Update app.py

app.py

@@ -5,6 +5,7 @@ import requests, time
 from data_extractor import extract_data, find_product, get_product
 from nutrient_analyzer import analyze_nutrients
 from rda import find_nutrition
+from typing import Dict, Any
 
 #Used the @st.cache_resource decorator on this function. 
 #This Streamlit decorator ensures that the function is only executed once and its result (the OpenAI client) is cached. 
@@ -37,54 +38,92 @@ def get_product_list(product_name_by_user, data_extractor_url):
     return response
 
 
-def rda_analysis(product_info_from_db_nutritionalInformation, product_info_from_db_servingSize):
-    data_for_rda_analysis = {'nutritionPerServing': {}, 'userServingSize': None}
-    nutrient_name_list = ['energy', 'protein', 'carbohydrates', 'addedSugars', 'dietaryFiber', 'totalFat', 
-                          'saturatedFat', 'monounsaturatedFat', 'polyunsaturatedFat', 'transFat', 'sodium']
+def rda_analysis(product_info_from_db_nutritionalInformation: Dict[str, Any], 
+                product_info_from_db_servingSize: float) -> Dict[str, Any]:
+    """
+    Analyze nutritional information and return RDA analysis data in a structured format.
+    
+    Args:
+        product_info_from_db_nutritionalInformation: Dictionary containing nutritional information
+        product_info_from_db_servingSize: Serving size value
+        
+    Returns:
+        Dictionary containing nutrition per serving and user serving size
+    """
+    nutrient_name_list = [
+        'energy', 'protein', 'carbohydrates', 'addedSugars', 'dietaryFiber',
+        'totalFat', 'saturatedFat', 'monounsaturatedFat', 'polyunsaturatedFat',
+        'transFat', 'sodium'
+    ]
+    
+    # Define the JSON schema
+    json_schema = {
+        "type": "object",
+        "properties": {
+            "energy": {"type": "number"},
+            "protein": {"type": "number"},
+            "carbohydrates": {"type": "number"},
+            "addedSugars": {"type": "number"},
+            "dietaryFiber": {"type": "number"},
+            "totalFat": {"type": "number"},
+            "saturatedFat": {"type": "number"},
+            "monounsaturatedFat": {"type": "number"},
+            "polyunsaturatedFat": {"type": "number"},
+            "transFat": {"type": "number"},
+            "sodium": {"type": "number"},
+            "servingSize": {"type": "number"}
+        },
+        "required": nutrient_name_list + ["servingSize"],
+        "additionalProperties": False
+    }
 
-    # Using the response_format to request output in structured JSON format
-    completion = client.chat.completions.create(
-        model="gpt-4o",
-        messages=[
-            {"role": "system", "content": "You will be given nutritional information of a food product."},
-            {"role": "user", "content": f"Nutritional content of food product is {json.dumps(product_info_from_db_nutritionalInformation)}. Extract the values of the following nutrients: {', '.join(nutrient_name_list)}."}
-        ],
-        response_format="json",  # This specifies we want structured JSON in return
-        functions=[
-            {
-                "name": "rda_analysis",
-                "description": "Return the extracted nutrient information in the correct format",
-                "parameters": {
-                    "type": "object",
-                    "properties": {
-                        "energy": {"type": "number"},
-                        "protein": {"type": "number"},
-                        "carbohydrates": {"type": "number"},
-                        "addedSugars": {"type": "number"},
-                        "dietaryFiber": {"type": "number"},
-                        "totalFat": {"type": "number"},
-                        "saturatedFat": {"type": "number"},
-                        "monounsaturatedFat": {"type": "number"},
-                        "polyunsaturatedFat": {"type": "number"},
-                        "transFat": {"type": "number"},
-                        "sodium": {"type": "number"},
-                        "servingSize": {"type": "number"}  # Ensures you get serving size too
-                    },
-                    "required": nutrient_name_list + ["servingSize"]
+    try:
+        completion = client.chat.completions.create(
+            model="gpt-4o",
+            messages=[
+                {
+                    "role": "system",
+                    "content": """You will be given nutritional information of a food product. 
+                                Return the data in the exact JSON format specified in the schema, 
+                                with all required fields."""
+                },
+                {
+                    "role": "user",
+                    "content": f"Nutritional content of food product is {json.dumps(product_info_from_db_nutritionalInformation)}. "
+                              f"Extract the values of the following nutrients: {', '.join(nutrient_name_list)}."
                 }
+            ],
+            response_format={
+                "type": "json_schema",
+                "schema": json_schema,
+                "strict": True
             }
-        ]
-    )
-
-    data_for_rda_analysis_part_1 = completion.choices[0].message["content"]
-    print(f"data_for_rda_analysis_part_1: {data_for_rda_analysis_part_1}")
-
-    data_for_rda_analysis = {
-        'nutritionPerServing': data_for_rda_analysis_part_1,
-        'userServingSize': product_info_from_db_servingSize
-    }
-    
-    return data_for_rda_analysis
+        )
+        
+        # Parse the JSON response
+        nutrition_data = json.loads(completion.choices[0].message["content"])
+        
+        # Validate that all required fields are present
+        missing_fields = [field for field in nutrient_name_list + ["servingSize"] 
+                         if field not in nutrition_data]
+        if missing_fields:
+            print(f"Missing required fields in API response: {missing_fields}")
+        
+        # Validate that all values are numbers
+        non_numeric_fields = [field for field, value in nutrition_data.items() 
+                            if not isinstance(value, (int, float))]
+        if non_numeric_fields:
+            raise ValueError(f"Non-numeric values found in fields: {non_numeric_fields}")
+        
+        return {
+            'nutritionPerServing': nutrition_data,
+            'userServingSize': product_info_from_db_servingSize
+        }
+        
+    except Exception as e:
+        # Log the error and raise it for proper handling
+        print(f"Error in RDA analysis: {str(e)}")
+        raise
 
 
 def find_product_nutrients(product_info_from_db):