shivanis14's picture
Update app.py
239bcd0 verified
import streamlit as st
from openai import OpenAI
import json, os
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
from calc_cosine_similarity import find_cosine_similarity, find_embedding , find_relevant_file_paths
import pickle
#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.
#Subsequent calls to this function will return the cached client, avoiding unnecessary recreation.
@st.cache_resource
def get_openai_client():
#Enable debug mode for testing only
return True, OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
@st.cache_resource
def get_backend_urls():
data_extractor_url = "https://data-extractor-67qj89pa0-sonikas-projects-9936eaad.vercel.app/"
return data_extractor_url
debug_mode, client = get_openai_client()
data_extractor_url = get_backend_urls()
assistant_default_doc = None
def extract_data_from_product_image(image_links, data_extractor_url):
response = extract_data(image_links)
return response
def get_product_list(product_name_by_user, data_extractor_url):
response = find_product(product_name_by_user)
return response
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'
]
try:
response = 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", "json_schema": {
"name": "Nutritional_Info_Label_Reader",
"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
},
"strict": True
}}
)
# Parse the JSON response
nutrition_data = json.loads(response.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):
#GET Response: {'_id': '6714f0487a0e96d7aae2e839',
#'brandName': 'Parle', 'claims': ['This product does not contain gold'],
#'fssaiLicenseNumbers': [10013022002253],
#'ingredients': [{'metadata': '', 'name': 'Refined Wheat Flour (Maida)', 'percent': '63%'}, {'metadata': '', 'name': 'Sugar', 'percent': ''}, {'metadata': '', 'name': 'Refined Palm Oil', 'percent': ''}, {'metadata': '(Glucose, Levulose)', 'name': 'Invert Sugar Syrup', 'percent': ''}, {'metadata': 'I', 'name': 'Sugar Citric Acid', 'percent': ''}, {'metadata': '', 'name': 'Milk Solids', 'percent': '1%'}, {'metadata': '', 'name': 'Iodised Salt', 'percent': ''}, {'metadata': '503(I), 500 (I)', 'name': 'Raising Agents', 'percent': ''}, {'metadata': '1101 (i)', 'name': 'Flour Treatment Agent', 'percent': ''}, {'metadata': 'Diacetyl Tartaric and Fatty Acid Esters of Glycerol (of Vegetable Origin)', 'name': 'Emulsifier', 'percent': ''}, {'metadata': 'Vanilla', 'name': 'Artificial Flavouring Substances', 'percent': ''}],
#'nutritionalInformation': [{'name': 'Energy', 'unit': 'kcal', 'values': [{'base': 'per 100 g','value': 462}]},
#{'name': 'Protein', 'unit': 'g', 'values': [{'base': 'per 100 g', 'value': 6.7}]},
#{'name': 'Carbohydrate', 'unit': 'g', 'values': [{'base': 'per 100 g', 'value': 76.0}, {'base': 'of which sugars', 'value': 26.9}]},
#{'name': 'Fat', 'unit': 'g', 'values': [{'base': 'per 100 g', 'value': 14.6}, {'base': 'Saturated Fat', 'value': 6.8}, {'base': 'Trans Fat', 'value': 0}]},
#{'name': 'Total Sugars', 'unit': 'g', 'values': [{'base': 'per 100 g', 'value': 27.7}]},
#{'name': 'Added Sugars', 'unit': 'g', 'values': [{'base': 'per 100 g', 'value': 26.9}]},
#{'name': 'Cholesterol', 'unit': 'mg', 'values': [{'base': 'per 100 g', 'value': 0}]},
#{'name': 'Sodium', 'unit': 'mg', 'values': [{'base': 'per 100 g', 'value': 281}]}],
#'packagingSize': {'quantity': 82, 'unit': 'g'},
#'productName': 'Parle-G Gold Biscuits',
#'servingSize': {'quantity': 18.8, 'unit': 'g'},
#'servingsPerPack': 3.98,
#'shelfLife': '7 months from packaging'}
product_type = None
calories = None
sugar = None
total_sugar = None
added_sugar = None
salt = None
serving_size = None
if product_info_from_db["servingSize"]["unit"].lower() == "g":
product_type = "solid"
elif product_info_from_db["servingSize"]["unit"].lower() == "ml":
product_type = "liquid"
serving_size = product_info_from_db["servingSize"]["quantity"]
for item in product_info_from_db["nutritionalInformation"]:
