update

app.py

@@ -4,12 +4,13 @@ import openai
 import streamlit as st
 from PIL import Image
 from dotenv import load_dotenv
-import time
+import torch
+from transformers import AutoProcessor, AutoModelForImageClassification
 
 # Load environment variables from .env file
 load_dotenv()
 
-# Set up the Hugging Face API URL and your API key
+# Set up the Hugging Face API for emotion detection
 emotion_model_url = "https://api-inference.huggingface.co/models/trpakov/vit-face-expression"
 headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_KEY')}"}
 
@@ -17,35 +18,33 @@ headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_KEY')}"}
 openai.api_key = os.getenv('OPENAI_API_KEY')
 
 # Function to query the facial expression recognition model
-def query_emotion(filename):
-    with open(filename, "rb") as f:
-        data = f.read()
-    response = requests.post(emotion_model_url, headers=headers, data=data)
+def query_emotion(image):
+    # Load the processor and model
+    processor = AutoProcessor.from_pretrained("trpakov/vit-face-expression")
+    model = AutoModelForImageClassification.from_pretrained("trpakov/vit-face-expression")
+
+    # Preprocess the image
+    inputs = processor(images=image, return_tensors="pt")
+
+    # Perform inference
+    with torch.no_grad():
+        outputs = model(**inputs)
     
-    if response.status_code == 200:
-        return response.json()
-    else:
-        st.error("Error detecting facial expression: " + response.text)
-        return None
+    # Get predicted class index
+    logits = outputs.logits
+    predicted_class_idx = torch.argmax(logits, dim=-1).item()
+
+    # Get the predicted label
+    predicted_label = processor.decode(predicted_class_idx)
+    return predicted_label
 
 # Function to generate a response using OpenAI based on detected emotion
 def generate_text_based_on_mood(emotion):
     try:
         # Create a dynamic prompt based on the detected emotion
-        if emotion == "happy":
-            prompt = "Give a motivational quote to celebrate happiness."
-        elif emotion == "sad":
-            prompt = "Provide a comforting message for someone feeling sad."
-        elif emotion == "angry":
-            prompt = "Suggest a way to calm down someone feeling angry."
-        elif emotion == "fear":
-            prompt = "Give an encouraging message for someone feeling fearful."
-        elif emotion == "surprised":
-            prompt = "Offer a fun fact or light-hearted comment for someone feeling surprised."
-        elif emotion == "neutral":
-            prompt = "Provide a general motivational quote."
+        prompt = f"Generate a light-hearted joke or motivational message for someone who is feeling {emotion}."
         
-        # Call OpenAI's API using the new interface
+        # Call OpenAI's API using GPT-4
         response = openai.ChatCompletion.create(
             model="gpt-4",  # Specify the GPT-4 model
             messages=[
@@ -85,25 +84,18 @@ if uploaded_file is not None:
     image = Image.open(uploaded_file)
     st.image(image, caption='Uploaded Image', use_column_width=True)
 
-    # Save the uploaded file temporarily
-    with open("uploaded_image.jpg", "wb") as f:
-        f.write(uploaded_file.getbuffer())
-
     # Detect facial expression
-    expression_output = query_emotion("uploaded_image.jpg")
-    if expression_output:
-        # Assuming the response has a 'label' field with the detected emotion
-        emotion = expression_output[0]['label']  # Adjust based on response structure
-        st.write(f"Detected emotion: {emotion}")
-
-        # Generate text based on detected emotion
-        message = generate_text_based_on_mood(emotion)
-        st.write("Here's something to remind you :")
-        st.write(message)
-
-        # Convert the generated message to audio
-        audio_file = text_to_speech(message)
-
-        # Provide an audio player in the Streamlit app if audio file exists
-        if audio_file:
-            st.audio(audio_file)  # Streamlit will handle playback
+    emotion = query_emotion(image)
+    st.write(f"Detected emotion: {emotion}")
+
+    # Generate text based on detected emotion
+    message = generate_text_based_on_mood(emotion)
+    st.write("Here's something to cheer you up:")
+    st.write(message)
+
+    # Convert the generated message to audio
+    audio_file = text_to_speech(message)
+
+    # Provide an audio player in the Streamlit app if audio file exists
+    if audio_file:
+        st.audio(audio_file)  # Streamlit will handle playback