Update app.py

app.py

@@ -1,41 +1,34 @@
 import os
 import tensorflow as tf
 from tensorflow import keras
-from fastapi import FastAPI
-from pydantic import BaseModel
-import uvicorn
-
-# Suppress TensorFlow logging
-os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
+import gradio as gr
 
 # Load the model
-model_path = 'doctor_ai_model.h5'  # Use the correct path and file format
+model_path = 'doctor_ai_model.h5'  # or .h5
 model = keras.models.load_model(model_path)
 
-# Create FastAPI app
-app = FastAPI()
-
-# Define the request model
-class InputData(BaseModel):
-    input_data: list
+def predict(input_data):
+    # Convert input to tensor
+    input_tensor = tf.convert_to_tensor(input_data)
 
-# Define the prediction endpoint
-@app.post('/predict')
-async def predict(data: InputData):
-    input_array = tf.convert_to_tensor(data.input_data)
-
-    # Check input shape
-    if input_array.shape[1] != 27:
-        return {'error': 'Input data must have shape: (None, 27)'}
+    # Ensure the input shape is correct
+    if input_tensor.shape[1] != 27:  # Adjust based on your input shape
+        return "Input data must have shape: (None, 27)"
     
-    # Expand dimensions if necessary
-    input_array = tf.expand_dims(input_array, axis=0)
+    # Expand dimensions for the model
+    input_tensor = tf.expand_dims(input_tensor, axis=0)
 
-    prediction = model.predict(input_array)
+    # Make prediction
+    prediction = model.predict(input_tensor)
     predicted_class = tf.argmax(prediction, axis=1).numpy().tolist()
     
-    return {'predicted_class': predicted_class}
+    return predicted_class
+
+# Create Gradio interface
+iface = gr.Interface(fn=predict, 
+                     inputs=gr.inputs.Dataframe(type='numpy', 
+                                                 label='Input Data (should have shape (None, 27))'), 
+                     outputs='json')
 
-# Start the FastAPI server
-if __name__ == '__main__':
-    uvicorn.run(app, host='127.0.0.1', port=8000)
+# Launch the Gradio app
+iface.launch()