🧾 Patient Health Report Analyzer (v2.0)

Prototype rebuilt with Mistral LLM, LangChain, and a modern UI.

This is an enhanced and efficient version of the Patient Health Report Analyzer project I initially created as a prototype. The earlier version focused on basic PDF parsing and keyword matching. In this release, I've rebuilt it using cutting-edge technologies like:

• 🧠 Mistral LLM (via Ollama) – for local, privacy-preserving AI inference
• 🔗 LangChain – for prompt management and chaining logic
• 📄 PDF Parsing – clean extraction of health report text using pdfplumber
• 💻 Streamlit UI – for an intuitive and interactive frontend

---

🧠 Features

• Upload and analyze health/lab PDF reports
• Extracts and summarizes:
  • 🔍 Key medical observations
  • ⚠️ Critical abnormalities or conditions
  • 📋 Suggested follow-ups or tests
  • ❓ Missing or ambiguous information
• All AI processing done locally via Mistral (no cloud model dependency)

---

🚀 Getting Started

1. Clone the repository:

   git clone https://github.com/rahulprajapati08/PHR-Analyzer.git
   cd PHR-Analyzer

2. Install dependencies:

   pip install -r requirements.txt

3. Install Ollama if not already installed, then run:

   ollama run mistral

4. Launch the App:

   streamlit run app.py
   

---

🖼️ Screenshots

---

📌 Notes

• This project is fully local and private — no external LLM APIs are used.
• You can deploy the Streamlit app online and expose your backend using ngrok or localtunnel.

---

🛠️ Future Improvements

• Add PDF export of summary
• Build patient history dashboard
• Integrate OCR for scanned reports