if 'energy' in item['name'].lower():
calories = item['values'][0]['value']
if 'total sugar' in item['name'].lower():
total_sugar = item['values'][0]['value']
if 'added sugar' in item['name'].lower():
added_sugar = item['values'][0]['value']
if 'sugar' in item['name'].lower() and 'added sugar' not in item['name'].lower() and 'total sugar' not in item['name'].lower():
sugar = item['values'][0]['value']
if 'salt' in item['name'].lower():
if salt is None:
salt = 0
salt += item['values'][0]['value']
if salt is None:
salt = 0
for item in product_info_from_db["nutritionalInformation"]:
if 'sodium' in item['name'].lower():
salt += item['values'][0]['value']
if added_sugar is not None and added_sugar > 0 and sugar is None:
sugar = added_sugar
elif total_sugar is not None and total_sugar > 0 and added_sugar is None and sugar is None:
sugar = total_sugar
return product_type, calories, sugar, salt, serving_size
# Initialize assistants and vector stores
# Function to initialize vector stores and assistants
@st.cache_resource
def initialize_assistants_and_vector_stores():
#Processing Level
global client
assistant1 = client.beta.assistants.create(
name="Processing Level",
instructions="You are an expert dietician. Use your knowledge base to answer questions about the processing level of food product.",
model="gpt-4o",
tools=[{"type": "file_search"}],
temperature=0,
top_p = 0.85
)
#Harmful Ingredients
assistant3 = client.beta.assistants.create(
name="Misleading Claims",
instructions="You are an expert dietician. Use your knowledge base to answer questions about the misleading claims about food product.",
model="gpt-4o",
tools=[{"type": "file_search"}],
temperature=0,
top_p = 0.85
)
# Create a vector store
vector_store1 = client.beta.vector_stores.create(name="Processing Level Vec")
# Ready the files for upload to OpenAI
file_paths = ["Processing_Level.docx"]
file_streams = [open(path, "rb") for path in file_paths]
# Use the upload and poll SDK helper to upload the files, add them to the vector store,
# and poll the status of the file batch for completion.
file_batch1 = client.beta.vector_stores.file_batches.upload_and_poll(
vector_store_id=vector_store1.id, files=file_streams
)
# You can print the status and the file counts of the batch to see the result of this operation.
print(file_batch1.status)
print(file_batch1.file_counts)
# Create a vector store
vector_store3 = client.beta.vector_stores.create(name="Misleading Claims Vec")
# Ready the files for upload to OpenAI
file_paths = ["MisLeading_Claims.docx"]
file_streams = [open(path, "rb") for path in file_paths]
# Use the upload and poll SDK helper to upload the files, add them to the vector store,
# and poll the status of the file batch for completion.
file_batch3 = client.beta.vector_stores.file_batches.upload_and_poll(
vector_store_id=vector_store3.id, files=file_streams
)
# You can print the status and the file counts of the batch to see the result of this operation.
print(file_batch3.status)
print(file_batch3.file_counts)
#Processing Level
assistant1 = client.beta.assistants.update(
assistant_id=assistant1.id,
tool_resources={"file_search": {"vector_store_ids": [vector_store1.id]}},
)
#Misleading Claims
assistant3 = client.beta.assistants.update(
assistant_id=assistant3.id,
tool_resources={"file_search": {"vector_store_ids": [vector_store3.id]}},
)
embeddings_titles = []
if not os.path.exists('embeddings.pkl'):
#Find embeddings of titles from titles.txt
titles = []
#if embedding_titles.pkl is absent
with open('titles.txt', 'r') as file:
lines = file.readlines()
titles = [line.strip() for line in lines]
embeddings_titles = find_embedding(titles, lim=50)
#Save embeddings_titles to embedding_titles.pkl
data = {
'sentences': titles[:50],
'embeddings': embeddings_titles
}
with open('embeddings.pkl', 'wb') as f:
pickle.dump(data, f)
if os.path.exists("embeddings.pkl"):
print("embeddings.pkl successfully written!")
