convert.py

#!/usr/bin/env python3
import argparse
import json
import os


def parse_args():
    parser = argparse.ArgumentParser(description="Convert DICOM files to BIDS format.")
    parser.add_argument("input_dir", type=str, help="Input directory containing DICOM files.")
    parser.add_argument("output_dir", type=str, help="Output directory for BIDS formatted files.")
    parser.add_argument(
        "--heuristic",
        type=str,
        help="Path to the heuristic JSON file.",
        required=True
    )
    parser.add_argument(
        "--subject-map",
        type=str,
        default=None,
        help="Optional CSV/XLS/XLSX mapping file for source identifiers -> bids_subject/session_id.",
    )
    parser.add_argument(
        "--use-session-dates",
        action=argparse.BooleanOptionalAction,
        default=None,
        help="Use acquisition dates for session names (ses-YYYYMMDD).",
    )
    parser.add_argument(
        "--combine-sessions",
        action=argparse.BooleanOptionalAction,
        default=None,
        help="Combine all scans into a single no-session directory per subject.",
    )
    parser.add_argument(
        "--source-id",
        action="append",
        choices=["patient_name", "patient_id"],
        default=None,
        help=(
            "Repeatable source identifier(s) used to match rows in --subject-map. "
            "patient_id is DICOM PatientID (0010,0020; same concept as dcm2niix %%i) and is not guaranteed to be MRN. "
            "patient_name uses DICOM PatientName (0010,0010) with exact string matching."
        ),
    )
    parser.add_argument(
        "--no-anonymize",
        action="store_true",
        help="Disable anonymization of sensitive patient data during conversion (not recommended).",
    )
    parser.add_argument("--verbose", action="store_true", help="Enable verbose output.")
    parser.add_argument("--debug", action="store_true", help="Enable debug output.")
    return parser.parse_args()


def main():
    from bidsmanager.read.dicom_reader import convert_dicom_directory
    args = parse_args()
    # Load heuristic from JSON file
    with open(args.heuristic, 'r') as f:
        heuristic = json.load(f)

    # CLI args override heuristic values when provided.
    subject_map = args.subject_map if args.subject_map else heuristic.get("subject_map")
    # Backward compatibility for existing heuristic files.
    if not subject_map:
        subject_map = heuristic.get("subject_map_csv") or heuristic.get("subject_map_excel")

    use_session_dates = args.use_session_dates
    if use_session_dates is None:
        use_session_dates = heuristic.get("use_session_dates", False)

    combine_sessions = args.combine_sessions
    if combine_sessions is None:
        combine_sessions = heuristic.get("combine_sessions", False)

    source_ids = args.source_id if args.source_id else heuristic.get("source_id")
    if isinstance(source_ids, str):
        source_ids = [source_ids]

    input_dir = os.path.abspath(args.input_dir)
    output_dir = os.path.abspath(args.output_dir)
    verbose = args.verbose
    if args.debug:
        verbose = True
        # Create a CSV file to store the DICOM files found
        output_file = os.path.join(output_dir, "source", "dicom_files.csv")
        os.makedirs(os.path.dirname(output_file), exist_ok=True)

    convert_dicom_directory(input_directory=input_dir,
                            heuristic=heuristic,
                            anonymize=not args.no_anonymize,
                            bids_directory=output_dir,
                            delete_intermediates=True,
                            verbose=verbose,
                            use_session_dates=use_session_dates,
                            combine_sessions=combine_sessions,
                            subject_map=subject_map,
                            source_ids=source_ids,
                            cleanup_temp_directory=not args.debug)


if __name__ == "__main__":
    main()