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'

# Load the model
model_path = 'doctor_ai_model'
model = keras.models.load_model(model_path)

# Create FastAPI app
app = FastAPI()

# Define the request model
class InputData(BaseModel):
    input_data: list

# 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)'}
    
    # Expand dimensions if necessary
    input_array = tf.expand_dims(input_array, axis=0)

    prediction = model.predict(input_array)
    predicted_class = tf.argmax(prediction, axis=1).numpy().tolist()
    
    return {'predicted_class': predicted_class}

# Start the FastAPI server
if __name__ == '__main__':
    uvicorn.run(app, host='127.0.0.1', port=8000)