import csv import os import datasets logger = datasets.logging.get_logger(__name__) _CITATION = """""" _DESCRIPTION = """Persian Licensee plate dataset. Primarily taken from AmirKabir University Challenge. Annotation are provided by the authors""" _DOWNLOAD_URLS = { "train": "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_train.csv", "val": "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_val.csv", "test": "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_test.csv", 'train_dataset': "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_train.zip", 'val_dataset': "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_val.zip", 'test_dataset': "https://huggingface.co/datasets/hezarai/persian-license-plate-v1/resolve/main/persian_license_plate_test.zip", } class PersianLicensePlateConfig(datasets.BuilderConfig): def __init__(self, **kwargs): super(PersianLicensePlateConfig, self).__init__(**kwargs) class PersianLicensePlate(datasets.GeneratorBasedBuilder): BUILDER_CONFIGS = [ PersianLicensePlateConfig( name="PersianLicensePlate", version=datasets.Version("1.0.0"), description=_DESCRIPTION, ), ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "label": datasets.Value("string"), "image_path": datasets.Value("string"), } ), citation=_CITATION, ) def _split_generators(self, dl_manager): """ Return SplitGenerators. """ train_path = dl_manager.download_and_extract(_DOWNLOAD_URLS["train"]) val_path = dl_manager.download_and_extract(_DOWNLOAD_URLS['val']) test_path = dl_manager.download_and_extract(_DOWNLOAD_URLS["test"]) archive_path = dl_manager.download(_DOWNLOAD_URLS['train_dataset']) train_extracted_path = dl_manager.extract(archive_path) if not dl_manager.is_streaming else "" archive_path = dl_manager.download(_DOWNLOAD_URLS['val_dataset']) val_extracted_path = dl_manager.extract(archive_path) if not dl_manager.is_streaming else "" archive_path = dl_manager.download(_DOWNLOAD_URLS['test_dataset']) test_extracted_path = dl_manager.extract(archive_path) if not dl_manager.is_streaming else "" return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, gen_kwargs={"filepath": train_path, "dataset_dir": train_extracted_path} ), datasets.SplitGenerator( name=datasets.Split.TEST, gen_kwargs={"filepath": test_path, "dataset_dir": test_extracted_path} ), datasets.SplitGenerator( name=datasets.Split.VALIDATION, gen_kwargs={"filepath": val_path, "dataset_dir": val_extracted_path} ), ] def _generate_examples(self, filepath, dataset_dir): logger.info("⏳ Generating examples from = %s", filepath) with open(filepath, encoding="utf-8") as csv_file: csv_reader = csv.reader(csv_file, quotechar='"', skipinitialspace=True) # Skip header next(csv_reader, None) for id_, row in enumerate(csv_reader): label, filename = row image_path = os.path.join(dataset_dir, filename) yield id_, {"label": label, "image_path": image_path}