@lojban/semantic-search-mcp 1.0.14 → 1.0.16

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lojban/semantic-search-mcp",
-  "version": "1.0.14",
+  "version": "1.0.16",
   "description": "Local-first MCP server for semantic search using transformers.js and SQLite",
   "type": "module",
   "scripts": {

package/src/index.ts

@@ -6,16 +6,17 @@ import {
   ListToolsRequestSchema,
 } from '@modelcontextprotocol/sdk/types.js';
 import path from 'path';
+import os from 'os';
 import { getEmbedding, getBatchEmbeddings } from './embeddings.js';
 import { createVectorStorage, type SearchResult } from './storage.js';
 import { normalizePath } from './path-util.js';
 import { listFilesInDirectories, readFileWithMetadata } from './scanner.js';
 import { getSampuVlasteDir } from './download-sampu-vlaste.js';
 
-// Data dir: use env, or project cwd so each workspace has its own index when run via npx from Cursor
+// Data dir: use env, or $HOME/.semantic-search/data so the same index is used across MCP version upgrades (npx cwd can change per version)
 const dataDir =
   process.env.SEMANTIC_SEARCH_DATA_DIR ||
-  path.join(process.cwd(), '.semantic-search', 'data');
+  path.join(os.homedir(), '.semantic-search', 'data');
 const DB_PATH = path.join(dataDir, 'vectors.db');
 
 // Background indexing state (progress for get_index_stats)

package/src/storage.ts

@@ -128,11 +128,11 @@ export class VectorStorage {
         l.file_path,
         l.line_number,
         l.content,
-        vec_distance_cosine(v.embedding, ?1) AS distance
+        vec_distance_cosine(v.embedding, ?) AS distance
       FROM vec_lines v
       INNER JOIN lines l ON v.line_id = l.id
       ORDER BY distance
-      LIMIT ?2
+      LIMIT ?
     `);
     const rows = stmt.all([queryEmbedding.buffer, limit]) as Array<{
       file_path: string;