# BOM-MCP
Electric Submersible Pump (ESP) Parts and Bill of Materials (BOM) database with MCP server for AI assistants.
## Features
- SQLite database with ESP parts, assemblies, and units
- MCP server (SSE transport) for AI assistants
- Command-line interface
- REST API server (optional)
## Installation
```bash
pip install -r requirements.txt
```
## Usage
### MCP Server (SSE transport - default)
```bash
python run_mcp.py --port 8080
```
### REST API (optional)
```bash
python api.py [port] # Default port: 5000
```
### CLI
```bash
python cli.py esp list
python cli.py esp get ESP-001
python cli.py esp bom ESP-001
python cli.py parts get ESP-MTR-001
python cli.py assemblies get ASM-MTR-001
```
## MCP Tools
| Tool | Description |
|------|-------------|
| `list_esps` | List all ESP pump models |
| `get_esp` | Get complete ESP with assemblies and parts |
| `get_esp_bom` | Get flat BOM parts list for an ESP |
| `get_bom_summary` | Get BOM summary (weight, counts, critical) |
| `get_parts_by_category` | Filter parts by category |
| `get_critical_parts` | List all critical parts |
| `get_assembly` | Get assembly with its parts |
| And more... |
## Data Model
- **Parts**: Individual components (part_number, name, category, material, weight_kg, is_critical, uom)
- **Assemblies**: Groups of parts
- **ESP Units**: Top-level pumps linking assemblies
## Docker
### Local Development
```bash
# Build and run MCP server on port 8080
docker-compose up --build
# MCP server available at http://localhost:8080
```
### Google Cloud Run Deployment
```bash
# Deploy directly from source
gcloud run deploy bom-mcp \
--source . \
--platform managed \
--region us-central1 \
--allow-unauthenticated
```
The MCP server will be available at your Cloud Run URL.
## License
MIT
