import streamlit as st import cv2 import numpy as np from tensorflow.keras.models import load_model import google.generativeai as genai import os from dotenv import load_dotenv load_dotenv() # Configure Genai Key genai.configure(api_key=os.getenv('GOOGLE_API_KEY')) # Initialize Streamlit app st.set_page_config(page_title="Recipe Generator App", page_icon="🍔", layout="centered") st.header("Recipe Generator") model_chat = genai.GenerativeModel("gemini-pro") chat = model_chat.start_chat(history=[]) if 'chat_history' not in st.session_state: st.session_state['chat_history'] = [] # Function to get a response from Gemini with the specified context and question def get_gemini_response(prompt): response = chat.send_message(prompt) return response # Function to get an answer to the user's question about the recipe def get_answer(question, recipe): # Create a prompt to get an answer to the question about the recipe prompt = f"Here is a recipe: {recipe}\n\n{question}. Answer in the friendly manner. The response should not contain ** at the beginning and end and also do not bold any text" response = get_gemini_response(prompt) return response.text def get_gemini_response_for_receipe(question, prompt): model = genai.GenerativeModel('gemini-pro') response = model.generate_content([prompt, question]) return response.text # Load the pre-trained model model_apna = load_model('my_model.h5') # Define class labels class_labels = ['Biryani', 'Burger', 'Dosa', 'Idly', 'Pizza'] def preprocess_image(image): resized_image = cv2.resize(image, (150, 150)) normalized_image = resized_image / 255.0 input_image = np.expand_dims(normalized_image, axis=0) return input_image def predict_class(image): input_image = preprocess_image(image) predictions = model_apna.predict(input_image) predicted_class_index = np.argmax(predictions, axis=1)[0] predicted_class_label = class_labels[predicted_class_index] return predicted_class_label def get_recipe(dish_name): prompt = f"Please provide a recipe for {dish_name}." response = get_gemini_response_for_receipe(dish_name, prompt) return response uploaded_image = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"]) # hard coding the image for now # uploaded_image = open("1.jpg", "rb") if uploaded_image is not None: image = cv2.imdecode(np.frombuffer(uploaded_image.read(), np.uint8), 1) # # Convert BGR image to RGB rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) st.image(rgb_image, caption='Uploaded Image', width=300) predicted_class = predict_class(image) st.success(f'The predicted recipe category is: {predicted_class}') # Get the recipe for the predicted dish category recipe = get_recipe(predicted_class) if recipe: st.subheader("Recipe:") st.write(recipe) user_question = st.text_input("You: ") if st.button("Ask"): # st.write("User question:", user_question) # Check user question # st.write("Recipe:", recipe) # Check recipe response = get_answer(user_question, recipe) if response: st.write(f"AI : {response}") else: st.write("Gemini response is empty") # if uploaded_image is not None: # image = cv2.imdecode(np.frombuffer(uploaded_image.read(), np.uint8), 1) # # Convert BGR image to RGB # rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) # st.image(rgb_image, caption='Uploaded Image', width=300) # if st.button('Predict'): # predicted_class = predict_class(image) # st.success(f'The predicted recipe category is: {predicted_class}') # # Get the recipe for the predicted dish category # recipe = get_recipe(predicted_class) # if recipe: # st.subheader("Recipe:") # st.write(recipe) # user_question = st.text_input("You: ") # if user_question.strip() != "" and st.button("Ask"): # print("User question:", user_question) # Check user question # print("Recipe:", recipe) # Check recipe # response = get_answer(user_question, recipe) # st.write(f"AI : {response}") # print("Gemini response:", response) # Check Gemini response