else:
print("Reading embeddings.pkl")
# Load both sentences and embeddings
with open('embeddings.pkl', 'rb') as f:
loaded_data = pickle.load(f)
embeddings_titles = loaded_data['embeddings']
return assistant1, assistant3, embeddings_titles
assistant1, assistant3, embeddings_titles = initialize_assistants_and_vector_stores()
def get_files_with_ingredient_info(ingredient, N=1):
file_paths = []
#Find embedding for title of all files
global embeddings_titles
with open('titles.txt', 'r') as file:
lines = file.readlines()
titles = [line.strip() for line in lines]
#Apply cosine similarity between embedding of ingredient name and title of all files
file_paths_abs, file_titles = find_relevant_file_paths(ingredient, embeddings_titles, titles, N=N)
#Fine top N titles that are the most similar to the ingredient's name
#Find file names for those titles
if len(file_paths_abs) == 0:
file_paths.append("Ingredients.docx")
else:
for file_path in file_paths_abs:
file_paths.append(f"articles/{file_path}")
print(f"Titles are {file_titles}")
return file_paths
def get_assistant_for_ingredient(ingredient, N=2):
global client
global assistant_default_doc
#Harmful Ingredients
assistant2 = client.beta.assistants.create(
name="Harmful Ingredients",
instructions=f"You are an expert dietician. Use your knowledge base to answer questions about the ingredient {ingredient} in a food product.",
model="gpt-4o",
tools=[{"type": "file_search"}],
temperature=0,
top_p = 0.85
)
# Create a vector store
vector_store2 = client.beta.vector_stores.create(
name="Harmful Ingredients Vec",
chunking_strategy={
"type": "static",
"static": {
"max_chunk_size_tokens": 400, # Set your desired max chunk size
"chunk_overlap_tokens": 200 # Set your desired overlap size
}
}
)
# Ready the files for upload to OpenAI.
file_paths = get_files_with_ingredient_info(ingredient, N)
if file_paths[0] == "Ingredients.docx" and assistant_default_doc:
#print(f"Using Ingredients.docx for analyzing ingredient {ingredient}")
return assistant_default_doc
print(f"DEBUG : Creating vector store for files {file_paths} to analyze ingredient {ingredient}")
file_streams = [open(path, "rb") for path in file_paths]
# Use the upload and poll SDK helper to upload the files, add them to the vector store,
# and poll the status of the file batch for completion.
file_batch2 = client.beta.vector_stores.file_batches.upload_and_poll(
vector_store_id=vector_store2.id, files=file_streams
)
# You can print the status and the file counts of the batch to see the result of this operation.
print(file_batch2.status)
print(file_batch2.file_counts)
#harmful Ingredients
assistant2 = client.beta.assistants.update(
assistant_id=assistant2.id,
tool_resources={"file_search": {"vector_store_ids": [vector_store2.id]}},
)
if file_paths[0] == "Ingredients.docx" and assistant_default_doc is None:
assistant_default_doc = assistant2
return assistant2
def analyze_nutrition_icmr_rda(nutrient_analysis, nutrient_analysis_rda):
global debug_mode, client
system_prompt = """
Task: Analyze the nutritional content of the food item and compare it to the Recommended Daily Allowance (RDA) or threshold limits defined by ICMR. Provide practical, contextual insights based on the following nutrients:
Nutrient Breakdown and Analysis:
Calories:
Compare the calorie content to a well-balanced meal.
Calculate how many meals' worth of calories the product contains, providing context for balanced eating.
Sugar & Salt:
Convert the amounts of sugar and salt into teaspoons to help users easily understand their daily intake.
Explain whether the levels exceed the ICMR-defined limits and what that means for overall health.
Fat & Calories:
Analyze fat content, specifying whether it is high or low in relation to a balanced diet.
Offer insights on how the fat and calorie levels may impact the user’s overall diet, including potential risks or benefits.
Contextual Insights:
For each nutrient, explain how its levels (whether high or low) affect health and diet balance.
Provide actionable recommendations for the user, suggesting healthier alternatives or adjustments to consumption if necessary.
Tailor the advice to the user's lifestyle, such as recommending lower intake if sedentary or suggesting other dietary considerations based on the product's composition.
Output Structure:
For each nutrient (Calories, Sugar, Salt, Fat), specify if the levels exceed or are below the RDA or ICMR threshold.
Provide clear, concise comparisons (e.g., sugar exceeds the RDA by 20%, equivalent to X teaspoons).
