dataset-qc

#!/usr/bin/env python

import os
import json
import argparse
from dsf.data import lemnatec


def options():
    parser = argparse.ArgumentParser(description='Output dataset statistics.',
                                     formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    parser.add_argument("-d", "--dataset", help="Dataset directory.", required=True)
    args = parser.parse_args()

    return args


def main():
    # Read user options
    args = options()

    # Load the dataset metadata
    meta = lemnatec.load_dataset(dataset_dir=args.dataset)

    # Good/bad images
    stats = {
        "VIS-SV": {
            "total": 0,
            "incomplete": 0
        },
        "VIS-TV": {
            "total": 0,
            "incomplete": 0
        },
        "NIR-SV": {
            "total": 0,
            "incomplete": 0
        },
        "NIR-TV": {
            "total": 0,
            "incomplete": 0
        },
        "FLUO-TV": {
            "total": 0,
            "incomplete": 0
        }
    }
    for img in meta["images"]:
        imgtype = meta["images"][img].get("imgtype")
        imgtype = imgtype.upper()
        camera = meta["images"][img].get("camera")
        stats[f"{imgtype}-{camera}"]["total"] += 1
        if not os.path.exists(os.path.join(args.dataset, img)):
            stats[f"{imgtype}-{camera}"]["incomplete"] += 1

    print(json.dumps(stats, indent=4))


if __name__ == "__main__":
    main()