Python Medical Recommendation Engine -- 2

I have a raw set of patient records and want to turn it into a functional medicine-recommendation engine built in Python. The system should read symptom text typed by a user, run it through a pre-trained machine-learning model, and return a ranked list of suitable medications.

Scope and key tasks
• Data pipeline: remove duplicate records and normalise numeric or categorical values so every observation is model-ready.
• Feature selection & training: experiment with at least Random Forest and Logistic Regression; pick the best model using accuracy, precision and recall.
• Evaluation report: include confusion matrices and a short note explaining why the chosen model performs best.
• Lightweight dashboard: a simple Flask or Streamlit interface with one free-text field for symptoms and a results pane listing the suggested medicines.
• Clean, well-commented code plus a README describing setup, usage and how to extend the model with medical history or allergy data later.

Everything should run end-to-end on my local machine with clear install instructions (virtualenv or conda).