<<<<<<< HEAD

A Python tool that extracts table metadata from Metabase API and generates useful KPIs using OpenAI's LLM.

• 🔐 Metabase Authentication: Secure API access with session management
• 📊 Table Metadata Extraction: Comprehensive table and column information
• 🔗 Related Table Discovery: Automatically finds and includes related table information
• 🤖 AI-Powered KPI Generation: Uses OpenAI GPT-4 to generate business-relevant KPIs
• 📝 SQL Query Generation: Provides ready-to-use SQL queries for each KPI
• 💾 JSON Output: Structured output for easy integration with other tools

• Python 3.8+
• Metabase instance with API access
• OpenAI API key
• Network access to both Metabase and OpenAI APIs

1. Clone or download the project files

2. Install dependencies:

   pip install -r config/requirements.txt

3. Set up environment variables: Create a .env file in the project root with:

   # Metabase Configuration
   METABASE_URL=http://your-metabase-instance.com
   METABASE_USERNAME=your_username
   METABASE_PASSWORD=your_password
   
   # OpenAI Configuration
   OPENAI_API_KEY=your_openai_api_key_here

Run the main script with default table names:

python src/kpi_extractor.py

Modify the table_names list in the main() function:

table_names = [
    "your_table_name",
    "another_table",
    "custom_table"
]

from src.kpi_extractor import MetabaseKPIExtractor

# Initialize
extractor = MetabaseKPIExtractor()

# Process specific tables
results = extractor.process_table_list(["users", "orders", "products"])

# Save results
extractor.save_results(results, "my_kpis.json")

The project includes individual test scripts for each step:

python test_step1_auth.py

Tests Metabase connection and authentication.

python test_step2_auth.py

Tests table discovery and metadata extraction.

python test_step3_llm.py

Tests OpenAI integration and KPI generation.

python test_step4_integration.py

Tests the complete end-to-end workflow.

The tool generates a structured JSON output:

{
  "table_name": {
    "table_name": "string",
    "table_id": "number",
    "metadata": {
      "table_info": {...},
      "fields": [...],
      "related_tables": [...]
    },
    "kpis": [
      {
        "kpi_name": "string",
        "description": "string",
        "business_value": "string",
        "sql_query": "string",
        "output_format": "string"
      }
    ]
  }
}

• POST /api/session - Authentication
• GET /api/database - List databases
• GET /api/database/{id}/tables - List tables in database
• GET /api/table/{id} - Get table metadata
• GET /api/table/{id}/query_metadata - Get table fields

The tool includes comprehensive error handling for:

• Authentication failures
• Network connectivity issues
• API rate limiting
• Invalid responses
• LLM generation failures

Built-in delays (2 seconds) between API calls to respect Metabase rate limits.

• Credentials are stored in environment variables
• Session tokens are managed securely
• No sensitive data is logged

1. Authentication Failed

   • Verify Metabase URL, username, and password
   • Check if Metabase instance is accessible
   • Ensure user has API access permissions

2. No Tables Found

   • Verify table names exist in your Metabase instance
   • Check database permissions
   • Ensure tables are not hidden or restricted

3. LLM Generation Failed

   • Verify OpenAI API key is valid
   • Check API quota and billing
   • Ensure network access to OpenAI API

4. Rate Limiting

   • Increase delays in the code if needed
   • Check Metabase server logs for rate limit errors

Enable verbose logging by modifying the print statements in the code.

KPI_Creation/
├── 📚 docs/                           # Documentation
│   ├── ARCHITECTURE.md                # Technical architecture
│   ├── PROJECT_SUMMARY.md             # Project overview
│   ├── WORKING_KPIS_REPORT.md         # Working KPIs analysis
│   ├── INVALID_SQLS_ANALYSIS.md       # Error analysis
│   └── CLEANUP_SUMMARY.md             # Cleanup documentation
├── 🐍 src/                            # Source code
│   ├── kpi_extractor.py               # Core KPI generation
│   └── kpi_registrar.py               # KPI registration
├── 📊 data/                           # Data files
│   └── kpis_clean.json                # Clean KPI dataset
├── ⚙️  config/                         # Configuration
│   ├── requirements.txt                # Python dependencies
│   └── env_example.txt                # Environment template
├── 📖 README.md                       # Main project documentation
└── 🗑️  archive/                       # Archived files

This project is provided as-is for educational and business use.

For issues and questions:

1. Check the troubleshooting section
2. Review the test scripts
3. Check Metabase and OpenAI documentation
4. Create an issue with detailed error information =======

Automated KPI creating using Metabase

b1479c8048f2abce619116ba88b5d51af8469ecf