code.txt

"""
CrewAI Medical Research Assistant
Compare this with your LangGraph approach
"""

import os
from crewai import Agent, Task, Crew, Process
from langchain_openai import AzureChatOpenAI

# ============== CONFIGURATION ==============
os.environ["AZURE_OPENAI_ENDPOINT"] = "https://your-resource.openai.azure.com/"
os.environ["AZURE_OPENAI_API_KEY"] = "your-key"
os.environ["OPENAI_API_VERSION"] = "2024-02-01"

llm = AzureChatOpenAI(
    deployment_name="gpt-4",
    temperature=0.7,
    model_name="gpt-4"
)

# ============== DEFINE AGENTS ==============

# Agent 1: Medical Researcher
researcher = Agent(
    role="Medical Researcher",
    goal="Research clinical information on medical conditions and treatments",
    backstory="""You are an experienced medical researcher with 15 years of 
    experience in evidence-based medicine. You excel at finding and synthesizing 
    clinical evidence from authoritative sources.""",
    verbose=True,
    allow_delegation=False,  # Won't delegate to other agents
    llm=llm
)

# Agent 2: Clinical Analyst
analyst = Agent(
    role="Clinical Analyst",
    goal="Analyze medical research and identify key clinical insights",
    backstory="""You are a clinical analyst who specializes in interpreting 
    medical research. You can identify important clinical implications, 
    contraindications, and treatment protocols.""",
    verbose=True,
    allow_delegation=False,
    llm=llm
)

# Agent 3: Medical Writer
writer = Agent(
    role="Medical Writer",
    goal="Create clear, accurate clinical summaries for healthcare professionals",
    backstory="""You are a medical writer who creates concise, accurate clinical 
    documentation. You organize information logically and highlight critical 
    clinical details.""",
    verbose=True,
    allow_delegation=False,
    llm=llm
)

# ============== DEFINE TASKS ==============

def create_research_task(condition: str) -> Task:
    """Create research task for given condition"""
    return Task(
        description=f"""Research the medical condition: {condition}
        
        Find and compile information on:
        1. Definition and pathophysiology
        2. Common symptoms and presentations
        3. Current treatment protocols
        4. Evidence-based guidelines
        
        Focus on recent clinical evidence (last 5 years preferred).
        Cite authoritative sources where possible.""",
        agent=researcher,
        expected_output="""Comprehensive research summary including:
        - Condition overview
        - Clinical presentations
        - Treatment options
        - Key evidence sources"""
    )

def create_analysis_task(condition: str) -> Task:
    """Create analysis task"""
    return Task(
        description=f"""Analyze the research findings for {condition}.
        
        Your analysis should include:
        1. First-line treatment recommendations
        2. Important contraindications or warnings
        3. Clinical decision points
        4. Patient considerations
        
        Prioritize information by clinical relevance.""",
        agent=analyst,
        expected_output="""Clinical analysis including:
        - Treatment priorities
        - Clinical warnings
        - Decision framework
        - Patient factors"""
    )

def create_writing_task(condition: str) -> Task:
    """Create writing task"""
    return Task(
        description=f"""Create a clinical summary document for {condition}.
        
        The summary should be:
        - Clear and concise (500-800 words)
        - Organized with clear sections
        - Focused on actionable clinical information
        - Written for healthcare professionals
        
        Include sections:
        1. Overview
        2. Clinical Presentation
        3. Treatment Approach
        4. Key Considerations""",
        agent=writer,
        expected_output="""Professional clinical summary ready for use by 
        healthcare providers, with clear sections and actionable information."""
    )

# ============== CREATE AND RUN CREW ==============

def run_medical_research(condition: str) -> str:
    """
    Run the medical research crew
    """
    print(f"\n{'='*60}")
    print(f"Starting research on: {condition}")
    print(f"{'='*60}\n")
    
    # Create tasks
    research_task = create_research_task(condition)
    analysis_task = create_analysis_task(condition)
    writing_task = create_writing_task(condition)
    
    # Create crew
    crew = Crew(
        agents=[researcher, analyst, writer],
        tasks=[research_task, analysis_task, writing_task],
        process=Process.sequential,  # Tasks execute in order
        verbose=True
    )
    
    # Execute
    result = crew.kickoff()
    
    return result

# ============== MAIN ==============

if __name__ == "__main__":
    # Test with a medical condition
    condition = "Type 2 Diabetes"
    
    result = run_medical_research(condition)
    
    print("\n" + "="*60)
    print("FINAL OUTPUT:")
    print("="*60)
    print(result)
```

---

### **What This Demo Shows:**

**Sequential Process:**
```
User provides condition
    ↓
Researcher Agent → Gathers information
    ↓
Analyst Agent → Analyzes findings (receives researcher's output)
    ↓
Writer Agent → Creates summary (receives analyst's output)
    ↓
Final clinical summary