FastAPI_garbage.py

import uvicorn
import numpy as np
import tensorflow as tf
from fastapi import FastAPI, File, UploadFile
from fastapi.middleware.cors import CORSMiddleware
from tensorflow.keras.preprocessing import image
import io
from PIL import Image

# -------------------------------
# Loading trained model
# -------------------------------
MODEL_PATH = "garbage_mobilenet_V3_finetuned.h5"
model = tf.keras.models.load_model(MODEL_PATH)

# class names
class_names = ["Bio", "metal", "paper", "plastic"]

# -------------------------------
# FastAPI app setup
# -------------------------------
app = FastAPI()

# Enabling CORS (so frontend can calls API)
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],  #restricting this to ["http://127.0.0.1:5500"] --->ip address
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# -------------------------------
# Prediction endpoint
# -------------------------------
@app.post("/predict")
async def predict(file: UploadFile = File(...)):
    try:
        # Read image file
        contents = await file.read()
        img = Image.open(io.BytesIO(contents)).convert("RGB")
        img = img.resize((224, 224))  # MobileNet input size

        # Preprocess image
        img_array = image.img_to_array(img) / 255.0
        img_array = np.expand_dims(img_array, axis=0)

        # Prediction
        preds = model.predict(img_array, verbose=0)
        pred_class = class_names[np.argmax(preds)]
        confidence = float(np.max(preds) * 100)

        return {"class": pred_class, "confidence": confidence}

    except Exception as e:
        return {"error": str(e)}

# -------------------------------
# Running server
# -------------------------------
if __name__ == "__main__":
    uvicorn.run(app, host="127.0.0.1", port=8000)