter } };
       }
     } else if (typeof doc_filter === 'object' && Object.keys(doc_filter).length > 0) {
@@ -232,6 +180,7 @@ class DatabaseAdapter {
       doc_params = []
     } = context;
     
+    // Parse doc_params from string if needed
     let params = doc_params;
     if (typeof doc_params === 'string') {
       try {
@@ -266,7 +215,7 @@ class DatabaseAdapter {
 // ✅ LOAD DOCUMENTS FROM DATABASE (if configured)
 async function loadDocumentsFromDatabase(context) {
   if (!context.db_type && !context.db_path && !context.MONGO_URI && !context.POSTGRES_URL) {
-    return null;
+    return null; // No database configured
   }
   
   const dbAdapter = new DatabaseAdapter();
@@ -283,11 +232,13 @@ async function loadDocumentsFromDatabase(context) {
 async function loadAllDocuments(context) {
   const documents = [];
   
+  // 1. Load from database first (if configured)
   const dbDocs = await loadDocumentsFromDatabase(context);
   if (dbDocs) {
     documents.push(...dbDocs);
   }
   
+  // 2. Load from file system (existing behavior)
   const baseDir = context.doc_root 
     ? safeResolve(process.cwd(), context.doc_root)
     : path.join(process.cwd(), "docs");
@@ -311,7 +262,7 @@ async function loadAllDocuments(context) {
   return documents;
 }
 
-// ✅ MAIN SEARCH FUNCTION
+// ✅ MAIN SEARCH FUNCTION (Your existing logic + universal docs)
 async function performDocQA(query, context = {}) {
   const { doc_root, stream = false } = context;
   const options = context.options || {};
@@ -325,6 +276,7 @@ async function performDocQA(query, context = {}) {
     return { text: "Missing required input: query" };
   }
 
+  // Load documents from both database and files
   const allDocs = await loadAllDocuments(context);
   if (!allDocs || !allDocs.length) {
     return { text: "No documents available." };
@@ -444,7 +396,7 @@ async function performDocQA(query, context = {}) {
   };
 }
 
-// ✅ O-Lang Resolver Interface - COMMONJS STYLE
+// ✅ O-Lang Resolver Interface (Your existing interface - converted to CommonJS)
 async function docSearchResolver(action, context) {
   if (action.startsWith('Ask doc-search ')) {
     const match = action.match(/"(.*)"|'(.*)'/);
@@ -454,8 +406,8 @@ async function docSearchResolver(action, context) {
   return undefined;
 }
 
-// ✅ Resolver name
+// ✅ Resolver name matches package name: @o-lang/doc-search → doc-search
 docSearchResolver.resolverName = 'doc-search';
 
-// ✅ COMMONJS EXPORT
+// ✅ COMMONJS EXPORT (this is the key change)
 module.exports = docSearchResolver;