ML-Assessment for AI-Engineer position

[damtaba]

Summary

This PR implements a FastAPI-based text summarization service that generates summaries and actionable CTAs using OpenAI's LLM. The implementation follows clean architecture principles with ports/adapters pattern and includes comprehensive error handling and documentation.

Key Features Implemented

• ✅ Single Summary Generation - Synchronous endpoint for processing individual texts
• ✅ Summary Retrieval - Fetch previously generated summaries by ID
• ✅ Batch Async Processing - Concurrent processing of multiple texts with individual error handling
• ✅ In-memory Storage - Repository pattern for storing generated summaries
• ✅ Error Handling - Proper HTTP status codes and OpenAI error handling
• ✅ API Documentation - Complete API documentation with examples (README_API.md)

Technical Stack

• Python 3.12
• FastAPI with OpenAPI documentation
• OpenAI API integration
• Poetry for dependency management
• Pytest with coverage
• Pre-commit hooks for code quality

Architecture Highlights

• Ports & Adapters Pattern - Clean separation between domain logic and infrastructure
• Dependency Injection - Using FastAPI's DI system
• Type Safety - Pydantic models for request/response validation
• Async Support - Batch processing with asyncio.gather for optimal performance

API Endpoints

1. GET /summary_maker/get_summary_and_ctas - Generate summary and CTAs
2. GET /summary_maker/get_summary_and_ctas_by_id - Retrieve by ID
3. POST /summary_maker/async_get_summary_and_ctas - Batch async processing
4. GET /extras/get_summaries_ids - List all summary IDs

Documentation

• Interactive API docs at /docs (Swagger UI)
• Comprehensive README_API.md with setup instructions and examples

Notes

• GET vs POST: The first endpoint uses GET as specified in requirements, though POST would be more semantically appropriate for content generation
• Storage: Currently uses in-memory storage; Redis caching suggested for production
• Pre-commit: Added pre-commit hooks for code quality enforcement