app.py

import streamlit as st
from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

# Load pre-trained emotion detection model
model_name = "j-hartmann/emotion-english-distilroberta-base"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# Get the actual emotion labels from the model
emotion_labels = model.config.id2label

# Emotion styles (emoji + colors)
emotion_styles = {
    "joy": {"emoji": "😃", "color": "#FFD700"},        # Gold
    "sadness": {"emoji": "😢", "color": "#3498DB"},    # Blue
    "anger": {"emoji": "😡", "color": "#8E44AD"},      # Red
    "fear": {"emoji": "😨", "color": "#E74C3C"},       # Purple
    "surprise": {"emoji": "😲", "color": "#F1C40F"},   # Yellow
    "disgust": {"emoji": "🤢", "color": "#2ECC71"},    # Green
    "neutral": {"emoji": "😐", "color": "#95A5A6"}     # Gray
}

# Streamlit UI
st.set_page_config(page_title="Emotion Detection", layout="centered")

# 🎨 Change Background Color to Black
st.markdown(
    """
    <style>
    body {
        background-color: #000000; /* Black Background */
    }
    </style>
    """,
    unsafe_allow_html=True
)

# 🎯 Add Image at the Top
st.image("innomatics_logo.png", use_container_width=True, width=300)  # Replace with your image filename

# 🏷️ Title
st.markdown("<h1 style='text-align: center; color: white;'>🔍 Emotion Detection 😊</h1>", unsafe_allow_html=True)

# 📌 **Business Context**
st.markdown(
    """
    <h3 style='color: #00FFFF;'> Business Problem</h3>
    <p style='text-align: justify; font-size: 16px; color: ##1E90FF;'>
        Organizations aim to enhance customer experiences, mental health support, and marketing strategies 
        by analyzing human emotions from text, speech, and facial expressions. However, accurately detecting 
        emotions remains a challenge due to their subjective nature.
    </p>
    
    <h3 style='color: #00FFFF;'> Business Objective</h3>
    <p style='text-align: justify; font-size: 16px; color: #D3D3D3;'>
        The project aims to create a Machine Learning-based Emotion Detection System that can accurately 
        identify human emotions. This will aid businesses in analyzing customer sentiment, enhancing 
        user engagement, and improving real-time decision-making.
    </p>
    
    <h3 style='color: #00FFFF;'> Business Constraints</h3>
    <p style='text-align: justify; font-size: 16px; color: #D3D3D3;'>
        The model should provide high accuracy in detecting emotions, but complex models may be difficult 
        for business users to interpret. Real-time processing and latency are crucial for applications 
        like chatbots and customer support. Data quality and availability are essential, with a balanced 
        dataset and handling of missing or noisy data. The model should be scalable, ethical, and 
        privacy-conscious, integrating seamlessly with existing business tools. Cost constraints include 
        compute costs for training and cloud deployment. The model should be robust and context-aware, 
        working across multiple platforms.
    </p>
    """,
    unsafe_allow_html=True
)

st.markdown("<h3 style='text-align: left; color: #00FFFF;'>📝 Enter the text:</h3>", unsafe_allow_html=True)

# User Input
user_text = st.text_input("", placeholder="Type your text here...")

if st.button("Submit"):
    if user_text:
        # Tokenize input text
        inputs = tokenizer(user_text, return_tensors="pt")

        # Get model predictions
        with torch.no_grad():
            outputs = model(**inputs)

        # Get the predicted emotion
        scores = outputs.logits[0]
        predicted_label_id = torch.argmax(scores).item()
        predicted_emotion = emotion_labels[predicted_label_id].lower()

        # Get emoji & color
        emotion_data = emotion_styles.get(predicted_emotion, {"emoji": "😐", "color": "#95A5A6"})
        emoji_display = emotion_data["emoji"]
        text_color = emotion_data["color"]

        # Display Results with Color
        st.markdown(
            f"""
            <div style="text-align: center; padding: 10px; border-radius: 10px; background-color: {text_color}; color: white; font-size: 24px;">
                <b>Detected Emotion:</b> {predicted_emotion.capitalize()} {emoji_display}
            </div>
            """,
            unsafe_allow_html=True
        )
    else:
        st.warning("Please enter some text!")