"""
user_prompt = f"""
Nutrition Analysis :
{nutrient_analysis}
{nutrient_analysis_rda}
"""
if debug_mode:
print(f"\nuser_prompt : \n {user_prompt}")
completion = client.chat.completions.create(
model="gpt-4o", # Make sure to use an appropriate model
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": user_prompt}
]
)
return completion.choices[0].message.content
def analyze_processing_level(ingredients, assistant_id):
global debug_mode, client
thread = client.beta.threads.create(
messages=[
{
"role": "user",
"content": "Categorize food product that has following ingredients: " + ', '.join(ingredients) + " into Group A, Group B, or Group C based on the document. The output must only be the group category name (Group A, Group B, or Group C) alongwith the reason behind assigning that respective category to the product. If the group category cannot be determined, output 'NOT FOUND'.",
}
]
)
run = client.beta.threads.runs.create_and_poll(
thread_id=thread.id,
assistant_id=assistant_id,
include=["step_details.tool_calls[*].file_search.results[*].content"]
)
# Polling loop to wait for a response in the thread
messages = []
max_retries = 10 # You can set a maximum retry limit
retries = 0
wait_time = 2 # Seconds to wait between retries
while retries < max_retries:
messages = list(client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id))
if messages: # If we receive any messages, break the loop
break
retries += 1
time.sleep(wait_time)
# Check if we got the message content
if not messages:
raise TimeoutError("Processing Level : No messages were returned after polling.")
message_content = messages[0].content[0].text
annotations = message_content.annotations
#citations = []
for index, annotation in enumerate(annotations):
message_content.value = message_content.value.replace(annotation.text, "")
#if file_citation := getattr(annotation, "file_citation", None):
# cited_file = client.files.retrieve(file_citation.file_id)
# citations.append(f"[{index}] {cited_file.filename}")
if debug_mode:
print(message_content.value)
processing_level_str = message_content.value
return processing_level_str
def analyze_harmful_ingredients(ingredient, assistant_id):
global debug_mode, client
thread = client.beta.threads.create(
messages=[
{
"role": "user",
"content": "A food product has the ingredient: " + ingredient + ". Is this ingredient safe to eat? The output must be in JSON format: {<ingredient_name>: <information from the document about why ingredient is harmful>}. If information about an ingredient is not found in the documents, the value for that ingredient must start with the prefix '(NOT FOUND IN DOCUMENT)' followed by the LLM's response based on its own knowledge.",
}
]
)
run = client.beta.threads.runs.create_and_poll(
thread_id=thread.id,
assistant_id=assistant_id,
include=["step_details.tool_calls[*].file_search.results[*].content"],
tools=[{
"type": "file_search",
"file_search": {
"max_num_results": 5
}
}]
)
# List run steps to get step IDs
run_steps = client.beta.threads.runs.steps.list(
thread_id=thread.id,
run_id=run.id
)
# Initialize a list to store step IDs and their corresponding run steps
all_steps_info = []
# Iterate over each step in run_steps.data
for step in run_steps.data: # Access each RunStep object
step_id = step.id # Get the step ID (use 'step_id' instead of 'id')
# Retrieve detailed information for each step using its ID
run_step_detail = client.beta.threads.runs.steps.retrieve(
thread_id=thread.id,
run_id=run.id,
step_id=step_id,
include=["step_details.tool_calls[*].file_search.results[*].content"]
)
# Append a tuple of (step_id, run_step_detail) to the list
all_steps_info.append((step_id, run_step_detail))
# Print all step IDs and their corresponding run steps
for step_id, run_step_detail in all_steps_info:
print(f"Step ID: {step_id}")
print(f"Run Step Detail: {run_step_detail}\n")
# Polling loop to wait for a response in the thread
messages = []
max_retries = 10 # You can set a maximum retry limit
retries = 0
wait_time = 2 # Seconds to wait between retries
while retries < max_retries:
messages = list(client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id))
if messages: # If we receive any messages, break the loop
break
retries += 1
time.sleep(wait_time)
# Check if we got the message content
if not messages:
raise TimeoutError("Processing Ingredients : No messages were returned after polling.")
