/**
* Job submission and management tool definitions
* Extracted from main server file for better organization
*/
export const jobTools = [
// New tool: submit Hive query
{
name: 'submit_hive_query',
description:
'Submit a Hive query to a Dataproc cluster with enhanced result discovery.\n\n' +
'**🚀 QUICK START EXAMPLES:**\n' +
'• `SHOW DATABASES` - List all databases\n' +
'• `SELECT COUNT(*) FROM my_table` - Get row count\n' +
'• `DESCRIBE my_table` - Show table schema\n' +
'• `SHOW TABLES IN my_database` - List tables in database\n\n' +
'**📊 RESULT ACCESS:**\n' +
'After submission, use query_knowledge with jobId to get actual results:\n' +
'`query_knowledge("jobId:YOUR_JOB_ID contentType:query_results")`\n\n' +
'**⚡ ASYNC MODE:**\n' +
'Set async:true for long-running queries, then monitor with get_job_status',
inputSchema: {
type: 'object',
properties: {
clusterName: { type: 'string', description: 'Name of the cluster to run the query on' },
query: { type: 'string', description: 'Hive query to execute' },
async: {
type: 'boolean',
description:
'Optional: Whether to wait for query completion (false) or return immediately (true)',
},
verbose: {
type: 'boolean',
description: 'Optional: Return full response without filtering (default: false)',
},
queryOptions: {
type: 'object',
description: 'Optional: Query configuration options',
properties: {
timeoutMs: { type: 'number', description: 'Optional: Timeout in milliseconds' },
parameters: { type: 'object', description: 'Optional: Query parameters' },
properties: { type: 'object', description: 'Optional: Query properties' },
},
},
},
required: ['clusterName', 'query'],
},
},
// New tool: get query status
{
name: 'get_query_status',
description: 'Get the status of a Hive query job',
inputSchema: {
type: 'object',
properties: {
jobId: { type: 'string', description: 'Job ID to check' },
},
required: ['jobId'],
},
},
// Enhanced tool: get query results with async support and semantic search
{
name: 'get_query_results',
description:
'Get the results of a completed Hive query with enhanced async support and semantic search integration',
inputSchema: {
type: 'object',
properties: {
jobId: { type: 'string', description: 'Job ID to get results for' },
maxResults: {
type: 'number',
description: 'Optional: Maximum number of rows to display in the response (default: 10)',
},
pageToken: { type: 'string', description: 'Optional: Page token for pagination' },
},
required: ['jobId'],
},
},
// New tool: submit Dataproc job (generic)
{
name: 'submit_dataproc_job',
description:
'Submit a Dataproc job (Hive, Spark, PySpark, Presto, etc.) to a cluster with enhanced monitoring.\n\n' +
'**🔧 SUPPORTED JOB TYPES:**\n' +
'• **hive** - SQL queries on Hadoop data\n' +
'• **spark** - Scala/Java Spark applications\n' +
'• **pyspark** - Python Spark jobs\n' +
'• **presto** - Fast SQL analytics\n' +
'• **hadoop** - MapReduce jobs\n\n' +
'**📝 JOB CONFIG EXAMPLES:**\n' +
'• Hive: `{"query": "SELECT COUNT(*) FROM table"}`\n' +
'• PySpark: `{"mainPythonFileUri": "{@./test-spark-job.py}", "args": ["arg1"]}`\n' +
'• Spark: `{"mainClass": "com.example.Main", "jarFileUris": ["{@./app.jar}"]}`\n\n' +
'**🔧 LOCAL FILE STAGING:**\n' +
'• Use `{@./relative/path}` for files relative to config directory\n' +
'• Use `{@/absolute/path}` for absolute file paths\n' +
'• Files are automatically staged to GCS and cleaned up after job completion\n' +
'• Supports .py, .jar, .sql, .R file extensions\n\n' +
'**🎯 RESULT WORKFLOW:**\n' +
'1. Submit job → Get jobId\n' +
'2. Monitor: get_job_status(jobId)\n' +
'3. Results: query_knowledge("jobId:YOUR_ID contentType:query_results")',
inputSchema: {
type: 'object',
properties: {
clusterName: { type: 'string', description: 'Name of the cluster to run the job on' },
jobType: {
type: 'string',
