braintumor.py

import os
import shutil
import json
from pathlib import Path
from tqdm import tqdm
from glob import glob
from typing import Dict, Any, List, Union, Iterator

import yaml
from yaml.loader import SafeLoader
import datasets
from datasets.download.download_manager import DownloadManager, ArchiveIterable
from pylabel import importer


_DESCRIPTION = """\
Training image sets and labels/bounding box coordinates for detecting brain
tumors in MR images.
- The datasets JPGs exported at their native size and are separated by plan
(Axial, Coronal and Sagittal).
- Tumors were hand labeled using https://makesense.ai
- Bounding box coordinates and MGMT positive labels were marked on ~400 images
for each plane in the T1wCE series from the RSNA-MICCAI competition data.
"""

_URLS = {
    "yolo": "https://huggingface.co./datasets/chanelcolgate/brain-tumors-object-detection-datasets/resolve/main/data/archive.zip"
}

_CLASSES = ["negative", "positive"]


# move all into one folder
def copy_yolo_files(from_folder, to_folder, images_labels, train_test):
    from_path = os.path.join(from_folder, images_labels, train_test)
    to_path = os.path.join(to_folder, images_labels, train_test)
    os.makedirs(to_path, exist_ok=True)
    # get files
    file_ext = "*.jpg" if images_labels == "images" else "*.txt"
    files = glob(os.path.join(from_path, file_ext))
    # move files
    for file in tqdm(files):
        shutil.copy(file, to_path)


def yolo_to_coco(input_folder, output_folder, train_test):
    labels_path = os.path.join(input_folder, "labels", train_test)
    images_path = os.path.join(input_folder, "images", train_test)
    coco_dir = os.path.join(output_folder, train_test)
    os.makedirs(coco_dir, exist_ok=True)

    txt_files = glob(os.path.join(labels_path, "*.txt"))
    img_files = glob(os.path.join(images_path, "*.jpg"))
    # copy annotations
    for f in tqdm(txt_files):
        shutil.copy(f, coco_dir)
    # copy images
    for f in tqdm(img_files):
        shutil.copy(f, coco_dir)

    # get the classes
    with open(os.path.join(input_folder, "classes.txt"), "r") as f:
        classes = f.read().split("\n")

    # load dataset
    dataset = importer.ImportYoloV5(
        path=coco_dir, cat_names=classes, name="brain tumors"
    )
    # export
    coco_file = os.path.join(coco_dir, "_annotations.coco.json")
    # Detection requires starting index from 1
    dataset.export.ExportToCoco(coco_file, cat_id_index=0)
    # now delete yolo annotations in coco set
    for f in txt_files:
        os.remove(f.replace(labels_path, coco_dir))


def round_box_values(box, decimals=2):
    return [round(val, decimals) for val in box]


class COCOHelper:
    """Helper class to load COCO annotations"""

    def __init__(self, annotation_path: Path, images_dir: Path) -> None:
        with open(annotation_path, "r") as file:
            data = json.load(file)
        self.data = data

        dict_id2annot: Dict[int, Any] = {}
        for annot in self.annotations:
            dict_id2annot.setdefault(annot["image_id"], []).append(annot)

        # Sort by id
        dict_id2annot = {
            k: list(sorted(v, key=lambda a: a["id"]))
            for k, v in dict_id2annot.items()
        }

        self.dict_path2annot: Dict[str, Any] = {}
        self.dict_path2id: Dict[str, Any] = {}
        for img in self.images:
            path_img = os.path.join(images_dir, img["file_name"])
            path_img_str = os.path.normpath(path_img)
            idx = int(img["id"])
            annot = dict_id2annot.get(idx, [])
            self.dict_path2annot[path_img_str] = annot
            self.dict_path2id[path_img_str] = img["id"]

    def __len__(self) -> int:
        return len(self.data["images"])

    @property
    def images(self) -> List[Dict[str, Union[str, int]]]:
        return self.data["images"]

    @property
    def annotations(self) -> List[Any]:
        return self.data["annotations"]

    @property
    def categories(self) -> List[Dict[str, Union[str, int]]]:
        return self.data["categories"]

    def get_annotations(self, image_path: str) -> List[Any]:
        return self.dict_path2annot.get(image_path, [])

    def get_image_id(self, image_path: str) -> int:
        return self.dict_path2id.get(image_path, -1)


class COCOBrainTumor(datasets.GeneratorBasedBuilder):
    """COCO Brain Tumor dataset"""

    VERSION = datasets.Version("1.0.1")

    def _info(self) -> datasets.DatasetInfo:
        """
        Return the dataset metadata and features.

