app.py

import os
from flask import Flask, request, jsonify
from dotenv import load_dotenv
from google.cloud import aiplatform
from google.protobuf import json_format
from google.protobuf.struct_pb2 import Value
from prompt import create_prompt

# Load environment variables from .env file
load_dotenv()

app = Flask(__name__)

# Configure Google Cloud credentials
PROJECT_ID = os.getenv("GCP_PROJECT_ID")
LOCATION = os.getenv("GCP_LOCATION")

@app.route("/api/chat", methods=["POST"])
def chat_handler():
    """
    Handles chat requests by sending them to the Gemini model.
    """
    # Get user query from the request body
    data = request.get_json()
    user_query = data.get("query")

    if not user_query:
        return jsonify({"error": "Query not provided"}), 400

    # Create a prompt for the model
    prompt = create_prompt(user_query)

    # Initialize Vertex AI client
    aiplatform.init(project=PROJECT_ID, location=LOCATION)
    
    # Set up the prediction client
    client_options = {"api_endpoint": f"{LOCATION}-aiplatform.googleapis.com"}
    client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)

    # Define the model endpoint
    endpoint = f"projects/{PROJECT_ID}/locations/{LOCATION}/publishers/google/models/gemini-1.0-pro-vision-001"

    # Create the instance payload
    instance = {"prompt": prompt}
    instances = [json_format.ParseDict(instance, Value())]

    # Set model parameters
    parameters = {
        "temperature": 0.2,
        "maxOutputTokens": 256,
        "topP": 0.8,
        "topK": 40,
    }
    
    try:
        # Send the request to the model
        response = client.predict(
            endpoint=endpoint, instances=instances, parameters=parameters
        )

        # Extract the model's response
        model_response = response.predictions[0]

        return jsonify({"response": model_response})

    except Exception as e:
        return jsonify({"error": str(e)}), 500

if __name__ == "__main__":
    app.run(debug=True, port=5000)