Project Overview

A robust and efficient automated downloader that fetches tick-level historical data (bid & ask) from Dukascopy’s public JForex feed and stores it in highly optimized Parquet files.

Ideal for quantitative traders, backtesters, and anyone needing clean, reliable tick data.

Key Features

• High-precision tick data (bid & ask)
• Incremental updates — only downloads missing data
• Hourly Parquet storage (excellent compression & query speed)
• Configurable retention (keeps last N days fresh)
• Smart scheduling with intelligent sleep logic
• Full Docker + docker-compose support
• Centralized logging system

Technical Highlights

• Built with Python + Pandas
• Efficient Parquet format for large time-series data
• Robust error handling and retry logic
• Supports multiple major forex pairs and metals
• Production-ready with Docker

This tool powers reliable data infrastructure for trading systems and algorithmic research.