Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+__pycache__/app.cpython-312.pyc filter=lfs diff=lfs merge=lfs -text

README.md

@@ -15,6 +15,8 @@ tags:
   - drug-interactions
   - medication-safety
   - hackathon
+  - pharmacogenomics
+  - clinical-decision-support
 ---
 
 # 🏥 MedGuard
@@ -23,29 +25,69 @@ tags:
 
 ## MCP 1st Birthday Hackathon Submission
 
-MedGuard is a healthcare multi-agent system that detects dangerous drug interactions, personalizes recommendations based on patient factors, and ensures clinical guideline compliance.
+MedGuard is a production-grade healthcare multi-agent system that:
+- Detects dangerous drug interactions (25+ known DDIs + CYP metabolic conflicts)
+- Personalizes recommendations based on pharmacogenomics (CYP2D6, CYP2C9, CYP2C19)
+- Applies Beers Criteria for elderly patients (AGS 2023 complete list)
+- Ensures clinical guideline compliance (AHA/ACC, ADA, ESC)
+- Optimizes medication costs with generic alternatives
 
-### 🤖 Agent Architecture
+### 🤖 Agent Architecture (5 Specialized Agents)
 
-1. **Drug Interaction Agent** - Detects harmful drug combinations using Neo4j knowledge graph
-2. **Personalization Agent** - Adjusts recommendations for age, renal function, allergies
-3. **Guideline Compliance Agent** - Validates against clinical practice guidelines
-4. **Cost Optimization Agent** - Suggests cost-effective alternatives
-5. **Explanation Agent** - Generates human-readable clinical summaries
+| Agent | Role |
+|-------|------|
+| **DrugInteractionAgentEnhanced** | DDI detection, CYP enzyme conflicts, ML prediction, PubMed enhancement |
+| **PersonalizationAgent** | Renal/hepatic adjustments, pharmacogenomics, Beers Criteria, polypharmacy |
+| **GuidelineComplianceAgent** | Evidence-based guideline checking, polypharmacy detection |
+| **CostOptimizationAgent** | Generic substitution, therapeutic alternatives, formulary optimization |
+| **ExplanationAgent** | Clinical synthesis, prioritization, patient-friendly summaries |
+
+### 🔄 LangGraph Orchestration
+
+```
+START → interaction_check → [conditional_route]
+  ├─ "critical" → human_review_node → explanation → END
+  ├─ "parallel_analysis" → [personalization, guideline, cost] → explanation → END
+  └─ "low_risk" → streamlined_analysis → explanation → END
+```
 
 ### 🛠️ Technology Stack
 
-- **MCP Server**: Python `mcp` SDK with stdio transport (10 tools, 3 resources)
-- **Orchestration**: LangGraph with conditional severity-based routing
-- **LLM**: Google Gemini 1.5 Flash
-- **Databases**: PostgreSQL, Neo4j, Qdrant (vector search)
-- **Embeddings**: BioBERT for clinical text
+| Component | Technology |
+|-----------|------------|
+| **MCP Server** | Python `mcp` SDK with stdio transport (10 tools, 3 resources) |
+| **Orchestration** | LangGraph StateGraph with conditional severity routing |
+| **LLM** | Google Gemini 1.5 Flash (production) / Rule-based (demo) |
+| **Knowledge Graph** | Neo4j for drug interaction network |
+| **Vector Search** | Qdrant + BioBERT embeddings |
+| **API** | FastAPI with HIPAA-compliant audit logging |
+| **Databases** | PostgreSQL, Redis for session management |
+
+### 🧬 Key Features
+
+- **25+ Known Drug Interactions** with mechanisms and management strategies
+- **CYP Enzyme Analysis** (2D6, 2C9, 2C19, 3A4, 1A2 substrates and inhibitors)
+- **Pharmacogenomics** (poor/intermediate/ultrarapid metabolizer phenotypes)
+- **Complete Beers Criteria** (30+ medications inappropriate in elderly)
+- **Polypharmacy Detection** (5+ and 10+ medication thresholds)
+- **Clinical Guidelines** (6 major conditions with AHA/ACC/ADA recommendations)
+
+### 📊 Demo Patients
+
+| Patient | Demonstrates |
+|---------|--------------|
+| P001 - John Smith | Warfarin + Aspirin DDI, CKD, CYP2C9 variant |
+| P002 - Maria Garcia | Serotonin syndrome risk (SSRI + Tramadol) |
+| P003 - Robert Chen | 8 medications, hyperkalemia risk, HF management |
+| P004 - Sarah Johnson | Statin + CCB interaction (CYP3A4) |
+| P005 - James Wilson | Beers Criteria violations, CYP2D6 poor metabolizer |
 
 ### 🔗 Links
 
-- **GitHub**: [SmartGridsML/mcp1stbirthday_hack](https://github.com/SmartGridsML/mcp1stbirthday_hack)
-- **Claude Desktop Integration**: Works via MCP stdio transport
+- **GitHub**: [mcp1stbirthday_hack](https://github.com/SmartGridsML/mcp1stbirthday_hack)
+- **Claude Desktop Integration**: ✅ Works via MCP stdio transport
+- **Hackathon Track**: Track 1 (Building MCP) + Track 2 (Consumer Use)
 
 ---
 
-*Built for the MCP 1st Birthday Hackathon*
+*Built for the MCP 1st Birthday Hackathon ($50,000 Prize Pool)*

__pycache__/app.cpython-312.pyc

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4d5a629e7fe3a3e8ba5f1d264c4e7c91a1a73394939ae6dc9be8afad56c2d15
+size 133206

requirements.txt

@@ -1,5 +1 @@
 gradio
-langchain
-langchain-google-genai
-langchain-anthropic
-langchain-core