from transformers import CLIPModel, CLIPProcessor
from typing import Dict, List, Any
from PIL import Image
from transformers import pipeline
import requests
import torch

class EndpointHandler():
    def __init__(self, path=""):
        """
        path: 
        """
        self.device = "cuda" if torch.cuda.is_available() else "cpu"
        self.processor = CLIPProcessor.from_pretrained(path)
        self.model = CLIPModel.from_pretrained(path).to(self.device)
        
    def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
        """
        data args:
            inputs (:obj: `str` | `PIL.Image` | `np.array`)
            kwargs
        Return:
            A :obj:`list` | `dict`: will be serialized and returned
        """
        result = {}
        inputs = data.pop("inputs", data)
        image_url = inputs['image_url']
        image = Image.open(requests.get(image_url, stream=True).raw).convert('RGB')
        processed_image = self.processor(images=image, return_tensors="pt").to(self.device)
        output = self.model.get_image_features(processed_image["pixel_values"])[0].tolist()
        result["embedding"] = output
        return result