Update app.py

app.py

@@ -3,12 +3,14 @@ import google.generativeai as genai
 from PIL import Image
 import io
 import json
-
+import re
+import os
+import uvicorn
 
 app = FastAPI()
 
-# Set your API key
-genai.configure(api_key="AIzaSyAlvHXHO7xFFjfEyaNOmyZvn9FFPPJdIx4")
+# Set your API key (Make sure to store it securely, e.g., using environment variables)
+genai.configure(api_key=os.getenv("GOOGLE_API_KEY", "YOUR_API_KEY_HERE"))
 
 @app.get("/")
 def read_root():
@@ -21,12 +23,15 @@ async def root_head():
 @app.post("/analyze")
 async def analyze_food_image(file: UploadFile = File(...)):
     """Analyzes food items in the image using Gemini 1.5 Flash"""
+    
+    # Read the uploaded image file
     image_bytes = await file.read()
     img = Image.open(io.BytesIO(image_bytes))
 
-    model = genai.GenerativeModel("gemini-1.5-flash")  # Best for real-time image processing
+    # Load the generative model
+    model = genai.GenerativeModel("gemini-1.5-flash")  # Optimized for real-time image processing
 
-    # Updated prompt for structured JSON response
+    # Updated structured JSON prompt
     prompt = """
     Identify all food items in the given image and determine their approximate quantity. Then, provide nutritional information 
     in valid JSON format following this structure:
@@ -71,17 +76,24 @@ async def analyze_food_image(file: UploadFile = File(...)):
     """
 
     # Get response from API
-    response = model.generate_content([prompt, img])
+    response = model.generate_content([prompt, image_bytes])  # Pass raw image bytes
+
+    # Debugging: Print the raw response
+    print("Raw Response from Gemini:", response.text)
 
     # Extract JSON response
     try:
-        json_response = response.text.strip("```json").strip("```")  # Remove markdown formatting
-        return json.loads(json_response)  # Convert to Python dictionary
+        # Use regex to extract JSON block
+        match = re.search(r'```json\s*(\{.*?\})\s*```', response.text, re.DOTALL)
+        if match:
+            json_response = match.group(1)  # Extract JSON content
+            return json.loads(json_response)  # Convert to Python dictionary
+        else:
+            return {"error": "Response does not contain valid JSON format", "raw_response": response.text}
     except json.JSONDecodeError:
-        return {"error": "Invalid response format from AI model"}
-
-import os
+        return {"error": "Failed to parse JSON response", "raw_response": response.text}
 
+# Entry point for Render deployment
 if __name__ == "__main__":
     port = int(os.getenv("PORT", 7860))  # Render assigns PORT dynamically
-    uvicorn.run(app, host="0.0.0.0", port=port)
+    uvicorn.run(app, host="0.0.0.0", port=port)