Update app.py

app.py

@@ -1,19 +1,8 @@
 import torch
 import joblib
 import gradio as gr
-import matplotlib.pyplot as plt
-import numpy as np
-import shap
-import lime
-import lime.lime_text
-import logging
-import os
 
-# Configure logging for debugging
-LOG_FILE = "prediction_logs.txt"
-logging.basicConfig(filename=LOG_FILE, level=logging.INFO, format="%(asctime)s - %(message)s")
-
-# Load Model & Preprocessors
+# Load the Model and Dependencies
 tfidf_vectorizer = joblib.load("tfidf_vectorizer.pkl")
 category_encoder = joblib.load("category_encoder.pkl")
 team_encoder = joblib.load("team_encoder.pkl")
@@ -21,7 +10,7 @@ team_encoder = joblib.load("team_encoder.pkl")
 multi_label_classifier = torch.load("multi_label_classifier.pth")
 multi_label_classifier.eval()
 
-# Dummy function to get keywords per category
+# Function to get keywords (dummy implementation)
 def get_top_keywords_per_category(category, n=5):
     keywords_dict = {
         "UX Issue": ["mobile", "responsive", "alignment", "css", "layout"],
@@ -30,83 +19,31 @@ def get_top_keywords_per_category(category, n=5):
     }
     return keywords_dict.get(category, ["No keywords found"])
 
-# Function to explain predictions using SHAP
-def explain_with_shap(phrase):
-    background_data = ["UI button is not working", "Server error while processing", "Page alignment issue"]
-    feature_data = tfidf_vectorizer.transform(background_data + [phrase])
-    
-    explainer = shap.Explainer(multi_label_classifier, feature_data[:-1])
-    shap_values = explainer(feature_data[-1])
-
-    # Generate SHAP force plot
-    shap_html = shap.plots.text(shap_values, display=False)
-    return shap_html
-
-# Function to predict, explain, and generate logs
-def predict_with_visuals(phrase):
+# Prediction function
+def predict_with_keywords(phrase):
     text_features = tfidf_vectorizer.transform([phrase])
-    
-    predictions = multi_label_classifier.predict_proba(text_features)
-    predicted_labels = np.argmax(predictions, axis=1)
-
-    predicted_category = category_encoder.inverse_transform([predicted_labels[0]])[0]
-    predicted_team = team_encoder.inverse_transform([predicted_labels[0]])[0]
-    team_email = f"support@{predicted_team.replace(' ', '').lower()}.com"
+    predicted_labels = multi_label_classifier.predict(text_features)
 
+    predicted_category = category_encoder.inverse_transform([predicted_labels[0][0]])[0]
+    predicted_team = team_encoder.inverse_transform([predicted_labels[0][1]])[0]
+    team_email = "support@" + predicted_team.replace(" ", "").lower() + ".com"
     keywords = get_top_keywords_per_category(predicted_category, n=5)
 
-    # Create Bar Chart for Confidence Scores
-    category_names = category_encoder.classes_
-    category_probs = predictions[0]
-
-    fig, ax = plt.subplots(facecolor='#121212')
-    ax.barh(category_names, category_probs, color=["#ff9999", "#66b3ff", "#99ff99"])
-    ax.set_xlabel("Confidence Score", color="white")
-    ax.set_title("Prediction Confidence by Category", color="white")
-    ax.tick_params(colors="white")
-
-    # Log the prediction
-    log_entry = f"Input: {phrase} | Predicted Category: {predicted_category} | Confidence: {category_probs.max():.2f}"
-    logging.info(log_entry)
-
-    # SHAP explanation
-    shap_explanation = explain_with_shap(phrase)
-
-    result = f"""
-    <div style='font-size: 18px; font-family: Arial; color: white; background-color: #121212; padding: 10px; border-radius: 8px;'>
-    <strong>📌 Predicted Category:</strong> <span style='color:#4CAF50;'>{predicted_category}</span><br>
-    <strong>👨‍💻 Assigned Team:</strong> <span style='color:#2196F3;'>{predicted_team}</span><br>
-    <strong>📧 Team Email:</strong> <span style='color:#FF5722;'>{team_email}</span><br>
-    <strong>🔑 Top Keywords:</strong> <span style='color:#FFEB3B;'>{', '.join(keywords)}</span>
-    </div>
+    return f"""
+    **Predicted Category:** {predicted_category}  
+    **Predicted Assigned Team:** {predicted_team}  
+    **Team Email:** {team_email}  
+    **Top Keywords:** {', '.join(keywords)}
     """
 
-    return result, fig, shap_explanation
-
-# Function to download logs
-def download_logs():
-    with open(LOG_FILE, "r") as f:
-        return f.read()
-
-# Gradio Interface with SHAP, Logs, and Faster Inference
+# Gradio Interface
 interface = gr.Interface(
-    fn=predict_with_visuals,
+    fn=predict_with_keywords,
     inputs=gr.Textbox(lines=2, placeholder="Enter defect description..."),
-    outputs=["html", "plot", "html"],
-    title="🔎 AI Defect Ticket Classifier",
-    description="Enter a defect description to predict its **Category, Assigned Team, and relevant Keywords**. See **SHAP explanations, logs, and confidence scores!**",
-    theme="dark"
-)
-
-# Add logs download button
-download_interface = gr.Interface(
-    fn=download_logs,
-    inputs=[],
-    outputs="text",
-    title="📂 Download Prediction Logs",
-    description="Click the button to download all logs.",
+    outputs="markdown",
+    title="Defect Ticket Classifier",
+    description="Enter a defect description to predict its Category, Assigned Team, and relevant Keywords."
 )
 
-# Launch both interfaces
-interface.launch(share=True)
-download_interface.launch(share=True)
+# Launch app
+interface.launch()