import pandas as pd
from huggingface_hub import hf_hub_url
import datasets
import os

_VERSION = datasets.Version("0.0.2")

_DESCRIPTION = "Your dataset description here"
_HOMEPAGE = "Your dataset homepage URL here"
_LICENSE = "Your dataset license here"
_CITATION = "Your dataset citation here"

_FEATURES = datasets.Features(
    {
        "image": datasets.Image(),
        "conditioning_image": datasets.Image(),
        "text": datasets.Value("string"),
    },
)

# Update these URLs to point to your dataset and images
METADATA_URL = hf_hub_url(
    "chaturyaganne05/imgtovideo",
    filename="dataset(1).json",
    repo_type="dataset",
)

IMAGES_URL = hf_hub_url(
    "chaturyaganne05/imgtovideo",
    filename="images.zip",
    repo_type="dataset",
)

CONDITIONING_IMAGES_URL = hf_hub_url(
    "chaturyaganne05/imgtovideo",
    filename="conditioning_images.zip",
    repo_type="dataset",
)

_DEFAULT_CONFIG = datasets.BuilderConfig(name="default", version=_VERSION)

class Fill50k(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [_DEFAULT_CONFIG]
    DEFAULT_CONFIG_NAME = "default"

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=_FEATURES,
            supervised_keys=None,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        metadata_path = dl_manager.download(METADATA_URL)
        images_dir = dl_manager.download_and_extract(IMAGES_URL)
        conditioning_images_dir = dl_manager.download_and_extract(
            CONDITIONING_IMAGES_URL
        )

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "metadata_path": metadata_path,
                    "images_dir": images_dir,
                    "conditioning_images_dir": conditioning_images_dir,
                },
            ),
        ]

    def _generate_examples(self, metadata_path, images_dir, conditioning_images_dir):
        metadata = pd.read_json(metadata_path, lines=True)

        for _, row in metadata.iterrows():
            text = row["text"]

            image_path = row["image"]
            image_path = os.path.join(images_dir, image_path)
            with open(image_path, "rb") as img_file:
                image = img_file.read()

            conditioning_image_path = row["conditioning_image"]
            conditioning_image_path = os.path.join(
                conditioning_images_dir, conditioning_image_path
            )
            with open(conditioning_image_path, "rb") as cond_img_file:
                conditioning_image = cond_img_file.read()

            yield row["image"], {
                "text": text,
                "image": {
                    "path": image_path,
                    "bytes": image,
                },
                "conditioning_image": {
                    "path": conditioning_image_path,
                    "bytes": conditioning_image,
                },
            }