Update app.py

app.py

@@ -8,7 +8,159 @@ import plotly.graph_objs as go
 from typing import List, Dict, Any
 import logging
 
-# ... (previous code remains unchanged)
+# Set up logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+
+# Load environment variables
+load_dotenv()
+
+# Initialize Groq client
+client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
+
+def parse_non_json_response(text: str) -> Dict[str, Any]:
+    # Attempt to extract structured information from non-JSON text
+    cancer_types = re.findall(r"(?:cancer type|Cancer Type):\s*(.+?)(?:\n|$)", text, re.IGNORECASE)
+    risk_levels = re.findall(r"(?:risk level|Risk Level):\s*(.+?)(?:\n|$)", text, re.IGNORECASE)
+    descriptions = re.findall(r"(?:description|Description):\s*(.+?)(?:\n|$)", text, re.IGNORECASE)
+    
+    potential_cancer_types = [
+        {"name": ct, "risk_level": rl, "description": desc}
+        for ct, rl, desc in zip(cancer_types, risk_levels, descriptions)
+    ]
+    
+    recommended_tests = re.findall(r"(?:Test Name):\s*(.+?)\n(?:description|Description):\s*(.+?)(?:\n|$)", text, re.IGNORECASE)
+    recommended_tests = [{"name": name, "description": desc} for name, desc in recommended_tests]
+    
+    general_advice = re.search(r"(?:General Advice|general advice):\s*(.+?)(?:\n|$)", text, re.DOTALL | re.IGNORECASE)
+    general_advice = general_advice.group(1) if general_advice else "No general advice provided."
+    
+    disclaimer = re.search(r"(?:DISCLAIMER|Disclaimer):\s*(.+?)(?:\n|$)", text, re.DOTALL | re.IGNORECASE)
+    disclaimer = disclaimer.group(1) if disclaimer else "No disclaimer provided. This tool is for educational purposes only and should not replace professional medical advice."
+    
+    return {
+        "potential_cancer_types": potential_cancer_types,
+        "recommended_tests": recommended_tests,
+        "general_advice": general_advice,
+        "disclaimer": disclaimer
+    }
+
+def get_diagnosis(age: int, gender: str, symptoms: str, medical_history: str) -> Dict[str, Any]:
+    prompt = f"""
+    Given the following patient information, provide a preliminary analysis of potential cancer risks and recommended tests. 
+    Give professional medical advice to the best of your ability.
+    
+    Patient Information:
+    Age: {age}
+    Gender: {gender}
+    Symptoms: {symptoms}
+    Medical History: {medical_history}
+    
+    Please provide a response with the following structure:
+    Potential Cancer Types:
+    - Cancer Type: [Name]
+      Risk Level: [Low/Medium/High]
+      Description: [Brief description of why this cancer type is considered]
+    Recommended Tests:
+    - Test Name: [Name]
+      Description: [Brief description of why this test is recommended]
+    General Advice: [General health advice for the patient]
+    DISCLAIMER: [A strong disclaimer about the limitations of this assessment]
+    
+    Ensure the response emphasizes the importance of consulting with a medical professional for accurate diagnosis and treatment.
+    """
+    
+    try:
+        chat_completion = client.chat.completions.create(
+            messages=[
+                {
+                    "role": "user",
+                    "content": prompt,
+                }
+            ],
+            model="llama-3.1-8b-instant",
+            temperature=0.5,
+            max_tokens=1500,
+        )
+        
+        response_content = chat_completion.choices[0].message.content
+        
+        try:
+            response = json.loads(response_content)
+        except json.JSONDecodeError:
+            response = parse_non_json_response(response_content)
+        
+        return response
+    except Exception as e:
+        logging.error(f"Error in get_diagnosis: {str(e)}")
+        return {"error": f"An error occurred while communicating with the API: {str(e)}"}
+
+def plot_risk(potential_cancer_types: List[Dict[str, str]]) -> go.Figure:
+    if not potential_cancer_types:
+        return None
+
+    names = [c["name"] for c in potential_cancer_types]
+    risk_levels = [1 if c["risk_level"].lower() == "low" else 2 if c["risk_level"].lower() == "medium" else 3 for c in potential_cancer_types]
+    colors = ["green" if rl == 1 else "yellow" if rl == 2 else "red" for rl in risk_levels]
+
+    fig = go.Figure(data=[go.Bar(
+        x=names,
+        y=risk_levels,
+        marker_color=colors,
+        text=risk_levels,
+        textposition='auto',
+    )])
+    fig.update_layout(
+        title="Cancer Risk Levels",
+        yaxis_title="Risk Level",
+        xaxis_tickangle=-45,
+        yaxis=dict(
+            tickmode='array',
+            tickvals=[1, 2, 3],
+            ticktext=['Low', 'Medium', 'High']
+        )
+    )
+
+    return fig
+
+def format_output(response: Dict[str, Any]) -> str:
+    if "error" in response:
+        return f"Error: {response['error']}"
+
+    output = "HealthScan AI: Personalized Cancer Risk Insights\n\n"
+    
+    output += "Potential Cancer Types:\n"
+    for cancer in response.get("potential_cancer_types", []):
+        output += f"- {cancer.get('name', 'N/A')} (Risk Level: {cancer.get('risk_level', 'N/A')})\n"
+        output += f"  {cancer.get('description', 'No description provided.')}\n\n"
+    
+    output += "Recommended Tests:\n"
+    for test in response.get("recommended_tests", []):
+        output += f"- {test.get('name', 'N/A')}: {test.get('description', 'No description provided.')}\n\n"
+    
+    output += f"General Advice:\n{response.get('general_advice', 'No general advice provided.')}\n\n"
+    output += f"DISCLAIMER:\n{response.get('disclaimer', 'No disclaimer provided. This tool is for educational purposes only and should not replace professional medical advice.')}"
+    
+    return output
+
+def validate_input(age: int, gender: str, symptoms: str, medical_history: str) -> List[str]:
+    errors = []
+    if not (0 < age < 120):
+        errors.append("Please enter a valid age between 1 and 120.")
+    if not symptoms.strip():
+        errors.append("Please enter at least one symptom.")
+    return errors
+
+def process_input(age: int, gender: str, symptoms: str, medical_history: str) -> tuple:
+    errors = validate_input(age, gender, symptoms, medical_history)
+    if errors:
+        return "\n".join(errors), None
+
+    diagnosis = get_diagnosis(age, gender, symptoms, medical_history)
+    output = format_output(diagnosis)
+    
+    risk_plot = plot_risk(diagnosis.get("potential_cancer_types", []))
+    
+    return output, risk_plot
 
 def clear_inputs():
     return gr.Number(value=None), gr.Radio(value=None), gr.Textbox(value=""), gr.Textbox(value="")