stained-glass.py

"""Dataset class for stained-glass dataset."""

import datasets
import os
import pandas as pd

_DESCRIPTION = """Dataset consists of 21 high-resolution images of stained glass art, accompanied by corresponding captions"""
_HOMEPAGE = """https://huggingface.co/datasets/abinthomasonline/stained-glass"""
_ADJECTIVES = ["", "good", "cropped", "clean", "bright", "cool", "nice", "small", "large", "dark", "weird"]


class StainedGlass(datasets.GeneratorBasedBuilder):
    """Stained Glass Images dataset"""

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION, 
            features=datasets.Features(
                {
                    "image": datasets.Image(),
                    "caption": datasets.Value(dtype='string'),
                }
            ),
            supervised_keys=("image", "caption"),
            homepage=_HOMEPAGE,
        )

    def _split_generators(self, dl_manager):
        captions_path = dl_manager.download("captions.csv")
        df = pd.read_csv(captions_path, header=None)
        image_paths = {row[0]: os.path.join("images", row[0]) for _, row in df.iterrows()}
        image_paths = dl_manager.download(image_paths)

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "captions_path": captions_path,
                    "image_paths": image_paths,
                },
            )
        ]

    def _generate_examples(self, captions_path, image_paths):
        df = pd.read_csv(captions_path, header=None)
        captions = {row[0]: row[1].replace('"', '') for _, row in df.iterrows()}
        for adjective in _ADJECTIVES:
            for key, image_path in image_paths.items():
                yield key + adjective, {
                    "image": image_path,
                    "caption": captions[key].format(token='<token>', adjective=adjective),
                }