FrankChen021 commented on code in PR #18143:
URL: https://github.com/apache/druid/pull/18143#discussion_r2184211165
##########
mcp-server/src/index.ts:
##########
@@ -0,0 +1,390 @@
+#!/usr/bin/env node
+
+import { Server } from '@modelcontextprotocol/sdk/server/index.js';
+import { StdioServerTransport } from
'@modelcontextprotocol/sdk/server/stdio.js';
+import {
+ CallToolRequestSchema,
+ ListResourcesRequestSchema,
+ ListToolsRequestSchema,
+ ReadResourceRequestSchema,
+ CallToolRequest,
+ CallToolResult,
+ ListResourcesResult,
+ ListToolsResult,
+ ReadResourceRequest,
+ ReadResourceResult,
+ Tool,
+ Resource,
+} from '@modelcontextprotocol/sdk/types.js';
+import { DruidClient, DruidConfig } from './druid-client.js';
+
+/**
+ * Get configuration from environment variables
+ */
+function getConfig(): DruidConfig {
+ const url = process.env.DRUID_URL || 'http://localhost:8888';
+ const username = process.env.DRUID_USERNAME;
+ const password = process.env.DRUID_PASSWORD;
+ const timeout = process.env.DRUID_TIMEOUT ?
parseInt(process.env.DRUID_TIMEOUT) : 30000;
+
+ return {
+ url,
+ username,
+ password,
+ timeout,
+ };
+}
+
+/**
+ * Display help message
+ */
+function showHelp(): void {
+ console.error(`
+Apache Druid MCP Server v1.1.0
+
+USAGE:
+ apache-druid-mcp [OPTIONS]
+
+OPTIONS:
+ --help Show this help message
+
+ENVIRONMENT VARIABLES:
+ DRUID_URL Druid broker URL (default: http://localhost:8888)
+ DRUID_USERNAME Optional username for authentication
+ DRUID_PASSWORD Optional password for authentication
+ DRUID_TIMEOUT Request timeout in milliseconds (default: 30000)
+
+EXAMPLES:
+ # Default configuration
+ apache-druid-mcp
+
+ # Custom Druid URL
+ DRUID_URL=http://production-druid:8888 apache-druid-mcp
+
+ # With authentication
+ DRUID_URL=https://secure-druid.example.com:8888 \\
+ DRUID_USERNAME=admin \\
+ DRUID_PASSWORD=secret \\
+ apache-druid-mcp
+`);
+}
+
+// Initialize Druid client
+const config = getConfig();
+const druidClient = new DruidClient(config);
+
+/**
+ * List available tools
+ */
+async function listTools(): Promise<ListToolsResult> {
+ const tools: Tool[] = [
+ {
+ name: 'execute_sql_query',
+ description: 'Execute a SQL query against Apache Druid and return
results',
+ inputSchema: {
+ type: 'object',
+ properties: {
+ query: {
+ type: 'string',
+ description: 'SQL query to execute (e.g., SELECT * FROM datasource
LIMIT 10)',
+ },
+ context: {
+ type: 'object',
+ description: 'Optional query context parameters',
+ additionalProperties: true,
+ },
+ },
+ required: ['query'],
+ },
+ },
+ {
+ name: 'list_datasources',
+ description: 'Get a list of all available datasources in Druid',
+ inputSchema: {
+ type: 'object',
+ properties: {},
+ },
+ },
+ {
+ name: 'get_datasource_metadata',
+ description: 'Get detailed metadata for a specific datasource including
schema, size, and segments',
+ inputSchema: {
+ type: 'object',
+ properties: {
+ datasource: {
+ type: 'string',
+ description: 'Name of the datasource to get metadata for',
+ },
+ },
+ required: ['datasource'],
+ },
+ },
+ {
+ name: 'test_connection',
+ description: 'Test connection to the Druid cluster',
+ inputSchema: {
+ type: 'object',
+ properties: {},
+ },
+ },
+ ];
+
+ return { tools };
+}
+
+/**
+ * Handle tool calls
+ */
+async function callTool(request: CallToolRequest): Promise<CallToolResult> {
+ try {
+ switch (request.params.name) {
+ case 'execute_sql_query':
+ return await handleExecuteSqlQuery(request);
+ case 'list_datasources':
+ return await handleListDatasources();
+ case 'get_datasource_metadata':
+ return await handleGetDatasourceMetadata(request);
+ case 'test_connection':
+ return await handleTestConnection();
+ default:
+ throw new Error(`Unknown tool: ${request.params.name}`);
+ }
+ } catch (error) {
+ const errorMessage = error instanceof Error ? error.message : 'Unknown
error occurred';
+ return {
+ content: [
+ {
+ type: 'text',
+ text: `Error: ${errorMessage}`,
+ },
+ ],
+ isError: true,
+ };
+ }
+}
+
+/**
+ * List available resources
+ */
+async function listResources(): Promise<ListResourcesResult> {
