AI4Finance-Foundation/FinRL ? reverse-engineered prompt

I want to build a Python project that uses AI for algorithmic stock trading. The core idea is to use deep reinforcement learning to train a trading bot.

The system should be able to pull historical data for a list of stocks, like the DOW 30, from a public source like Yahoo Finance. I'd like it to process this data and also calculate common technical indicators like MACD and RSI to use as features. Then, I want to train the reinforcement learning agent on a historical dataset, letting it learn a profitable trading strategy through trial and error.

After the agent is trained, I need to be able to run a backtest on a separate period of data to evaluate its performance. The output should show me how the strategy would have performed, including the final portfolio value. Please use popular libraries like PyTorch or TensorFlow for the AI models. You can look up the FinRL framework online as a guide for the overall train-test-trade pipeline.