How harsh-p-shukla/phrs works
Project type and purpose
• This is a full-stack, Flask-based web application that serves as a Personalized Medical Recommendation System.
• It exposes a one-screen dashboard where users enter symptoms; the backend runs ML inference and returns predictions plus disease descriptions, precautions, diet plans, workout routines, and medications.
File structure
• Top level:
• main.Harsh.py – entry point for the Flask app
• model/ – pretrained pickled models (rf.pkl, svc.pkl)
• static/ – CSS (styles.css) and upload folder for user-submitted files
• templates/ – Jinja2 HTML templates (index, about, blog, contact, developer)
• dataset/ – source CSVs and chest X-ray image folders used during model training
• notebooks & archive/ – exploratory and historical data-science artifacts (not part of runtime)
• venv/ – isolated Python environment
• Patterns & clues:
• MVC style: templates as Views, main.Harsh.py routes as Controllers, data-loading & ML modules as Models
• No external database—data is read from CSVs at runtime
• Configuration files (.gitignore) and local environment (venv) imply simple file-system deployment
README for additional insights
• Confirms Flask API backend, use of scikit-learn, pandas, numpy, textblob for text correction, and pickle for model serialization
• Describes core features: synonym dictionary, auto-spelling correction, and ML predictions feeding comprehensive recommendations