Update app.py

app.py

@@ -1,18 +1,39 @@
 from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
 import tensorflow as tf
 import gradio as gr
+import os
+import re
+from nltk.corpus import stopwords
+from nltk.stem import PorterStemmer
+
+# Ensure you have the necessary nltk resources
+import nltk
+nltk.download('stopwords')
+
+# Caching the model locally to avoid re-downloading
+MODEL_NAME = "Zabihin/Symptom_to_Diagnosis"
+CACHE_DIR = "./cached_model"
+if not os.path.exists(CACHE_DIR):
+    os.makedirs(CACHE_DIR)
 
 # Load the tokenizer and model
-model_name = "Zabihin/Symptom_to_Diagnosis"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
+model = TFAutoModelForSequenceClassification.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
+
+# Initialize stopwords and stemmer
+stop_words = set(stopwords.words('english'))
+stemmer = PorterStemmer()
 
-# Clean the input text
+# Clean the input text with advanced preprocessing
 def clean_input(symptom_text):
-    # Remove unwanted characters or non-ASCII characters
+    # Remove non-ASCII characters and convert to lowercase
     symptom_text = ''.join([c for c in symptom_text if ord(c) < 128])
-    symptom_text = symptom_text.lower()  # Optional: Convert to lowercase
-    return symptom_text
+    symptom_text = symptom_text.lower().strip()  # Remove leading/trailing spaces
+    
+    # Remove stopwords and apply stemming
+    words = symptom_text.split()
+    filtered_words = [stemmer.stem(word) for word in words if word not in stop_words]
+    return ' '.join(filtered_words)
 
 # Define the predict function
 def predict(symptom_text, chat_history=[]):
@@ -58,7 +79,7 @@ def predict(symptom_text, chat_history=[]):
 
         # Add conversation history
         chat_history.append(("User", symptom_text))
-        chat_history.append(("AI", f"Predicted Diagnosis: {diagnosis}. {description} Please consult a doctor for more accurate results."))
+        chat_history.append(("AI", f"Predicted Diagnosis: **{diagnosis}**. {description} Please consult a doctor for more accurate results."))
 
     except Exception as e:
         chat_history.append(("AI", f"Error: {str(e)}"))
@@ -67,7 +88,7 @@ def predict(symptom_text, chat_history=[]):
 
 # Gradio UI
 with gr.Blocks() as interface:
-    gr.Markdown("<h1 style='text-align: center; font-size: 40px; margin-top: 30px; margin-bottom: 30px;'>Medi Mind - Your AI Health Assistant</h1>")
+    gr.Markdown("<h1 style='text-align: center; font-size: 50px; margin-top: 40px; margin-bottom: 40px;'>Medi Mind - Your AI Health Assistant</h1>")
     chatbot = gr.Chatbot()
     input_box = gr.Textbox(show_label=False, placeholder="Describe your symptoms here...")
     send_button = gr.Button("Send")