AI Lottery Number Prediction Bot

I want a fully-functional bot that forecasts lottery numbers by blending three pillars: historical draw data, pattern-recognition AI, and a controlled layer of RNG for fresh variation. The system should read large archives of past winning combinations, learn frequency trends, hot–cold sequences, gap analysis, and similar statistical markers, then generate the next suggested pick set in the correct format for any lottery I point it to.

Key requirements
• Core engine in a mainstream language (Python preferred) with clear, well-commented code
• Automatic ingestion or easy CSV upload of historical results so I can retrain whenever new draws occur
• Configurable rules (number pool size, pick count, bonus balls, etc.) so the same model can switch from a national 6/49 to a smaller state game without rewriting code
• Option to adjust the weight given to pure RNG versus learned patterns so I can experiment with different prediction styles
• Output ready to display in a simple local GUI or a lightweight web dashboard, plus a JSON endpoint for future expansion
• Brief documentation walking through setup, retraining, and parameter tuning, plus a short report explaining the modelling approach and performance metrics from back-tests

Deliver the source code, requirements.txt, and a demo dataset so I can test immediately. I’m open to libraries like pandas, scikit-learn, TensorFlow, or PyTorch—use whatever best serves speed and clarity.