import datasets
from datasets.tasks import QuestionAnsweringExtractive


logger = datasets.logging.get_logger(__name__)

_URL = "https://huggingface.co/datasets/yanmiamin/breastCancerEnhancedTest/resolve/main/cropped3c_test.tar.gz"


class breastcancer(datasets.GeneratorBasedBuilder):
    
    def _info(self):
        return datasets.DatasetInfo(
        
            features=datasets.Features(
                {
                    "image_file_path": datasets.Value("string"),
                    "image_name": datasets.Value("string"),
                    "image": datasets.Image(),
                    
                }
            ),
            # No default supervised_keys (as we have to pass both question
            # and context as input).
            supervised_keys=None,
            homepage="https://huggingface.co/datasets/yanmiamin/breastCancerEnhancedTest",
            
        )
        
    def _split_generators(self, dl_manager):
        
        downloaded_files = dl_manager.download(_URL)
        image_iters = dl_manager.iter_archive(downloaded_files)

        return [
            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"images": image_iters})
        ]


    def _generate_examples(self, images):
        """This function returns the examples in the raw (text) form."""
        
        idx = 0

        for filepath, image in images:
            name = filepath.split('/')[-1]
            name = name.split('.')[0]
            yield idx, {
                "image_file_path": filepath, 
                "image_name": name,
                "image": {
                    "bytes":image.read()
                },
            }
            idx += 1