app.py

import streamlit as st
#import google.generativeai as genai
import google-generativeai as genai
import os
from PIL import Image
import cv2
from io import BytesIO
import base64
from dotenv import load_dotenv
import numpy as np
from deepface import DeepFace  # Replacing FER with DeepFace
print("DeepFace is installed and ready to use!")
#import google.generativeai as genai
i#mport google.generativeai as genai

# Test if the module can be imported successfully
print("Google Generative AI module is successfully imported!")



load_dotenv()

genai.configure(api_key=("AIzaSyBP1kQr-80Aq_K5_9AVgD1MLJqs05Cg20Q"))

# gemini function for general content generation
def get_gemini_response(input):
    try:
        model = genai.GenerativeModel('gemini-pro')
        response = model.generate_content(input)
        return response
    except Exception as e:
        st.error(f"Error: {e}")
        return None

# Function to analyze image for depression and emotion detection using DeepFace
def detect_emotions(image):
    try:
        # Use DeepFace to analyze emotions
        analysis = DeepFace.analyze(image, actions=['emotion'], enforce_detection=False)
        # Return the dominant emotion and its score
        return analysis[0]['dominant_emotion'], analysis[0]['emotion']
    except Exception as e:
        st.error(f"Error during emotion detection: {e}")
        return None, None

# Function to analyze detected emotions with LLM
def analyze_emotions_with_llm(emotion, emotions):
    emotion_analysis = f"{emotion}: {emotions[emotion]:.2f}"
    
    analysis_prompt = f"""
    ### As a mental health and emotional well-being expert, analyze the following detected emotions.
    ### Detected Emotions:
    {emotion_analysis}
    ### Analysis Output:
    1. Identify any potential signs of depression based on the detected emotions.
    2. Explain the reasoning behind your identification.
    3. Provide recommendations for addressing any identified issues.
    """
    response = get_gemini_response(analysis_prompt)
    return response

# Function to capture live video frame for analysis
def capture_video_frame():
    video_capture = cv2.VideoCapture(0)
    if not video_capture.isOpened():
        st.error("Failed to access the webcam. Ensure you have allowed camera access in your browser.")
        return None
    ret, frame = video_capture.read()
    video_capture.release()
    if ret:
        frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        return Image.fromarray(frame_rgb)
    else:
        st.error("Failed to capture a frame from the webcam.")
        return None

# Function to parse and display response content
def display_response_content(response):
    st.subheader("Response Output")
    if response and response.candidates:
        response_content = response.candidates[0].content.parts[0].text if response.candidates[0].content.parts else ""
        sections = response_content.split('###')
        for section in sections:
            if section.strip():
                section_lines = section.split('\n')
                section_title = section_lines[0].strip()
                section_body = '\n'.join(line.strip() for line in section_lines[1:] if line.strip())
                if section_title:
                    st.markdown(f"**{section_title}**")
                if section_body:
                    st.write(section_body)
    else:
        st.write("No response received from the model or quota exceeded.")

## Streamlit App
st.title("AI-Powered Depression and Emotion Detection System")
st.text("Use the AI system for detecting depression and emotions from images and live video.")

# Tabs for different functionalities
tab1, tab2 = st.tabs(["Image Analysis", "Live Video Analysis"])

with tab1:
    st.header("Image Analysis")
    uploaded_file = st.file_uploader("Upload an image for analysis", type=["jpg", "jpeg", "png"], help="Please upload an image file.")
    submit_image = st.button('Analyze Image')
    
    if submit_image:
        if uploaded_file is not None:
            image = Image.open(uploaded_file)
            emotion, emotions = detect_emotions(image)
            if emotion:
                response = analyze_emotions_with_llm(emotion, emotions)
                # Parse and display response in a structured way
                display_response_content(response)
            else:
                st.write("No emotions detected in the image.")

with tab2:
    st.header("Live Video Analysis")
    capture_frame = st.button('Capture and Analyze Frame')
    
    if capture_frame:
        image = capture_video_frame()
        if image is not None:
            emotion, emotions = detect_emotions(image)
            if emotion:
                response = analyze_emotions_with_llm(emotion, emotions)
                # Parse and display response in a structured way
                display_response_content(response)
            else:
                st.write("No emotions detected in the video frame.")
        else:
            st.write("Failed to capture video frame.")