description: 'Type of job (hive, spark, pyspark, presto, etc.)',
},
jobConfig: { type: 'object', description: 'Job configuration object (type-specific)' },
async: {
type: 'boolean',
description: 'Whether to submit asynchronously (default: false)',
},
},
required: ['clusterName', 'jobType', 'jobConfig'],
},
},
// New tool: get Dataproc job status
{
name: 'get_job_status',
description:
'Get the status of a Dataproc job by job ID with smart result discovery.\n\n' +
'**📊 STATUS TYPES:**\n' +
'• PENDING - Job queued for execution\n' +
'• RUNNING - Job currently executing\n' +
'• DONE - Job completed successfully ✅\n' +
'• ERROR - Job failed with errors ❌\n' +
'• CANCELLED - Job was cancelled\n\n' +
'**🎯 WHEN STATUS = DONE:**\n' +
'Automatically shows result discovery hints:\n' +
'`query_knowledge("jobId:YOUR_ID contentType:query_results")`\n\n' +
'**💡 MONITORING WORKFLOW:**\n' +
'1. Submit job (async mode)\n' +
'2. Check status periodically\n' +
'3. When DONE, get actual results via query_knowledge',
inputSchema: {
type: 'object',
properties: {
jobId: { type: 'string', description: 'Job ID to check' },
verbose: {
type: 'boolean',
description: 'Optional: Return full response without filtering (default: false)',
},
},
required: ['jobId'],
},
},
// Enhanced tool: get Dataproc job results with smart discovery hints
{
name: 'get_job_results',
description:
'Get the results of a completed Dataproc job by job ID.\n\n' +
'**🎯 FOR COMPLETE RESULTS INCLUDING ACTUAL DATA:**\n' +
'Use query_knowledge with combined tags for better results:\n' +
'• `jobId:YOUR_JOB_ID contentType:query_results` - Get actual query results\n' +
'• `jobId:YOUR_JOB_ID type:query_result` - Alternative format\n' +
'• `jobId:YOUR_JOB_ID` - Get job metadata and result hints\n\n' +
'**💡 EXAMPLE:**\n' +
'query_knowledge("jobId:89feded7-902b-4698-b076-12008a8929a7 contentType:query_results")\n' +
'→ Returns actual data: ["220144"]\n\n' +
'**Note:** This tool returns optimization metadata. For actual query output data, use the query_knowledge patterns above.',
inputSchema: {
type: 'object',
properties: {
jobId: { type: 'string', description: 'Job ID to get results for' },
maxResults: {
type: 'number',
description: 'Optional: Maximum number of rows to display in the response (default: 10)',
},
},
required: ['jobId'],
},
},
// New tool: quick status check for active jobs
{
name: 'check_active_jobs',
description:
"🚀 Quick status check for all active and recent jobs - perfect for seeing what's running!",
inputSchema: {
type: 'object',
properties: {
includeCompleted: {
type: 'boolean',
description: 'Include recently completed jobs (default: false)',
},
},
required: [],
},
},
// New tool: cancel Dataproc job
{
name: 'cancel_dataproc_job',
description:
'Cancel a running Dataproc job with intelligent status handling and job tracking integration.\n\n' +
'**🛑 CANCELLATION WORKFLOW:**\n' +
'• Attempts to cancel jobs in PENDING or RUNNING states\n' +
'• Provides informative messages for jobs already in terminal states\n' +
'• Updates internal job tracking when cancellation succeeds\n\n' +
'**📊 STATUS HANDLING:**\n' +
'• PENDING/RUNNING → Cancellation attempted\n' +
'• DONE/ERROR/CANCELLED → Informative message returned\n' +
'• Job not found → Clear error message\n\n' +
'**💡 MONITORING:**\n' +
'After cancellation, use get_job_status("jobId") to confirm the job reaches CANCELLED state.',
inputSchema: {
type: 'object',
properties: {
jobId: { type: 'string', description: 'The ID of the Dataproc job to cancel' },
projectId: {
type: 'string',
description: 'Optional: Google Cloud Project ID (uses defaults)',
},
region: { type: 'string', description: 'Optional: Google Cloud region (uses defaults)' },
verbose: {
type: 'boolean',
description: 'Optional: Return full response without filtering (default: false)',
},
},
required: ['jobId'],
},
},
];