+ try {
+ const datasources = await druidClient.getDatasources();
+
+ const resources: Resource[] = [
+ {
+ uri: 'druid://cluster/status',
+ name: 'Cluster Status',
+ description: 'Current status and health of the Druid cluster',
+ mimeType: 'application/json',
+ },
+ {
+ uri: 'druid://datasources',
+ name: 'All Datasources',
+ description: 'List of all available datasources in the cluster',
+ mimeType: 'application/json',
+ },
+ ...datasources.map((ds) => ({
+ uri: `druid://datasource/${ds}`,
+ name: `Datasource: ${ds}`,
+ description: `Metadata and schema information for datasource '${ds}'`,
+ mimeType: 'application/json',
+ })),
+ ];
+
+ return { resources };
+ } catch (error) {
+ // Return basic resources if we can't connect to Druid
+ const resources: Resource[] = [
+ {
+ uri: 'druid://cluster/status',
+ name: 'Cluster Status',
+ description: 'Current status and health of the Druid cluster',
+ mimeType: 'application/json',
+ },
+ {
+ uri: 'druid://datasources',
+ name: 'All Datasources',
+ description: 'List of all available datasources in the cluster',
+ mimeType: 'application/json',
+ },
+ ];
+ return { resources };
+ }
+}
+
+/**
+ * Read a specific resource
+ */
+async function readResource(request: ReadResourceRequest):
Promise<ReadResourceResult> {
+ try {
+ const uri = request.params.uri;
+
+ if (uri === 'druid://cluster/status') {
+ const status = await druidClient.getStatus();
+ return {
+ contents: [
+ {
+ uri,
+ mimeType: 'application/json',
+ text: JSON.stringify(status, null, 2),
+ },
+ ],
+ };
+ }
+
+ if (uri === 'druid://datasources') {
+ const datasources = await druidClient.getDatasources();
+ return {
+ contents: [
+ {
+ uri,
+ mimeType: 'application/json',
+ text: JSON.stringify(datasources, null, 2),
+ },
+ ],
+ };
+ }
+
+ const datasourceMatch = uri.match(/^druid:\/\/datasource\/(.+)$/);
+ if (datasourceMatch) {
+ const datasource = datasourceMatch[1];
+ const metadata = await druidClient.getDatasourceMetadata(datasource);
+ return {
+ contents: [
+ {
+ uri,
+ mimeType: 'application/json',
+ text: JSON.stringify(metadata, null, 2),
+ },
+ ],
+ };
+ }
+
+ throw new Error(`Unknown resource URI: ${uri}`);
+ } catch (error) {
+ const errorMessage = error instanceof Error ? error.message : 'Unknown
error occurred';
+ throw new Error(`Failed to read resource: ${errorMessage}`);
+ }
+}
+
+// Tool handlers
+async function handleExecuteSqlQuery(request: CallToolRequest):
Promise<CallToolResult> {
+ const { query, context } = request.params.arguments as { query: string;
context?: any };
+ const result = await druidClient.executeSqlQuery(query, context);
+
+ return {
+ content: [
+ {
+ type: 'text',
+ text: JSON.stringify(result, null, 2),
+ },
+ ],
+ };
+}
+
+async function handleListDatasources(): Promise<CallToolResult> {
+ const datasources = await druidClient.getDatasources();
+
+ return {
+ content: [
+ {
+ type: 'text',
+ text: JSON.stringify(datasources, null, 2),
+ },
+ ],
+ };
+}
+
+async function handleGetDatasourceMetadata(request: CallToolRequest):
Promise<CallToolResult> {
+ const { datasource } = request.params.arguments as { datasource: string };
+ const metadata = await druidClient.getDatasourceMetadata(datasource);
+
+ return {
+ content: [
+ {
+ type: 'text',
+ text: JSON.stringify(metadata, null, 2),
+ },
+ ],
+ };
+}
+
+async function handleTestConnection(): Promise<CallToolResult> {
+ const status = await druidClient.getStatus();
+
+ return {
+ content: [
+ {
+ type: 'text',
+ text: `✅ Connection successful!\n${JSON.stringify(status, null, 2)}`,
+ },
+ ],
+ };
+}
+
+/**
+ * Main server function
+ */
+async function main(): Promise<void> {
+ // Check for help flag
+ if (process.argv.includes('--help') || process.argv.includes('-h')) {
+ showHelp();
+ process.exit(0);
+ }
+
+ console.error('Starting Apache Druid MCP Server...');
Review Comment:
should not use `error`, or the
[mcp-inspector](https://www.claudemcp.com/inspector) will recoginize this print
as error
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org
For additional commands, e-mail: commits-h...@druid.apache.org
