Update app.py

app.py

@@ -1,54 +1,47 @@
 import streamlit as st
-from transformers import pipeline
-from textblob import TextBlob
-from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
+import transformers
+import torch
 
+# Define sentiment analysis models
 models = {
-    'DistilBERT': pipeline('sentiment-analysis', model='distilbert-base-uncased-finetuned-sst-2-english'),
-    'TextBlob': None,
-    'VADER': None
+    "DistilBERT": transformers.pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english"),
+    "BERT": transformers.pipeline("sentiment-analysis", model="bert-base-uncased-finetuned-sst-2-english"),
+    "RoBERTa": transformers.pipeline("sentiment-analysis", model="roberta-base-openai-detector"),
 }
 
-def textblob_sentiment(text):
-    polarity = TextBlob(text).sentiment.polarity
-    if polarity > 0.2:
-        return 'Positive'
-    elif polarity < -0.2:
-        return 'Negative'
-    else:
-        return 'Neutral'
+# Define function to analyze sentiment using selected model
+def analyze_sentiment(text, model_name):
+    model = models[model_name]
+    result = model(text)[0]
+    return result['label'], result['score']
 
-def vader_sentiment(text):
-    analyzer = SentimentIntensityAnalyzer()
-    scores = analyzer.polarity_scores(text)
-    if scores['compound'] > 0.5:
-        return 'Positive'
-    elif scores['compound'] < -0.5:
-        return 'Negative'
-    else:
-        return 'Neutral'
+# Define Streamlit app
+def app():
+    st.title("Sentiment Analysis App")
 
-models['TextBlob'] = textblob_sentiment
-models['VADER'] = vader_sentiment
+    # User input
+    text = st.text_area("Enter text to analyze", max_chars=1024)
 
-st.title('Sentiment Analysis App')
+    # Sentiment analysis
+    if st.button("Analyze"):
+        st.write("Analyzing sentiment...")
+        with st.spinner("Wait for it..."):
+            results = []
+            for model_name in models:
+                label, score = analyze_sentiment(text, model_name)
+                results.append((model_name, label, score))
+        st.success("Sentiment analysis complete!")
+        st.write("Results:")
+        for model_name, label, score in results:
+            st.write(f"- {model_name}: {label} ({score:.2f})")
 
-text = st.text_input('Enter some text')
-model_name = st.selectbox('Select a model', list(models.keys()))
+    # Advanced features
+    if st.beta_expander("Advanced Options", expanded=False):
+        model_name = st.selectbox("Select model", list(models.keys()))
+        if st.button("Test model"):
+            label, score = analyze_sentiment("This is a test.", model_name)
+            st.write(f"Test result: {label} ({score:.2f})")
 
-if st.button('Predict'):
-    if model_name == 'TextBlob' or model_name == 'VADER':
-        sentiment = models[model_name](text)
-        st.write(f'Sentiment: {sentiment}')
-    else:
-        model = models[model_name]
-        results = model(text, return_all_scores=True)
-        st.write('Results:')
-        for result in results:
-            label = result['label']
-            score = result['score']
-            if score > 0.2:
-                st.write(f'{label} (Score: {score:.2f})')
-        mixed_score = 1 - sum([r['score'] for r in results])
-        if mixed_score > 0.2:
-            st.write(f'Mixed (Score: {mixed_score:.2f})')
+# Run Streamlit app
+if __name__ == "__main__":
+    app()