newellp88/V20pyPro ? reverse-engineered prompt

Build me a Python project for forex trading research that can take about 500 days of historical price data for a currency pair, add common technical indicators, and run a few different prediction models on it.

I want it to include examples for simple linear models, ARIMA, a small ensemble comparison, and a basic Keras or TensorFlow neural network, then show the results clearly with charts and saved output so I can compare how each approach performs. It should also have a simple way to store and load research data from a local database, and it would be great if the code is organized so I can run the research scripts separately and swap in different pairs later.

Please make it easy to understand and tweak, with sensible defaults, readable code, and a basic README that explains how to run the examples. If you need to look up current library docs online while building it, go ahead.