How dhanush-r-git/mh-analysis works
Project type and purpose
• Maṉa is a full-stack web application for mental‐health analysis via social media.
• It ingests user comments or API-fetched posts, runs sentiment analysis (MHRoberta/RoBERTa), then branches into two chat modes:
• MaṉaNow (DeepSeek-R1) for in-depth assessments and report generation
File structure
• Top level: README, requirement.txt, research notebooks, mental_health_docs (static docs)
• static/assets: front-end static files (CSS/JS/images)
• webapp_setup/
• templates/: Jinja2 HTML views (Landing, Overview, MaṉaChat, MaṉaNow, etc.)
• *.py modules:
– chatbot.py: entry point / FastAPI or Flask server
– Manachat.py: orchestrates MaṉaChat mode
– Mananow.py: orchestrates MaṉaNow mode
– web_scrapper.py: collects comments via social-media APIs
– error_logger.py: logs errors/reports
• No explicit database folder, so reports may be written to files or external store
README for insights
• Confirms FastAPI-powered chatbot backend
• Uses Hugging Face Inference API or local fallback models