tibkiss/huba-v1 — reverse-engineered prompt

Build me a Python learning project for pairs trading using statistical arbitrage. I want to load minute stock price data, clean it up, filter out stocks that are too illiquid or not tradeable, then search within equity sectors for pairs that historically move together.

The app should let me run a big pair scan on old data, pick promising pairs, test them on separate data, and then run a simple strategy that goes long one stock and short the other when the spread gets too far from its usual value. It should track trades, equity, profit and loss, and write useful logs so I can review what happened later.

Please make it practical rather than fancy. Use PyAlgoTrade if that still makes sense, and look up current docs online if needed. Include example config for traded pairs, backtesting, paper trading style runs, and enough comments so someone studying quantitative trading can understand the approach and its limitations.