        Returns:
            DatasetInfo: Metadata and features of the dataset.
        """
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "image": datasets.Image(),
                    "image_id": datasets.Value("int64"),
                    "objects": datasets.Sequence(
                        {
                            "id": datasets.Value("int64"),
                            "area": datasets.Value("float64"),
                            "bbox": datasets.Sequence(
                                datasets.Value("float32"), length=4
                            ),
                            "label": datasets.ClassLabel(names=_CLASSES),
                            "iscrowd": datasets.Value("bool"),
                        }
                    ),
                }
            ),
        )

    def _split_generators(
        self, dl_manager: DownloadManager
    ) -> List[datasets.SplitGenerator]:
        """
        Provides the split information and downloads the data.

        Args:
            dl_manager (DownloadManager): The DownloadManager to use for
                downloading and extracting data.

        Returns:
            List[SplitGenerator]: List of SplitGenrator objects representing
                the data splits.
        """
        archive_yolo = dl_manager.download(_URLS["yolo"])
        archive_yolo = dl_manager.extract(archive_yolo)
        data_folder = "braintumors"
        data_folder_yolo = data_folder + "_yolo"
        data_folder_coco = data_folder + "_coco"
        folders = os.listdir(str(archive_yolo))

        # copy
        for from_folder in folders:
            from_folder = os.path.join(archive_yolo, from_folder)
            to_folder = os.path.join(archive_yolo, data_folder_yolo)
            for images_labels in ["images", "labels"]:
                for train_test in ["train", "test"]:
                    copy_yolo_files(
                        from_folder, to_folder, images_labels, train_test
                    )

        # Open the file and load the file
        with open(
            os.path.join(archive_yolo, folders[0], folders[0] + ".yaml")
        ) as f:
            classes = yaml.load(f, Loader=SafeLoader)["names"]

        # Write classes.txt
        with open(
            os.path.join(archive_yolo, data_folder_yolo, "classes.txt"), "w"
        ) as f:
            f.write("\n".join(classes))

        data_folder_yolo = os.path.join(archive_yolo, data_folder_yolo)
        data_folder_coco = os.path.join(archive_yolo, data_folder_coco)
        yolo_to_coco(data_folder_yolo, data_folder_coco, "train")
        yolo_to_coco(data_folder_yolo, data_folder_coco, "test")

        name_ds = str(archive_yolo) + "/braintumors_coco"
        image_root_train = name_ds + "/train"
        image_root_test = name_ds + "/test"
        af = "_annotations.coco.json"
        json_file_train = name_ds + "/train/" + af
        json_file_test = name_ds + "/test/" + af

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "annotation_path": json_file_train,
                    "images_dir": image_root_train,
                    "images": dl_manager.iter_files(image_root_train),
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "annotation_path": json_file_test,
                    "images_dir": image_root_test,
                    "images": dl_manager.iter_files(image_root_test),
                },
            ),
        ]

    def _generate_examples(
        self, annotation_path: Path, images_dir: Path, images: ArchiveIterable
    ) -> Iterator:
        """
        Generates examples for the dataset.

        Args:
            annotation_path (Path): The path to the annotation file.
            images_dir (Path): The path to the directory containing the images.
            images: (ArchiveIterable): An iterable containing the images.

        Yields:
            Dict[str, Union[str, Image]]: A dictionary containing the
                generated examples.
        """
        coco_annotation = COCOHelper(annotation_path, images_dir)

        for image_path in images:
            image_path = os.path.normpath(image_path)
            if "_annotations.coco.json" not in image_path:
                f = open(image_path, "rb")
                annotations = coco_annotation.get_annotations(image_path)
                ret = {
                    "image": {"path": image_path, "bytes": f.read()},
                    "image_id": coco_annotation.get_image_id(image_path),
                    "objects": [
                        {
                            "id": annot["id"],
                            "area": annot["area"],
                            "bbox": round_box_values(
                                annot["bbox"], 2
                            ),  # [x, y, w, h]
                            "label": annot["category_id"],
                            "iscrowd": bool(annot["iscrowd"]),
                        }
                        for annot in annotations
                    ],
                }
                yield image_path, ret