message_content = messages[0].content[0].text
annotations = message_content.annotations
#citations = []
#print(f"Length of annotations is {len(annotations)}")
for index, annotation in enumerate(annotations):
if file_citation := getattr(annotation, "file_citation", None):
#cited_file = client.files.retrieve(file_citation.file_id)
#citations.append(f"[{index}] {cited_file.filename}")
message_content.value = message_content.value.replace(annotation.text, "")
if debug_mode:
ingredients_not_found_in_doc = []
print(message_content.value)
for key, value in json.loads(message_content.value.replace("```", "").replace("json", "")).items():
if value.startswith("(NOT FOUND IN DOCUMENT)"):
ingredients_not_found_in_doc.append(key)
print(f"Ingredients not found in database {','.join(ingredients_not_found_in_doc)}")
harmful_ingredient_analysis = json.loads(message_content.value.replace("```", "").replace("json", "").replace("(NOT FOUND IN DOCUMENT) ", ""))
harmful_ingredient_analysis_str = ""
for key, value in harmful_ingredient_analysis.items():
harmful_ingredient_analysis_str += f"{key}: {value}\n"
return harmful_ingredient_analysis_str
def analyze_claims(claims, ingredients, assistant_id):
global debug_mode, client
thread = client.beta.threads.create(
messages=[
{
"role": "user",
"content": "A food product named has the following claims: " + ', '.join(claims) + " and ingredients: " + ', '.join(ingredients) + """. Please evaluate the validity of each claim as well as assess if the product name is misleading.
The output must be in JSON format as follows:
{
<claim_name>: {
'Verdict': <A judgment on the claim's accuracy, ranging from 'Accurate' to varying degrees of 'Misleading'>,
'Why?': <A concise, bulleted summary explaining the specific ingredients or aspects contributing to the discrepancy>,
'Detailed Analysis': <An in-depth explanation of the claim, incorporating relevant regulatory guidelines and health perspectives to support the verdict>
}
}
"""
}
]
)
run = client.beta.threads.runs.create_and_poll(
thread_id=thread.id,
assistant_id=assistant_id,
include=["step_details.tool_calls[*].file_search.results[*].content"]
)
# Polling loop to wait for a response in the thread
messages = []
max_retries = 10 # You can set a maximum retry limit
retries = 0
wait_time = 2 # Seconds to wait between retries
while retries < max_retries:
messages = list(client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id))
if messages: # If we receive any messages, break the loop
break
retries += 1
time.sleep(wait_time)
# Check if we got the message content
if not messages:
raise TimeoutError("Processing Claims : No messages were returned after polling.")
message_content = messages[0].content[0].text
annotations = message_content.annotations
#citations = []
#print(f"Length of annotations is {len(annotations)}")
for index, annotation in enumerate(annotations):
if file_citation := getattr(annotation, "file_citation", None):
#cited_file = client.files.retrieve(file_citation.file_id)
#citations.append(f"[{index}] {cited_file.filename}")
message_content.value = message_content.value.replace(annotation.text, "")
#if debug_mode:
# claims_not_found_in_doc = []
# print(message_content.value)
# for key, value in json.loads(message_content.value.replace("```", "").replace("json", "")).items():
# if value.startswith("(NOT FOUND IN DOCUMENT)"):
# claims_not_found_in_doc.append(key)
# print(f"Claims not found in the doc are {','.join(claims_not_found_in_doc)}")
#claims_analysis = json.loads(message_content.value.replace("```", "").replace("json", "").replace("(NOT FOUND IN DOCUMENT) ", ""))
claims_analysis = {}
if message_content.value != "":
claims_analysis = json.loads(message_content.value.replace("```", "").replace("json", ""))
claims_analysis_str = ""
for key, value in claims_analysis.items():
claims_analysis_str += f"{key}: {value}\n"
return claims_analysis_str
def generate_final_analysis(brand_name, product_name, nutritional_level, processing_level, harmful_ingredient_analysis, claims_analysis, system_prompt):
global debug_mode, client
system_prompt_orig = """You are provided with a detailed analysis of a food product. Your task is to generate actionable insights to help the user decide whether to consume the product, at what frequency, and identify any potential harms or benefits. Consider the context of consumption to ensure the advice is personalized and practical.
Use the following criteria to generate your response:
1. **Nutrition Analysis:**
- How much do sugar, calories, or salt exceed the threshold limit?
- How processed is the product?
- How much of the Recommended Dietary Allowance (RDA) does the product provide for each nutrient?
2. **Harmful Ingredients:**
- Identify any harmful or questionable ingredients.
3. **Misleading Claims:**
- Are there any misleading claims made by the brand?
Additionally, consider the following while generating insights:
1. **Consumption Context:**
- Is the product being consumed for health reasons or as a treat?
