M&A x402 Service

A FastAPI service for Coinbase AgentKit that scrapes PR Newswire articles and extracts M&A (Mergers & Acquisitions) entities using OpenAI's GPT models.

Features

• 🔍 Scrapes PR Newswire article links from a given URL
• 📄 Extracts full article text content
• 🤖 Uses OpenAI API to extract structured M&A entities:
  • BUYER: Acquiring companies/entities
  • SELLER: Companies being sold/divested
  • FUND: Investment funds, PE, VC firms
  • LAW_FIRM: Legal firms involved
  • INTERMEDIARY: Investment banks, advisors
  • PROFESSIONAL: Individual professionals
  • MONEY: Deal values and financial figures
  • DATE: Transaction dates
  • DEAL_TYPE: Type of transaction

Tech Stack

• Python 3.11
• FastAPI: Modern web framework
• Uvicorn: ASGI server
• BeautifulSoup4: HTML parsing
• Newspaper3k: Article extraction
• OpenAI API: Entity extraction
• Pydantic: Data validation

Setup

1. Clone and Install Dependencies

cd mna-x402-service
pip install -r requirements.txt

2. Environment Variables

Copy .env.example to .env and add your OpenAI API key:

$cp .env.example .env

Edit .env:

OPENAI_API_KEY=sk-your-actual-key-here

3. Run Locally

$uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

The API will be available at http://localhost:8000

API Endpoints

Health Check

GET /
GET /health

Scrape Articles

POST /x402/scrape
Content-Type: application/json
{
  "site_url": "https://www.prnewswire.com/news-releases/financial-services-latest-news/acquisitions-mergers-and-takeovers-list/",
  "max_articles": 10
}

Parameters:

• site_url (required): URL of the PR Newswire page to scrape
• max_articles (optional): Maximum number of newest articles to process. Defaults to 10, maximum 100. Returns the most recently posted articles first. Helps prevent timeouts and control processing time.

Response:

{
  "site_url": "https://...",
  "articles": [
    {
      "url": "https://...",
      "content": "Full article text...",
      "entities": {
        "buyer": ["Company A"],
        "seller": ["Company B"],
        "fund": ["PE Fund XYZ"],
        "law_firm": ["Law Firm ABC"],
        "intermediary": ["Investment Bank DEF"],
        "professional": ["John Doe"],
        "money": ["$100M", "$50 million"],
        "date": ["2024-01-15", "Q1 2024"],
        "deal_type": "Acquisition"
      }
    }
  ],
  "count": 10,
  "total_found": 25,
  "processed": 10,
  "limit": 10
}

Response Fields:

• site_url: The URL that was scraped
• articles: Array of article results with extracted entities
• count: Number of articles in the response
• total_found: Total number of articles found on the page
• processed: Number of articles actually processed
• limit: The limit that was applied (max_articles parameter)

## Docker

### Build
```bash
docker build -t mna-x402-service .

Run

$docker run -p 8000:8000 --env-file .env mna-x402-service

API Documentation

Once running, visit:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Project Structure

mna-x402-service/
├── app/
│   ├── __init__.py
│   ├── main.py          # FastAPI application
│   ├── models.py        # Pydantic models
│   ├── scraper.py       # PR Newswire scraping
│   └── extractor.py     # OpenAI entity extraction
├── requirements.txt
├── Dockerfile
├── .env.example
└── README.md

Error Handling

The service includes comprehensive error handling:

• Network errors during scraping
• Article extraction failures
• OpenAI API errors
• JSON parsing errors

Failed articles are included in the response with an error field.

Production Considerations

• Add rate limiting
• Implement caching for repeated requests
• Add authentication/API keys
• Set up monitoring and logging
• Configure proper CORS origins
• Use environment-specific configurations

License

MIT