📘 Trade Opportunities API 📌 Overview
This project implements a FastAPI service that analyzes market data for a given sector in India and generates a Markdown report of trade opportunities using Google Gemini AI.
It was built as part of the Trade Opportunities API Developer Task.
⚙️ Features
✅ FastAPI backend with async support
✅ Single endpoint: GET /analyze/{sector}
✅ Collects recent market data from DuckDuckGo-Search
✅ Uses Google Gemini API to analyze and generate structured Markdown reports
✅ Built-in rate limiting (5 requests per minute per IP)
✅ Clean error handling (502 for data issues, 503 for AI issues)
✅ Returns Markdown report wrapped in JSON
📂 Project Structure trade_api/ │── main.py # FastAPI app (endpoint + rate limiting + data fetch) │── requirements.txt # Dependencies │── README.md # This file
🚀 Setup & Run
-
Clone & Install git clone cd trade_api pip install -r requirements.txt
-
Set Gemini API Key
You need a Google Gemini API key (from Google AI Studio or Google Cloud). Set it as an environment variable:
export GEMINI_API_KEY="your_api_key_here"
(Windows PowerShell)
$env:GEMINI_API_KEY="your_api_key_here"
- Run FastAPI uvicorn main:app --reload --host 127.0.0.1 --port 8000
📡 API Usage Endpoint GET /analyze/{sector}
Example Request GET http://127.0.0.1:8000/analyze/pharmaceuticals
Example Response { "report": "# Market Analysis — Pharmaceuticals\n\n- Key Trends...\n- Opportunities...\n- Risks...\n- Future Outlook...\n\nGenerated by AI" }
🔒 Security
Open access (no authentication required).
Rate limiting: 5 requests/minute per IP to prevent abuse.
502 → Market data fetch failed
503 → AI service unavailable
✅ Success Criteria
API accepts sector names (e.g., pharmaceuticals, technology, agriculture)
Returns structured Markdown report containing trade opportunities
Implements rate limiting and proper error handling
Clean, well-organized code with separation of concerns
📝 Notes
The data collection currently uses DuckDuckGo API for simplicity. In a production system, this can be replaced with a more robust data source (e.g., NewsAPI, scraping, financial APIs).
Gemini model used: gemini-2.5-flash (fast & lightweight).
✨ You’re all set! Run the server, test in Postman, and save the returned Markdown as .md files for reports.