import bitsandbytes as bnb
from transformers import (AutoModelForCausalLM,
                          AutoTokenizer,
                          BitsAndBytesConfig,
                          TrainingArguments,
                          pipeline)
import torch

class EndpointHandler():
    def __init__(self, path=""):
        bnb_config = BitsAndBytesConfig(
            load_in_8bit=True
        )

        model = AutoModelForCausalLM.from_pretrained(
            path,
            device_map="auto",
            torch_dtype="float16",
            quantization_config=bnb_config,
        )

        model.eval()

        tokenizer = AutoTokenizer.from_pretrained(path)
        pipe = pipeline(
            "text-generation",
            model=model,
            tokenizer=tokenizer,
            torch_dtype=torch.float16,
            device_map="auto",
        )
        self.model = model 
        self.tokenizer = tokenizer
        self.pipeline = pipe

    def __call__(self, data) :
        """
        Args:
            data (:obj:):
                includes the input data and the parameters for the inference.
        Return:
            A :obj:`list`:. The object returned should be a list of one list like [[{"label": 0.9939950108528137}]] containing :
                - "label": A string representing what the label/class is. There can be multiple labels.
                - "score": A score between 0 and 1 describing how confident the model is for this label/class.
        """
        inputs = data.pop("inputs", data)
        parameters = data.pop("parameters", None)
        output = self.pipeline(inputs , **parameters)
        return output