changed the logic to use a submodule

Author: zho

src/constants.ts

@@ -2,8 +2,6 @@
  * Constants for Pinecone Read-Only MCP
  */
 
-import { createRequire } from 'node:module';
-
 export const DEFAULT_INDEX_NAME = 'rag-hybrid';
 export const DEFAULT_RERANK_MODEL = 'bge-reranker-v2-m3';
 export const DEFAULT_TOP_K = 10;
@@ -34,52 +32,23 @@ export const MAX_QUERY_DOCUMENTS_TOP_K = 20;
 export const QUERY_DOCUMENTS_MAX_CHUNKS = 500;
 
 export const SERVER_NAME = 'Pinecone Read-Only MCP';
+export { SERVER_VERSION } from './server-version.js';
 
-/**
- * Read package.json at runtime so SERVER_VERSION (advertised via MCP serverInfo)
- * always matches the published package version.
- *
- * `createRequire` is used because:
- * - JSON imports with `with { type: 'json' }` clash with `rootDir: ./src`.
- * - Reading at runtime avoids a build-time codegen step.
- *
- * Falls back to `0.0.0-dev` only if package.json is unreachable (e.g. when
- * dist files are copied without the manifest); operators see a clearly fake
- * version instead of a stale hardcoded number.
- */
-function readPackageVersion(): string {
-  try {
-    const require = createRequire(import.meta.url);
-    const pkg = require('../package.json') as { version?: string };
-    return typeof pkg.version === 'string' && pkg.version.length > 0 ? pkg.version : '0.0.0-dev';
-  } catch {
-    return '0.0.0-dev';
-  }
-}
-
-export const SERVER_VERSION: string = readPackageVersion();
-
-export const SERVER_INSTRUCTIONS = `Quickstart for AI clients: for most user questions, call \`guided_query\` with the user's question — it does namespace routing, suggestion, and execution in one shot and returns a decision_trace you can show the user. For manual flows, call \`list_namespaces\` -> \`suggest_query_params\` -> \`query_fast\` / \`query_detailed\` / \`count\` (the suggest step is a mandatory gate). Use \`query_documents\` for full-document reading, \`keyword_search\` for exact-keyword retrieval against the sparse index, and \`generate_urls\` when records need URLs synthesized from metadata.
-
-A semantic search server that provides hybrid search capabilities over Pinecone vector indexes with automatic namespace discovery.
+export const SERVER_INSTRUCTIONS = `A semantic search server that provides hybrid search capabilities over Pinecone vector indexes with automatic namespace discovery.
 
 Features:
 - Hybrid Search: Combines dense and sparse embeddings for superior recall
-- Semantic Reranking: Uses a configurable reranker model (default bge-reranker-v2-m3) for improved precision
+- Semantic Reranking: Uses BGE reranker model for improved precision
 - Dynamic Namespace Discovery: Automatically discovers available namespaces
 - Metadata Filtering: Supports optional metadata filters for refined searches
 - Namespace Router: Suggests likely namespace(s) from natural-language intent
 - Count: Use the count tool for "how many X?" questions; it uses semantic search only and minimal fields (no content) for performance, returning unique document count.
-- URL Generation: Use generate_urls to synthesize URLs for namespaces that have a registered generator when metadata lacks url.
-- Document reassembly: Use query_documents to get whole documents (chunks grouped and merged by document_number/doc_id/url) for content analysis or summarization. query_documents always reranks for best document-level relevance.
-- Keyword search: Use keyword_search to query the sparse index for lexical/keyword-only retrieval without reranking.
+- URL Generation: Use generate_urls to synthesize URLs for namespaces that support it when metadata lacks url.
+- Document reassembly: Use query_documents to get whole documents (chunks grouped and merged by document_number/doc_id/url) for content analysis or summarization.
+- Keyword search: Use keyword_search to query the sparse index (default: rag-hybrid-sparse) for lexical/keyword-only retrieval without reranking.
 
 Usage:
-1. Use list_namespaces (cached) to discover available namespaces in the index. The response includes \`expires_at_iso\` so you know when to refresh.
-2. Optionally use namespace_router to choose candidate namespace(s) from user intent.
-3. Call suggest_query_params before query/count/query_documents tools (mandatory flow gate) to get suggested_fields and recommended_tool.
-4. Use count for count questions, query_fast/query_detailed for chunk-level retrieval, or query_documents for full-document content.
-
-Notes:
-- Result rows include both \`document_id\` (canonical) and \`paper_number\` (deprecated alias kept for one minor cycle; will be removed in the next major release). Prefer \`document_id\` in new code.
-- The server emits structured logs to stderr (text by default, set PINECONE_READ_ONLY_MCP_LOG_FORMAT=json for log aggregation).`;
+1. Use list_namespaces (cached for 30 minutes) to discover available namespaces in the index
+2. Optionally use namespace_router to choose candidate namespace(s) from user intent
+3. Call suggest_query_params before query/count/query_documents tools (mandatory flow gate) to get suggested_fields and recommended tool
+4. Use count for count questions, query_fast/query_detailed for chunk-level retrieval, or query_documents for full-document content`;