Update app.py

app.py

@@ -7,18 +7,12 @@ sentiment_pipeline = pipeline("sentiment-analysis", model="nlptown/bert-base-mul
 # Define a function for sentiment analysis
 def analyze_sentiment(text):
     result = sentiment_pipeline(text)
+    label = result[0]['label']  # Get the label, e.g., "1 star", "5 stars"
+    score = int(label.split()[0])  # Extract the numerical part of the label (e.g., 1, 2, ..., 5)
     
-    # Map the label to a sentiment score between -1 and 1
-    label = result[0]['label']
-    score = result[0]['score']
-    
-    # Convert label to sentiment score
-    if label == 'negative':
-        sentiment_score = -score  # Map negative to -1 to 0
-    elif label == 'neutral':
-        sentiment_score = 0.0      # Neutral is 0
-    elif label == 'positive':
-        sentiment_score = score     # Map positive to 0 to 1
+    # Map the score from 1-5 to -1 to 1:
+    # 1 star -> -1.0, 2 stars -> -0.5, 3 stars -> 0.0, 4 stars -> 0.5, 5 stars -> 1.0
+    sentiment_score = (score - 3) / 2
     
     return sentiment_score
 
@@ -38,7 +32,13 @@ interface = gr.Interface(
     outputs=gr.Number(label="Sentiment Score (-1 to 1)"),
     examples=examples,
     title="Sentiment Analysis with Hugging Face Transformers",
-    description="Enter a sentence to analyze its sentiment using a pre-trained model from Hugging Face."
+    description="""
+        Enter a sentence to analyze its sentiment. The model will output a sentiment score between -1 and 1:
+        - **-1**: Extremely Negative  
+        - **0**: Neutral  
+        - **+1**: Extremely Positive  
+        This uses the `nlptown/bert-base-multilingual-uncased-sentiment` model.
+    """
 )
 
 # Launch the Gradio app