- Could the consumer be overlooking hidden harms?
- If the product is something they could consume daily, should they?
- If they are consuming it daily, what potential harm are they not noticing?
- If the product is intended for health purposes, are there concerns the user might miss?
**Output:**
- Recommend whether the product should be consumed or avoided.
- If recommended, specify the appropriate frequency and intended functionality (e.g., treat vs. health).
- Highlight any risks or benefits at that level of consumption."""
user_prompt = f"""
Product Name: {brand_name} {product_name}
Nutrition Analysis :
{nutritional_level}
Processing Level:
{processing_level}
Ingredient Analysis:
{harmful_ingredient_analysis}
Claims Analysis:
{claims_analysis}
"""
if debug_mode:
print(f"\nuser_prompt : \n {user_prompt}")
completion = client.chat.completions.create(
model="gpt-4o", # Make sure to use an appropriate model
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": user_prompt}
]
)
return f"Brand: {brand_name}\n\nProduct: {product_name}\n\nAnalysis:\n\n{completion.choices[0].message.content}"
def analyze_product(product_info_raw, system_prompt):
global assistant1, assistant3
if product_info_raw:
product_info_from_db = product_info_raw
brand_name = product_info_from_db.get("brandName", "")
product_name = product_info_from_db.get("productName", "")
ingredients_list = [ingredient["name"] for ingredient in product_info_from_db.get("ingredients", [])]
claims_list = product_info_from_db.get("claims", [])
nutritional_information = product_info_from_db['nutritionalInformation']
serving_size = product_info_from_db["servingSize"]["quantity"]
nutrient_analysis_rda = ""
nutrient_analysis = ""
nutritional_level = ""
processing_level = ""
harmful_ingredient_analysis = ""
claims_analysis = ""
if nutritional_information:
product_type, calories, sugar, salt, serving_size = find_product_nutrients(product_info_from_db)
if product_type is not None and serving_size is not None and serving_size > 0:
nutrient_analysis = analyze_nutrients(product_type, calories, sugar, salt, serving_size)
else:
return "product not found because product information in the db is corrupt"
print(f"DEBUG ! nutrient analysis is {nutrient_analysis}")
nutrient_analysis_rda_data = rda_analysis(nutritional_information, serving_size)
print(f"DEBUG ! Data for RDA nutrient analysis is of type {type(nutrient_analysis_rda_data)} - {nutrient_analysis_rda_data}")
print(f"DEBUG : nutrient_analysis_rda_data['nutritionPerServing'] : {nutrient_analysis_rda_data['nutritionPerServing']}")
print(f"DEBUG : nutrient_analysis_rda_data['userServingSize'] : {nutrient_analysis_rda_data['userServingSize']}")
nutrient_analysis_rda = find_nutrition(nutrient_analysis_rda_data)
print(f"DEBUG ! RDA nutrient analysis is {nutrient_analysis_rda}")
#Call GPT for nutrient analysis
nutritional_level = analyze_nutrition_icmr_rda(nutrient_analysis, nutrient_analysis_rda)
if len(ingredients_list) > 0:
processing_level = analyze_processing_level(ingredients_list, assistant1.id) if ingredients_list else ""
for ingredient in ingredients_list:
assistant_id_ingredient = get_assistant_for_ingredient(ingredient, 2)
harmful_ingredient_analysis += analyze_harmful_ingredients(ingredient, assistant_id_ingredient.id) + "\n"
if len(claims_list) > 0:
claims_analysis = analyze_claims(claims_list, ingredients_list, assistant3.id) if claims_list else ""
final_analysis = generate_final_analysis(brand_name, product_name, nutritional_level, processing_level, harmful_ingredient_analysis, claims_analysis, system_prompt)
return final_analysis
#else:
