llm.py

import google.generativeai as genai
import os
import json
from dotenv import load_dotenv

load_dotenv()

# Support multiple Gemini keys (comma-separated or single key)
api_keys = os.getenv("GOOGLE_API_KEYS") or os.getenv("GOOGLE_API_KEY")
if not api_keys:
    raise ValueError("No Gemini API keys found in GOOGLE_API_KEYS or GOOGLE_API_KEY environment variable.")

api_keys = [k.strip() for k in api_keys.split(",") if k.strip()]
print(f"Loaded {len(api_keys)} Gemini API key(s)")

def query_gemini(questions, contexts, max_retries=3):
    import itertools

    context = "\n\n".join(contexts)
    questions_text = "\n".join([f"{i+1}. {q}" for i, q in enumerate(questions)])

    prompt = f"""
You are an expert insurance assistant generating formal yet user-facing answers to policy questions and Other Human Questions. Your goal is to write professional, structured answers that reflect the language of policy documents — but are still human-readable and easy to understand.
🧠 FORMAT & TONE GUIDELINES:
- Write in professional third-person language (no "you", no "we").
- Use clear sentence structure with proper punctuation and spacing.
- Do NOT write in legalese or robotic passive constructions.
- Include eligibility, limits, and waiting periods explicitly where relevant.
- Keep it factual, neutral, and easy to follow.
- First, try to answer each question using information from the provided context.
- If the question is NOT covered by the context Provide Then Give The General Answer It Not Be In Context if Nothing Found Give Normal Ai Answer for The Question Correctly
- Limit each answer to 2–3 sentences, and do not repeat unnecessary information.
- If a question can be answered with a simple "Yes", "No", "Can apply", or "Cannot apply", then begin the answer with that phrase, followed by a short supporting Statement In Natural Human Like response.So Give A Good Answer For The Question With Correct Information.
- Avoid giving  theory Based Long Long answers Try to Give Short Good Reasonable Answers.
🛑 DO NOT:
- Use words like "context", "document", or "text".
- Output markdown, bullets, emojis, or markdown code blocks.
- Say "helpful", "available", "allowed", "indemnified", "excluded", etc.
- Use overly robotic passive constructions like "shall be indemnified".
- Dont Give In Message Like "Based On The Context "Or "Nothing Refered In The context" Like That Dont Give In Response Try To Give Answer For The Question Alone
✅ DO:
- Write in clean, informative language.
- Give complete answers in 2–3 sentences maximum.
📤 OUTPUT FORMAT (strict):
Respond with only the following JSON — no explanations, no comments, no markdown:
{{
  "answers": [
    "Answer to question 1",
    "Answer to question 2",
    ...
  ]
}}
📚 CONTEXT:
{context}
❓ QUESTIONS:
{questions_text}
Your task: For each question, provide a complete, professional, and clearly written answer in 2–3 sentences using a formal but readable tone.
"""

    last_exception = None
    total_attempts = len(api_keys) * max_retries
    key_cycle = itertools.cycle(api_keys)

    for attempt in range(total_attempts):
        key = next(key_cycle)
        try:
            genai.configure(api_key=key)
            model = genai.GenerativeModel("gemini-2.5-flash-lite")
            response = model.generate_content(prompt)
            response_text = getattr(response, "text", "").strip()

            if not response_text:
                raise ValueError("Empty response received from Gemini API.")

            if response_text.startswith("```json"):
                response_text = response_text.replace("```json", "").replace("```", "").strip()
            elif response_text.startswith("```"):
                response_text = response_text.replace("```", "").strip()

            parsed = json.loads(response_text)
            if "answers" in parsed and isinstance(parsed["answers"], list):
                return parsed
            else:
                raise ValueError("Invalid response format received from Gemini.")

        except Exception as e:
            last_exception = e
            msg = str(e).lower()
            print(f"[Retry {attempt+1}/{total_attempts}] Gemini key {key[:8]}... failed: {e}")
            continue

    print(f"All Gemini API attempts failed. Last error: {last_exception}")
    return {"answers": [f"Error generating response: {str(last_exception)}"] * len(questions)}