📱 Tetro-Stack

Cross-Platform iPhone Marketplace Intelligence

A Python client to scrape, parse, and analyze iPhone listings from multiple marketplaces (OLX India, Facebook Marketplace) with AI-powered data extraction.

✨ Features

• Multi-Platform Scraping: OLX India + Facebook Marketplace
• AI-Powered Parsing: Gemini 2.0 Flash extracts iPhone specs from unstructured text
• Structured Data Extraction:
  • Series, Model, Storage, Color
  • Battery Health, Warranty info
  • Seller details, Price, Location
• Smart Filtering: 1-day, 3-day, 5-day, 7-day inventory age filters
• Export: CSV and JSON formats
• Cities Supported: Chennai, Bangalore, Coimbatore

🚀 Quick Start

1. Clone & Install

git clone https://github.com/Vi-Sri/Tetro-Stack.git
cd Tetro-Stack
pip install -r requirements.txt

2. Configure Environment

Create env.sh with your API keys:

#!/bin/bash
export APIFY_API_TOKEN="your_apify_token_here"
export GEMINI_API_KEY="your_gemini_api_key_here"

Then source it:

source env.sh

3. Run the Scraper

# Full cross-platform scrape (OLX + Facebook)
python test_ensemble.py

# Token validation & cached data (no credits used)
python test_integration.py

# Test AI parser only
python test_gemini_parser.py

🔑 API Keys Setup

Apify (Required)

1. Sign up at apify.com
2. Go to Settings → Integrations
3. Copy your API token
4. Add payment method for paid actors

Actors Used:

• natanielsantos/olx-india-scraper (OLX India)
• curious_coder/facebook-marketplace (Facebook Marketplace)

Gemini AI (Optional, enhances parsing)

1. Go to Google AI Studio
2. Create API key
3. Model used: gemini-2.0-flash

📁 Project Structure

tetro_stack/
├── clients/
│   ├── apify_client.py     # Apify scraper integration
│   ├── platforms.py        # Platform configs (OLX, FB)
│   └── base_client.py      # Abstract base client
├── models/
│   └── iphone_listing.py   # Pydantic data models
├── parsers/
│   └── iphone_parser.py    # Gemini AI + regex parsing
├── filters/
│   └── inventory_filter.py # Age-based filtering
├── exporters/
│   └── data_exporter.py    # CSV/JSON export
└── config.py               # Configuration

📊 Sample Output

┏━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Model             ┃ Storage ┃ Price    ┃ City          ┃
┡━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ iPhone 14 Pro Max │ 128GB   │ ₹54,500  │ Bengaluru, KA │
│ iPhone 13         │ 128GB   │ ₹27,000  │ Chennai, TN   │
│ iPhone 12 Pro     │ 256GB   │ ₹26,000  │ Bengaluru, KA │
│ iPhone 11         │ 64GB    │ ₹14,500  │ Coimbatore    │
└───────────────────┴─────────┴──────────┴───────────────┘

📈 Data Fields

Field	Description
series	iPhone generation (11, 12, 13, 14, 15, 16)
model	Specific model (Pro, Pro Max, Plus, Mini)
storage	Capacity (64GB, 128GB, 256GB, 512GB, 1TB)
color	Device color
battery_health	Battery percentage if available
warranty	Warranty status and type
price	Listed price in INR
seller.city	Seller location
days_active	Days since listing was posted
source	Platform (olx_india, facebook_marketplace)

⚡ Performance

• ~240 listings per run (100 OLX + 140 Facebook)
• ~55 seconds total scrape time
• 3 cities searched simultaneously
• AI parsing with regex fallback

📝 Notes

• Listings older than 5 days are filtered out (iPhones move fast!)
• Facebook Marketplace doesn't expose seller names in search results
• Gemini free tier has rate limits; falls back to regex when exceeded