# return "I'm sorry, product information could not be extracted from the url."
# Streamlit app
# Initialize session state
if 'messages' not in st.session_state:
st.session_state.messages = []
def chatbot_response(image_urls_str, product_name_by_user, data_extractor_url, system_prompt, extract_info = True):
# Process the user input and generate a response
processing_level = ""
harmful_ingredient_analysis = ""
claims_analysis = ""
image_urls = []
if product_name_by_user != "":
similar_product_list_json = get_product_list(product_name_by_user, data_extractor_url)
if similar_product_list_json and extract_info == False:
with st.spinner("Fetching product information from our database... This may take a moment."):
print(f"similar_product_list_json : {similar_product_list_json}")
if 'error' not in similar_product_list_json.keys():
similar_product_list = similar_product_list_json['products']
return similar_product_list, "Product list found from our database"
else:
return [], "Product list not found"
elif extract_info == True:
with st.spinner("Analyzing the product... This may take a moment."):
product_info_raw = get_product(product_name_by_user)
print(f"DEBUG product_info_raw from name: {type(product_info_raw)} {product_info_raw}")
if not product_info_raw:
return [], "product not found because product information in the db is corrupt"
if 'error' not in product_info_raw.keys():
final_analysis = analyze_product(product_info_raw, system_prompt)
return [], final_analysis
else:
return [], f"Product information could not be extracted from our database because of {product_info_raw['error']}"
else:
return [], "Product not found in our database."
elif "http:/" in image_urls_str.lower() or "https:/" in image_urls_str.lower():
# Extract image URL from user input
if "," not in image_urls_str:
image_urls.append(image_urls_str)
else:
for url in image_urls_str.split(","):
if "http:/" in url.lower() or "https:/" in url.lower():
image_urls.append(url)
with st.spinner("Analyzing the product... This may take a moment."):
product_info_raw = extract_data_from_product_image(image_urls, data_extractor_url)
print(f"DEBUG product_info_raw from image : {product_info_raw}")
if 'error' not in json.loads(product_info_raw).keys():
final_analysis = analyze_product(product_info_raw, system_prompt)
return [], final_analysis
else:
return [], f"Product information could not be extracted from the image because of {json.loads(product_info_raw)['error']}"
else:
return [], "I'm here to analyze food products. Please provide an image URL (Example : http://example.com/image.jpg) or product name (Example : Harvest Gold Bread)"
class SessionState:
"""Handles all session state variables in a centralized way"""
@staticmethod
def initialize():
initial_states = {
"messages": [],
"product_selected": False,
"product_shared": False,
"analyze_more": True,
"welcome_shown": False,
"yes_no_choice": None,
"welcome_msg": "Welcome to ConsumeWise! What product would you like me to analyze today?",
"system_prompt": "",
"similar_products": [],
"awaiting_selection": False,
"current_user_input": "",
"selected_product": None
}
for key, value in initial_states.items():
if key not in st.session_state:
st.session_state[key] = value
class SystemPromptManager:
"""Manages the system prompt input and related functionality"""
@staticmethod
def render_sidebar():
st.sidebar.header("System Prompt")
system_prompt = st.sidebar.text_area(
"Enter your system prompt here (required):",
value=st.session_state.system_prompt,
height=150,
key="system_prompt_input"
)
if st.sidebar.button("Submit Prompt"):
if system_prompt.strip():
st.session_state.system_prompt = system_prompt
SessionState.initialize() # Reset all states
st.rerun()
else:
st.sidebar.error("Please enter a valid system prompt.")
return system_prompt.strip()
class ProductSelector:
"""Handles product selection logic"""
@staticmethod
def handle_selection():
if st.session_state.similar_products:
# Create a container for the selection UI
selection_container = st.container()
with selection_container:
# Radio button for product selection
choice = st.radio(
"Select a product:",
st.session_state.similar_products + ["None of the above"],
key="product_choice"
)
# Confirm button
confirm_clicked = st.button("Confirm Selection")
msg = ""
# Only process the selection when confirm is clicked
if confirm_clicked:
st.session_state.awaiting_selection = False
if choice != "None of the above":
#st.session_state.selected_product = choice
st.session_state.messages.append({"role": "assistant", "content": f"You selected {choice}"})
_, msg = chatbot_response("", choice.split(" by ")[0], "", st.session_state.system_prompt, extract_info=True)
#Check if analysis couldn't be done because db had incomplete information
if msg != "product not found because product information in the db is corrupt":
#Only when msg is acceptable
st.session_state.messages.append({"role": "assistant", "content": msg})
with st.chat_message("assistant"):
st.markdown(msg)
st.session_state.product_selected = True
keys_to_keep = ["system_prompt", "messages", "welcome_msg"]
keys_to_delete = [key for key in st.session_state.keys() if key not in keys_to_keep]
for key in keys_to_delete:
del st.session_state[key]
st.session_state.welcome_msg = "What product would you like me to analyze next?"
if choice == "None of the above" or msg == "product not found because product information in the db is corrupt":
st.session_state.messages.append(
{"role": "assistant", "content": "Please provide the image URL of the product to analyze based on the latest information."}
)
with st.chat_message("assistant"):
st.markdown("Please provide the image URL of the product to analyze based on the latest information.")
