updated app.py code

App/app.py

@@ -70,120 +70,23 @@ def load_models():
         return None
     return models
 
-# Load the model and tokenizer
-tokenizer, model = load_model()
+# Load all models at startup
+models = load_models()
 
-# --- UI and Prediction Functions ---
-
-# Dictionary for mapping emotions to emojis
-emotions_emoji_dict = {
-    "anger": "😠", "disgust": "🤮", "fear": "😨", "joy": "😂",
-    "neutral": "😐", "sadness": "😔", "shame": "😳", "surprise": "😮"
-}
-
-def predict_emotions(docx):
-    """
-    Predicts the emotion of a single text string using the loaded BERT model.
-    """
-    if model is None or tokenizer is None:
-        return "Model not loaded"
-        
-    inputs = tokenizer(docx, return_tensors="pt", padding=True, truncation=True, max_length=128)
-    
-    with torch.no_grad():
-        logits = model(**inputs).logits
-        
-    predicted_class_id = torch.argmax(logits, dim=1).item()
-    return model.config.id2label[predicted_class_id]
-
-def get_prediction_proba(docx):
-    """
-    Gets the prediction probabilities for each emotion for a single text string.
-    """
-    if model is None or tokenizer is None:
-        return np.array([])
-
-    inputs = tokenizer(docx, return_tensors="pt", padding=True, truncation=True, max_length=128)
-    
-    with torch.no_grad():
-        logits = model(**inputs).logits
-    
-    probabilities = torch.nn.functional.softmax(logits, dim=1)
-    return probabilities.numpy()
-
-# --- Main Application ---
 
 def main():
-    """
-    The main function that runs the Streamlit application.
-    """
-    # --- KEY CHANGE: Reverted to the older 'use_column_width' argument ---
-    st.image('https://i.ibb.co/rsbYCsN/senttext-low-resolution-logo-white-on-black-background.png', use_column_width=True)
-    # --- End of Change ---
+    st.sidebar.image('https://i.ibb.co/rsbYCsN/senttext-low-resolution-logo-white-on-black-background.png', use_column_width=True)
     
-    st.title("SentText Emotion Analyzer")
-    st.subheader("Analyze the emotional tone of your text with a fine-tuned BERT model.")
-
     with st.sidebar:
         choice = option_menu(
-            "Menu", ["Home", "About"],
-            icons=['house', 'info-circle'],
-            menu_icon="cast",
-            default_index=0
+            "Menu", ["Emotion Analyzer", "Financial Analysis", "About"],
+            icons=['emoji-smile', 'bank', 'info-circle'],
+            menu_icon="cast", default_index=0
         )
 
-    if choice == "Home":
-        if model is not None:
-            with st.form(key='emotion_clf_form'):
-                raw_text = st.text_area("Type your text here...", height=150)
-                submit_text = st.form_submit_button(label='Analyze')
-
-            if submit_text:
-                if raw_text.strip() == "":
-                    st.warning("Please enter some text to analyze.")
-                else:
-                    col1, col2 = st.columns(2)
-
-                    prediction = predict_emotions(raw_text)
-                    probability = get_prediction_proba(raw_text)
-
-                    with col1:
-                        st.success("Original Text")
-                        st.write(raw_text)
-
-                        st.success("Prediction Probability")
-                        labels = list(model.config.id2label.values())
-                        proba_df = pd.DataFrame(probability, columns=labels)
-                        st.write(proba_df.T.rename(columns={0: 'Probability'}))
-
-                    with col2:
-                        st.success("Prediction")
-                        emoji_icon = emotions_emoji_dict.get(prediction, "🙂")
-                        st.metric(label="Predicted Emotion", value=f"{prediction.capitalize()} {emoji_icon}")
-                        st.metric(label="Confidence", value=f"{np.max(probability):.4f}")
-                        
-                        proba_df_clean = proba_df.T.reset_index()
-                        proba_df_clean.columns = ["emotions", "probability"]
-                        
-                        fig = alt.Chart(proba_df_clean).mark_bar().encode(
-                            x=alt.X('emotions', sort=None),
-                            y='probability',
-                            color='emotions',
-                            tooltip=['emotions', 'probability']
-                        ).properties(
-                            title="Emotion Probabilities"
-                        )
-                        st.altair_chart(fig, use_container_width=True)
-        else:
-            st.info("The model is not available. Please check the error messages above.")
-
-    else: # About page
-        st.header("About SentText")
-        st.markdown("""
-        This application provides nuanced emotional analysis that goes beyond simple positive or negative sentiment. By classifying text into granular categories such as joy, anger, sadness, and fear - SentText offers a deeper understanding of the underlying tone and intent of a message.
-
-        The analysis is powered by a DistilBERT model, a streamlined and efficient version of the powerful BERT language model. This model has been specifically fine-tuned on a large dataset of emotional texts to accurately identify and differentiate between various human emotions. The result is a tool that provides fast, reliable, and insightful analysis of textual communication.
-        """)
-
-if __name__ == '__main__':
-    main()
+    if choice == "Emotion Analyzer":
+        render_emotion_analyzer()
+    elif choice == "Financial Analysis":
+        render_financial_analyzer()
+    elif choice == "About":
+        render_about_page()