from fastapi import FastAPI, File, UploadFile import numpy as np import tensorflow as tf import keras from PIL import Image from io import BytesIO from fastapi.middleware.cors import CORSMiddleware import time app = FastAPI() app.title = "API de prueba" app.version = "0.0.1" origins = ["*"] clases = ["Industria", "Golpeada", "Partida"] app.add_middleware( CORSMiddleware, allow_origins=origins, allow_credentials=True, allow_methods=["*"], allow_headers=["*"], ) # Cargar el modelo una vez model = keras.saving.load_model("models/modelo_fin_effi.keras") @app.get("/") def api_home(): return {'detail': 'Tesis Uvillas'} @app.post("/analyze") async def analyze(file: UploadFile = File(...)): image = await file.read() return predecir(image) def predecir(imagen): start_time = time.time() image = Image.open(BytesIO(imagen)) image = image.resize((96, 96)) input_arr = keras.utils.img_to_array(image) input_arr = np.array([input_arr]) predictions = model.predict(input_arr) end_time = time.time() prediction_time = end_time - start_time predicted_class = clases[np.argmax(predictions[0])] prediction_percentages = {clase: float(pred) * 100 for clase, pred in zip(clases, predictions[0])} return { 'clase_predicha': predicted_class, 'porcentajes_de_precisión': prediction_percentages, 'time':prediction_time }