/**
* Knowledge base and analytics tool definitions
* Extracted from main server file for better organization
*/
export const knowledgeTools = [
// New tool: semantic query for stored cluster data
{
name: 'query_cluster_data',
description:
'Query stored cluster data using natural language (e.g., "pip packages", "machine types", "network config")',
inputSchema: {
type: 'object',
properties: {
projectId: { type: 'string', description: 'Optional: Google Cloud Project ID' },
region: { type: 'string', description: 'Optional: Google Cloud region' },
query: {
type: 'string',
description:
'Natural language query (e.g., "pip packages", "machine configuration", "network settings")',
},
clusterName: {
type: 'string',
description: 'Filter by specific cluster name (optional)',
},
limit: {
type: 'number',
description: 'Maximum number of results (default: 5)',
},
},
required: ['query'],
},
},
// New tool: get cluster discovery insights
{
name: 'get_cluster_insights',
description:
'📊 Get comprehensive insights about discovered clusters, machine types, components, and recent discoveries',
inputSchema: {
type: 'object',
properties: {
projectId: { type: 'string', description: 'Optional: Google Cloud Project ID' },
region: { type: 'string', description: 'Optional: Google Cloud region' },
},
required: [],
},
},
// New tool: get job analytics
{
name: 'get_job_analytics',
description:
'📈 Get analytics about job submissions, success rates, error patterns, and performance metrics',
inputSchema: {
type: 'object',
properties: {
projectId: { type: 'string', description: 'Optional: Google Cloud Project ID' },
region: { type: 'string', description: 'Optional: Google Cloud region' },
},
required: [],
},
},
// New tool: query knowledge base
{
name: 'query_knowledge',
description:
'🧠 Query the comprehensive knowledge base using natural language (clusters, jobs, errors, all)\n\n' +
'**Enhanced Features:**\n' +
'• Structured data retrieval with schema/rows separation for query results\n' +
'• Raw Qdrant document access with compression status\n' +
'• Proper clusterName display (no more "unknown" values)\n' +
'• Decompression support for compressed data\n' +
'• Detailed metadata including compression ratios and token counts\n' +
'• Tag-based search for exact field matching (solves UUID search problem)\n\n' +
'**Search Types:**\n' +
'• **Semantic Search**: `"hive query results"` - Natural language queries\n' +
'• **Tag-based Search**: `"jobId:20d1092f-9aa8-4f4d-b4e3-bfbdbdd8d431"` - Exact field matching\n' +
'• **Hybrid Search**: `"jobId:12345 SHOW DATABASES"` - Tags + semantic content\n\n' +
'**Supported Tags:**\n' +
'• `jobId:value` - Search by job ID\n' +
'• `clusterName:value` - Search by cluster name\n' +
'• `projectId:value` - Search by project ID\n' +
'• `region:value` - Search by region\n' +
'• `toolName:value` - Search by tool name\n' +
'• `type:value` - Search by data type\n\n' +
'**🔍 RESULT DISCOVERY EXAMPLES:**\n' +
'• `jobId:abc123-def456 contentType:query_results` - Get actual Hive/Spark results\n' +
'• `jobId:xyz789 type:query_result includeRawDocument:true` - Complete job output\n' +
'• `clusterName:analytics-prod contentType:query_results` - All results from cluster\n' +
'• `projectId:data-warehouse contentType:query_results` - Project query outputs\n\n' +
'**🎯 COMMON PATTERNS:**\n' +
'• Find table counts: `jobId:YOUR_ID contentType:query_results` → ["42857"]\n' +
'• Get schema info: `DESCRIBE TABLE contentType:query_results` → column details\n' +
'• Check job status: `jobId:YOUR_ID type:job` → execution metadata\n' +
'• Cluster analysis: `clusterName:spark-cluster machine types` → hardware info\n\n' +
'**🚀 POWER USER TIPS:**\n' +
'• Combine tags: `jobId:123 clusterName:prod contentType:query_results`\n' +
'• Time-based: `toolName:submit_hive_query 2024-12 contentType:query_results`\n' +
'• Error hunting: `ERROR contentType:query_results` → failed query outputs\n' +
'• Performance: `EXPLAIN contentType:query_results` → query plans',
inputSchema: {
type: 'object',
properties: {
projectId: { type: 'string', description: 'Optional: Google Cloud Project ID' },
region: { type: 'string', description: 'Optional: Google Cloud region' },
query: {
type: 'string',
description: 'Natural language query about clusters, jobs, or errors',
},
type: {
type: 'string',
enum: ['clusters', 'cluster', 'jobs', 'job', 'errors', 'error', 'all'],
description:
'Type of knowledge to search (default: all). Supports both singular and plural forms.',
},
limit: {
type: 'number',
description: 'Maximum number of results (default: 10)',
},
includeRawDocument: {
type: 'boolean',
description:
'Include raw Qdrant document with full payload, compression status, and metadata (default: false)',
},
},
required: ['query'],
},
},
];
