precisionBOM/python-agent at main · precision-bom/precisionBOM · GitHub
Skip to content

python-agent

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

python-agent

README.md

PrecisionBOM Python Agent

Multi-agent BOM processing service built with CrewAI and FastAPI.

Production URL: https://api.precisionbom.com

Overview

This service automates the review of electronic component sourcing through specialized AI agents:

  • Engineering Agent - Evaluates technical specifications and compliance
  • Sourcing Agent - Analyzes supply chain risk and availability
  • Finance Agent - Reviews pricing and budget constraints

Tech Stack

  • Framework: FastAPI + Uvicorn
  • AI: CrewAI with Anthropic Claude
  • CLI: Click + Rich
  • Data: SQLite stores + CrewAI memory

Getting Started

Prerequisites

  • Python 3.11+
  • uv package manager

Installation

uv sync

Environment Variables

Copy .env.example to .env and configure:

# LLM API Key (at least one required)
OPENAI_API_KEY=sk-...        # Required for gpt-5-nano (default)
ANTHROPIC_API_KEY=sk-ant-... # Required for Claude models

# Optional
CREWAI_MODEL=gpt-5-nano      # Model for agents
PORT=8000                     # API server port

Start the Server

uv run uvicorn bom_agent_service.main:app --reload

Server runs at http://localhost:8000.

CLI Usage

Process a BOM

uv run sourcing process sample_bom.csv --intake project_intake.yaml

Check Status

uv run sourcing status

View Trace

uv run sourcing trace <project_id>

Knowledge Base

uv run sourcing kb suppliers list

Project Structure

python-agent/
├── src/bom_agent_service/
│   ├── main.py           # FastAPI entry point
│   ├── cli.py            # Click CLI
│   ├── agents/           # CrewAI agent definitions
│   ├── flows/            # CrewAI Flow orchestration
│   ├── stores/           # SQLite data stores
│   ├── api/              # FastAPI routers
│   └── models/           # Pydantic models
├── tests/                # pytest tests
├── data/                 # Sample data files
├── demo/                 # Demo scripts
└── docs/                 # Documentation

API Endpoints

Endpoint Method Description
/health GET Health check
/projects GET List all projects
/projects POST Create new project
/projects/{id} GET Get project details
/knowledge/suppliers GET List suppliers
/process POST Process BOM file

Architecture

CrewAI Flows

Orchestration uses @start() and @listen() decorators for sequential agent execution:

@start()
def ingest_bom(self):
    # Parse BOM file
    pass

@listen(ingest_bom)
def engineering_review(self):
    # Technical evaluation
    pass

@listen(engineering_review)
def sourcing_review(self):
    # Supply chain analysis
    pass

Data Flow

CLI → HTTP → FastAPI → CrewAI Flow → Agents → Stores

Testing

# Run all tests
uv run pytest

# Verify imports
uv run python -c "from bom_agent_service.main import app; print('OK')"

# Test endpoints
curl http://localhost:8000/health

Scripts

Command Description
sourcing Main CLI
sourcing-server Start API server
sourcing-demo Run demo