Image-demo.py

import datasets
import json
import os

_CITATION = """\
@InProceedings{huggingface:dataset,
title = {Small image-text set},
author={James Briggs},
year={2022}
}
"""

_DESCRIPTION = """\
Demo dataset for testing or showing image-text capabilities.
"""
_HOMEPAGE = "https://huggingface.co/datasets/ppp121386/Image-demo"

_LICENSE = ""

_REPO_URL = "https://huggingface.co/datasets/ppp121386/Image-demo/resolve/main/images.tar.gz"
_CAPTION_URL = "https://huggingface.co/datasets/ppp121386/Image-demo/resolve/main/caption.json"

# _CAPTION =  ["a dog sitting on a bed looking at a pink wall","a brown dog sitting in front of a pink wall","two people are cross country skiing in the snow","a woman is cross country skiing in the snow","a woman is cross country skiing in the snow"]

class ImageSet(datasets.GeneratorBasedBuilder):
    """Small sample of image-text pairs"""

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    'text': datasets.Value("string"),
                    'image': datasets.Image(),
                }
            ),
            supervised_keys=None,
            homepage=_HOMEPAGE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        images_archive = dl_manager.download(_REPO_URL)
        image_iters = dl_manager.iter_archive(images_archive)
        filepath = dl_manager.download_and_extract(_CAPTION_URL)
        # caption_iters = dl_manager.iter_archive(caption_archive)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "images": image_iters,
                    "filepath": filepath
                }
            ),
        ]

    def _generate_examples(self, images, filepath):
        """ This function returns the examples in the raw (text) form."""
        _CAPTION = json.load(open(filepath, 'r'))
        for idx, (imgpath, image) in enumerate(images):
            # description = filepath.split('/')[-1][:-4]
            # description = description.replace('_', ' ')
            yield idx, {
                "image": {"path": imgpath, "bytes": image.read()},
                "text": _CAPTION[idx],
            }