Update app.py

app.py

@@ -1,39 +1,30 @@
-import gradio as gr
-import pandas as pd
-import numpy as np
-from sklearn.feature_extraction.text import CountVectorizer
-from sklearn.naive_bayes import MultinomialNB
-
-# Load the symptom-disease dataset
-df = pd.read_csv("Symptom-severity.csv")
+from transformers import AutoTokenizer, AutoModelForQuestionAnswering
 
-# Create a bag-of-words representation of the symptoms
-vectorizer = CountVectorizer()
-X = vectorizer.fit_transform(df["Symptom"].values.astype("U"))
-y = df["Disease"]
+tokenizer = AutoTokenizer.from_pretrained("microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext")
+model = AutoModelForQuestionAnswering.from_pretrained("microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext")
 
-# Train a Naive Bayes classifier on the symptom-disease dataset
-clf = MultinomialNB()
-clf.fit(X, y)
+import torch
 
-# Define the chatbot function
-def diagnose_disease(symptoms):
-    # Convert input symptoms to bag-of-words representation
-    X_new = vectorizer.transform([symptoms])
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
-    # Predict the disease using the trained classifier
-    disease = clf.predict(X_new)[0]
+def biomedical_chatbot(user_message):
+    # Tokenize the user's message
+    inputs = tokenizer.encode_plus(user_message, add_special_tokens=True, return_tensors="pt").to(device)
+    
+    # Generate a response using the pre-trained model
+    answer_start_scores, answer_end_scores = model(**inputs)
+    answer_start = torch.argmax(answer_start_scores)
+    answer_end = torch.argmax(answer_end_scores) + 1
+    answer = tokenizer.convert_tokens_to_string(tokenizer.convert_ids_to_tokens(inputs["input_ids"][0][answer_start:answer_end]))
+    
+    # Return the response
+    return answer
 
-    # Get the description of the predicted disease
-    description = df[df["Disease"] == disease]["Description"].values[0]
+import gradio as gr
 
-    return f"The most likely disease based on the symptoms entered is {disease}. {description}"
+gradio_interface = gr.Interface(fn=biomedical_chatbot, 
+                                inputs=gr.inputs.Textbox(placeholder="Enter your message here..."), 
+                                outputs=gr.outputs.Textbox())
 
-# Define the input and output interfaces
-input_text = gr.inputs.Textbox(label="Enter your symptoms separated by commas")
-output_text = gr.outputs.Textbox()
 
-# Create the Gradio interface
-gr.Interface(fn=diagnose_disease, inputs=input_text, outputs=output_text, 
-             title="Symptom-based Disease Diagnosis Chatbot", 
-             description="Enter your symptoms separated by commas, and the chatbot will predict the most likely disease.").launch()
+gradio_interface.launch()