#st.session_state.selected_product = None
st.rerun()
# Prevent further chat input while awaiting selection
return True # Indicates selection is in progress
return False # Indicates no selection in progress
class ChatManager:
"""Manages chat interactions and responses"""
@staticmethod
def process_response(user_input):
if not st.session_state.product_selected:
if "http:/" not in user_input and "https:/" not in user_input:
response, status = ChatManager._handle_product_name(user_input)
else:
response, status = ChatManager._handle_product_url(user_input)
return response, status
@staticmethod
def _handle_product_name(user_input):
st.session_state.product_shared = True
st.session_state.current_user_input = user_input
similar_products, _ = chatbot_response(
"", user_input, data_extractor_url,
st.session_state.system_prompt, extract_info=False
)
if len(similar_products) > 0:
st.session_state.similar_products = similar_products
st.session_state.awaiting_selection = True
return "Here are some similar products from our database. Please select:", "no success"
return "Product not found in our database. Please provide the image URL of the product.", "no success"
@staticmethod
def _handle_product_url(user_input):
is_valid_url = (".jpeg" in user_input or ".jpg" in user_input) and \
("http:/" in user_input or "https:/" in user_input)
if not st.session_state.product_shared:
return "Please provide the product name first"
if is_valid_url and st.session_state.product_shared:
_, msg = chatbot_response(
user_input, "", data_extractor_url,
st.session_state.system_prompt, extract_info=True
)
st.session_state.product_selected = True
if msg != "product not found because image is not clear" and "Product information could not be extracted from the image" not in msg:
response = msg
status = "success"
elif msg == "product not found because image is not clear":
response = msg + ". Please share clear image URLs!"
status = "no success"
else:
response = msg + ".Please re-try!!"
status = "no success"
return response, status
return "Please provide valid image URL of the product.", "no success"
def main():
#Initialize session state
SessionState.initialize()
# Display title
st.title("ConsumeWise - Your Food Product Analysis Assistant")
# Handle system prompt
system_prompt = SystemPromptManager.render_sidebar()
if not system_prompt:
st.warning("⚠️ Please enter a system prompt in the sidebar before proceeding.")
st.chat_input("Enter your message:", disabled=True)
return
# Show welcome message
if not st.session_state.welcome_shown:
st.session_state.messages.append({
"role": "assistant",
"content": st.session_state.welcome_msg
})
st.session_state.welcome_shown = True
# Display chat history
for message in st.session_state.messages:
with st.chat_message(message["role"]):
st.markdown(message["content"])
# Handle product selection if awaiting
selection_in_progress = False
if st.session_state.awaiting_selection:
selection_in_progress = ProductSelector.handle_selection()
# Only show chat input if not awaiting selection
if not selection_in_progress:
user_input = st.chat_input("Enter your message:", key="user_input")
if user_input:
# Add user message to chat
st.session_state.messages.append({"role": "user", "content": user_input})
with st.chat_message("user"):
st.markdown(user_input)
# Process response
response, status = ChatManager.process_response(user_input)
st.session_state.messages.append({"role": "assistant", "content": response})
with st.chat_message("assistant"):
st.markdown(response)
if status == "success":
SessionState.initialize() # Reset states for next product
#st.session_state.welcome_msg = "What is the next product you would like me to analyze today?"
keys_to_keep = ["system_prompt", "messages", "welcome_msg"]
keys_to_delete = [key for key in st.session_state.keys() if key not in keys_to_keep]
for key in keys_to_delete:
del st.session_state[key]
st.session_state.welcome_msg = "What product would you like me to analyze next?"
#elif response: # Only add response if it's not None
# print(f"DEBUG : st.session_state.awaiting_selection : {st.session_state.awaiting_selection}")
# print(f"response : {response}")
st.rerun()
else:
# Disable chat input while selection is in progress
st.chat_input("Please confirm your selection above first...", disabled=True)
# Clear chat history button
if st.button("Clear Chat History"):
st.session_state.clear()
st.rerun()
if __name__ == "__main__":
main()