loc_beyond_words.py

# Copyright 2022 Daniel van Strien
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Beyond Words"""

import collections
import json
import os
from typing import Any, Dict, List
import datasets
from pathlib import Path

_CITATION = "TODO"


_DESCRIPTION = "TODO"


_HOMEPAGE = "TODO"


_LICENSE = "Public Domain Mark 1.0"


class BeyondWords(datasets.GeneratorBasedBuilder):
    """Beyond Words Dataset"""

    def _info(self):
        features = datasets.Features(
            {
                "image_id": datasets.Value("int64"),
                "image": datasets.Image(),
                "width": datasets.Value("int32"),
                "height": datasets.Value("int32"),
            }
        )

        object_dict = {
            "bw_id": datasets.Value("string"),
            "category_id": datasets.ClassLabel(
                names=[
                    "Photograph",
                    "Illustration",
                    "Map",
                    "Comics/Cartoon",
                    "Editorial Cartoon",
                    "Headline",
                    "Advertisement",
                ]
            ),
            "image_id": datasets.Value("string"),
            "id": datasets.Value("int64"),
            "area": datasets.Value("int64"),
            "bbox": datasets.Sequence(datasets.Value("float32"), length=4),
            "iscrowd": datasets.Value(
                "bool"
            ),  # always False for stuff segmentation task
        }
        features["objects"] = [object_dict]

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        images = dl_manager.download_and_extract("data/images.zip")
        training = dl_manager.download("data/train_80_percent.json")
        validation = dl_manager.download("data/val_20_percent.json")
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "annotations_file": Path(training),
                    "image_dir": Path(images),
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "annotations_file": Path(validation),
                    "image_dir": Path(images),
                },
            ),
        ]

    def _get_image_id_to_annotations_mapping(
        self, annotations: List[Dict]
    ) -> Dict[int, List[Dict[Any, Any]]]:
        """
        A helper function to build a mapping from image ids to annotations.
        """
        image_id_to_annotations = collections.defaultdict(list)
        for annotation in annotations:
            image_id_to_annotations[annotation["image_id"]].append(annotation)
        return image_id_to_annotations

    def _generate_examples(self, annotations_file, image_dir):
        def _image_info_to_example(image_info, image_dir):
            image = image_info["file_name"]
            return {
                "image_id": image_info["id"],
                "image": os.path.join(image_dir, "images", image),
                "width": image_info["width"],
                "height": image_info["height"],
            }

        with open(annotations_file, encoding="utf8") as f:
            annotation_data = json.load(f)
            images = annotation_data["images"]
            annotations = annotation_data["annotations"]
            image_id_to_annotations = self._get_image_id_to_annotations_mapping(
                annotations
            )
            for idx, image_info in enumerate(images):
                example = _image_info_to_example(image_info, image_dir)

                annotations = image_id_to_annotations[image_info["id"]]
                objects = []
                for annotation in annotations:
                    objects.append(annotation)
                example["objects"] = objects
                yield (idx, example)