News Intelligence System

An AI-powered news aggregation, classification, and summarization system built as part of the AI Engineering Intern Assignment.

Overview:

This project implements an end-to-end news intelligence pipeline that:

• Collects recent news articles (last 7 days)
• Classifies them by a predefined company set
• Generates concise 30–40 word summaries
• Stores structured results in a PostgreSQL database
• Exposes the data through a FastAPI backend The system is designed with robust fallbacks, clean modular architecture, and reproducible setup.

---

Chosen Test Set:

Global IT Companies

• Microsoft
• Google
• Apple
• Meta Only articles related to the above companies are ingested.
  All unrelated news is filtered out.

---

System Architecture:

News API ↓ News Fetcher (last 7 days) ↓ Deduplication (URL + hash) ↓ Entity Classification (alias-based) ↓ Content Scraping + Fallback ↓ Summarization (LLM / Offline fallback) ↓ PostgreSQL Database ↓ FastAPI Backend

---

Tech Stack:

• Language: Python 3.10
• Database: PostgreSQL
• ORM: SQLAlchemy
• Backend: FastAPI
• News Source: NewsAPI
• Summarization:
  • OpenAI (if you have API Key)
  • Offline extractive fallback (for robustness)
• UI: Streamlit
• Version Control: Git & GitHub

---

Project Structure:

news ├── app/ │ ├── main.py # Pipeline orchestrator │ ├── api/ # FastAPI routes │ ├── config/ # Settings & company definitions │ ├── db/ # Database models & CRUD │ ├── services/ # Fetching, tagging, summarization │ └── ui/ # Streamlit UI ├── scripts/ │ └── export_to_csv.py # CSV export utility ├── requirements.txt ├── .env.example └── README.md

How To Run:

Install Dependencies- pip install -r requirements.txt Run the Ingestion Pipeline- python -m app.main Start the API Server- python -m uvicorn app.api.routes:app --reload Open API Documentation- http://127.0.0.1:8000/docs