NCBI Entrez MCP Server

import { z } from "zod"; import { BaseTool } from "./base.js"; import { getParserForTool } from "../lib/parsers.js"; export class EInfoTool extends BaseTool { register(): void { this.context.server.tool( "einfo", "Get comprehensive metadata about NCBI Entrez databases including field statistics, last update dates, and available links to other databases. Covers all 38+ Entrez databases from PubMed to protein sequences.\n\n📚 **Major Databases**: pubmed (literature), protein (sequences), nuccore (nucleotides), gene (gene records), structure (3D structures)\n🔗 **Use Cases**: \n• Discover searchable fields for precise queries\n• Find cross-database links for related data\n• Check database size and update status\n• Plan search strategies with field information", { db: z .string() .optional() .describe( "Database name (optional). If not provided, returns list of all databases", ), version: z .string() .optional() .describe("Version 2.0 for enhanced XML output"), retmode: z .enum(["xml", "json"]) .optional() .default("xml") .describe("Output format"), }, async ({ db, version, retmode }) => { try { // Validate database if provided if (db && !this.isValidDatabase(db)) { throw new Error(`Invalid database name: ${db}`); } const params = new URLSearchParams({ tool: this.context.defaultTool, email: this.context.defaultEmail, retmode: retmode || "xml", }); if (db) params.append("db", db); if (version) params.append("version", version); const url = this.buildUrl("einfo.fcgi", params); const response = await fetch(url); const data = await this.parseResponse(response, "EInfo", retmode); // Check if we should stage this response const stagingDecision = this.context.shouldStageResponse( data, "EInfo", ); if (stagingDecision.shouldStage) { // Parse and stage the response const parser = getParserForTool("EInfo", data); const parseResult = parser.parse(data); // Stage the data using the DO const env = this.getEnvironment(); if (!env?.JSON_TO_SQL_DO) { return { content: [ { type: "text", text: `EInfo Results (optimized, ${stagingDecision.estimatedTokens} tokens):\n\n${this.formatResponseData(data)}`, }, ], }; } const doId = env.JSON_TO_SQL_DO.newUniqueId(); const stub = env.JSON_TO_SQL_DO.get(doId); const stagingResponse = await stub.fetch("http://do/process", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify(parseResult), }); const stagingResult = (await stagingResponse.json()) as any; return { content: [ { type: "text", text: `✅ **EInfo Data Successfully Staged in SQL Database**\n\n🗃️ **Data Access ID**: \`${stagingResult.data_access_id}\`\n📊 **Records Staged**: ${stagingResult.processing_details?.total_rows || 0} rows across ${stagingResult.processing_details?.table_count || 0} tables\n📋 **Tables Created**: ${(stagingResult.processing_details?.tables_created || []).join(", ")}\n\n## 🚀 Quick Start Queries:\n1. \`SELECT name, full_name FROM searchable_field WHERE is_date = 1\` - Get date fields\n2. \`SELECT name, target_db FROM link_info ORDER BY target_db\` - Get available database links\n3. \`SELECT * FROM database_info\` - Get database metadata\n\n## 📋 Next Steps:\n• Use **\`query_staged_data\`** with the data_access_id above to run any SQL query\n• Use **\`get_staged_schema\`** to see full table structures and advanced query examples\n\n💡 **Pro tip**: Query field information to build better search strategies for this database!`, }, ], }; } return { content: [ { type: "text", text: `EInfo Results (${stagingDecision.estimatedTokens} estimated tokens):\n\n${this.formatResponseData(data)}`, }, ], }; } catch (error) { return { content: [ { type: "text", text: `Error in EInfo: ${error instanceof Error ? error.message : String(error)}`, }, ], }; } }, ); } }