gpt_analyzer.py

# gpt_analyzer.py
import json
from openai import OpenAI
from typing import Dict, List, Any

class GPTAnalyzer:
    def __init__(self, api_key: str):
        print("Initializing GPT Analyzer")
        self.client = OpenAI(api_key=api_key)
    
    def analyze_request(self, request_text: str, available_categories: List[str]) -> Dict[str, Any]:
        print(f"Analyzing request with available categories: {available_categories}")
        
        prompt = f"""
        As a hospital data analyst, analyze this data request:
        "{request_text}"

        Consider these available data sources in hospital Web Data system:
        {json.dumps(available_categories, indent=2, ensure_ascii=False)}

        You must return a response that exactly matches this JSON structure. Fields must exactly match the available report fields. DO NOT include any other text in your response except the JSON:

        {{
            "required_reports": [
                {{
                    "category": "OPD",
                    "report_type": "diagnosis",
                    "fields_needed": ["HN", "Patient_Name", "VisitDate", "Doctor_Name", "ICD_Code", "ICD_Name"],
                    "filters": {{
                        "date_range": "January 2024",
                        "other_filters": []
                    }}
                }}
            ],
            "interpretation": "Brief explanation of what data is needed",
            "confidence_score": "HIGH/MEDIUM/LOW"
        }}

        IMPORTANT:
        - report_type must match exactly one of the available report names
        - fields_needed must match exactly the field names in the reports
        - Return only the JSON, no other text
        """

        try:
            print("Sending request to GPT")
            response = self.client.chat.completions.create(
                messages=[
                    {
                        "role": "system",
                        "content": "You are a healthcare data analyst expert who understands hospital information systems. Always return exact matching report names and field names. Return only JSON, no other text."
                    },
                    {
                        "role": "user",
                        "content": prompt
                    }
                ],
                model="gpt-4o-mini",
            )
            
            response_text = response.choices[0].message.content.strip()
            print(f"GPT Response received: {response_text}")
            
            return json.loads(response_text)
            
        except json.JSONDecodeError as e:
            print(f"JSON Parse Error: {str(e)}")
            return {"error": f"Failed to parse GPT response as JSON: {str(e)}"}
        except Exception as e:
            print(f"General Error: {str(e)}")
            return {"error": str(e)}