Update app.py

app.py

@@ -1,63 +1,78 @@
-from transformers import AutoModelForSequenceClassification, AutoTokenizer
-import torch
 import gradio as gr
+from transformers import pipeline
 import matplotlib.pyplot as plt
+from wordcloud import WordCloud
+import json
+import os
 
-# Use a better model for sentiment analysis (Supports Neutral)
-MODEL_NAME = "cardiffnlp/twitter-roberta-base-sentiment"
+# Load sentiment analysis model
+sentiment_pipeline = pipeline("sentiment-analysis")
 
-# Load tokenizer and model
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
-model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME)
+# Store last input/output using localStorage
+STATE_FILE = "state.json"
 
-# Labels for classification
-LABELS_MAP = {0: "Negative", 1: "Neutral", 2: "Positive"}
+def save_state(text, sentiment, confidence):
+    with open(STATE_FILE, "w") as f:
+        json.dump({"text": text, "sentiment": sentiment, "confidence": confidence}, f)
 
-# Function to analyze sentiment
+def load_state():
+    if os.path.exists(STATE_FILE):
+        with open(STATE_FILE, "r") as f:
+            return json.load(f)
+    return {"text": "", "sentiment": "", "confidence": ""}
+
+# Function to analyze sentiment in real-time
 def analyze_sentiment(text):
-    inputs = tokenizer(text, return_tensors="pt")
-    outputs = model(**inputs)
+    if not text.strip():
+        return "Enter text to analyze.", "", None, ""
+    
+    result = sentiment_pipeline(text)[0]
+    sentiment, confidence = result['label'], result['score']
     
-    scores = torch.nn.functional.softmax(outputs.logits, dim=-1)
-    predicted_label = torch.argmax(scores).item()
-    sentiment = LABELS_MAP[predicted_label]
-    confidence = scores[0, predicted_label].item()
-
-    # Sentiment mapping for UI
     sentiment_map = {
-        "Positive": ("🟢 Positive 😊", "green"),
-        "Negative": ("🔴 Negative 😠", "red"),
-        "Neutral": ("🟡 Neutral 😐", "orange")
+        "POSITIVE": ("🟢 Positive 😊", "green"),
+        "NEGATIVE": ("🔴 Negative 😠", "red"),
+        "NEUTRAL": ("🟡 Neutral 😐", "orange")
     }
     
-    sentiment_label, color = sentiment_map[sentiment]
-
-    # Generate Confidence Score Bar Chart
-    fig, ax = plt.subplots(figsize=(4, 2))
-    ax.bar(sentiment, confidence, color=color)
-    ax.set_ylim([0, 1])
-    ax.set_ylabel("Confidence Score")
-    ax.set_title("Sentiment Confidence")
-
-    return sentiment_label, f"Confidence: {confidence:.2f}", fig
+    sentiment_label, color = sentiment_map.get(sentiment.upper(), ("⚪ Unknown ❓", "gray"))
+    
+    # Generate Word Cloud
+    wordcloud = WordCloud(width=400, height=200, background_color="white").generate(text)
+    fig, ax = plt.subplots()
+    ax.imshow(wordcloud, interpolation='bilinear')
+    ax.axis("off")
+    
+    save_state(text, sentiment_label, confidence)
+    
+    return sentiment_label, f"Confidence: {confidence:.2%}", fig, text
 
-# Gradio UI
+# UI Layout
 with gr.Blocks(theme=gr.themes.Soft()) as iface:
-    gr.Markdown("# 📢 Text Sentiment Analyzer")
-    gr.Markdown("Analyze the sentiment of your text input and visualize the confidence score.")
-
+    gr.Markdown("# 🎭 AI Sentiment Analyzer")
+    gr.Markdown("Analyze sentiment in real-time with enhanced visualization.")
+    
     with gr.Row():
-        text_input = gr.Textbox(lines=2, placeholder="Enter text here...", label="Your Input")
-
+        dark_mode = gr.Checkbox(label="🌙 Dark Mode")
+        
+    with gr.Row():
+        text_input = gr.Textbox(lines=3, placeholder="Type your text here...", label="Your Input", live=True)
+    
     analyze_button = gr.Button("Analyze Sentiment ✨")
-
+    reset_button = gr.Button("🔄 Reset")
+    
     with gr.Row():
-        sentiment_output = gr.Textbox(label="Sentiment Result", interactive=False)
-        confidence_output = gr.Textbox(label="Confidence Score", interactive=False)
+        sentiment_output = gr.Label(label="Sentiment Result")
+        confidence_output = gr.Label(label="Confidence Score")
     
-    chart_output = gr.Plot(label="Confidence Score Chart")
-
-    analyze_button.click(analyze_sentiment, inputs=text_input, outputs=[sentiment_output, confidence_output, chart_output])
+    wordcloud_output = gr.Plot(label="Word Cloud Visualization")
+    
+    # Load previous session state
+    state = load_state()
+    text_input.value, sentiment_output.value, confidence_output.value = state['text'], state['sentiment'], state['confidence']
+    
+    analyze_button.click(analyze_sentiment, inputs=text_input, outputs=[sentiment_output, confidence_output, wordcloud_output, text_input])
+    reset_button.click(lambda: ("", "", None, ""), inputs=[], outputs=[sentiment_output, confidence_output, wordcloud_output, text_input])
 
 # Launch the app
 if __name__ == "__main__":