Source code for syne_tune.blackbox_repository.conversion_scripts.blackbox_recipe

import logging

from syne_tune.blackbox_repository.conversion_scripts.utils import (
    upload_blackbox,
)
from syne_tune.util import catchtime




class BlackboxRecipe:
    def __init__(self, name: str, cite_reference: str):
        """
        Parent class for a blackbox recipe that allows to generate the blackbox data on disk, see ``FCNETRecipe`` or
        ``LCBenchRecipe`` classes for example on how to add a new benchmark.
        :param name: name of the blackbox
        :param cite_reference: name of the paper to be referenced. A message is prompted when generating the blackbox
        to ask the user to cite the relevant paper.
        """
        self.name = name
        self.cite_reference = cite_reference



    def generate(self, upload_on_hub: bool = True):
        """
        Generates the blackbox on disk then upload it on HuggingFace hub
        :return:
        """
        message = (
            f"Generating {self.name} blackbox locally, if you use this dataset in a publication, please cite "
            f'the following paper: "{self.cite_reference}"'
        )
        logging.info(message)
        self._generate_on_disk()

        if upload_on_hub:
            with catchtime(f"Uploading blackbox {self.name} to HuggingFace hub"):
                upload_blackbox(name=self.name)


    def _generate_on_disk(self):
        """
        Method to be overloaded by the child class that should generate the blackbox on disk (handling the donwloading
        and reformatting of external files).
        :return:
        """
        raise NotImplementedError()