renamevzs

#!/usr/bin/python3

import argparse
from pathlib import Path
import re
import shutil
import string
import textwrap


def rename(filepath, dest=None, move=False, width=3):
    """Rename dipsExport file."""
    (transaction, provenance) = get_info(filepath)
    transaction_ = fmt_transaction(transaction, width)
    provenance_ = fmt_provenance(provenance)
    new_name = f"{transaction_}_{provenance_}.dips.xml"
    old_path = Path(filepath)
    new_path = Path(dest) / new_name if dest else old_path.with_name(new_name)
    if move:
        shutil.move(old_path, new_path)
    else:
        shutil.copy(old_path, new_path)


regex_transaction = re.compile("<(?:[^:]+:)?transactionNumber>([^<]*)</")
regex_provenance = re.compile("<(?:[^:]+:)?provenanceName>([^<]*)</")


def get_info(filepath):
    """Retrieve transaction number and provenance from dipsExport file."""
    transaction = ""
    provenance = ""
    with open(filepath) as f:
        for line in f:
            if not transaction:
                match_transaction = regex_transaction.search(line)
                if match_transaction:
                    transaction = match_transaction.group(1)
            match_provenance = regex_provenance.search(line)
            if match_provenance:
                provenance = match_provenance.group(1)
                break  # Reading the whole file is not necessary.
    return (transaction, provenance)


def fmt_transaction(transaction, width=3):
    """Format dips-2023-12 as dips-2023-012."""
    parts = transaction.split("-")
    return f"{parts[0]}-{parts[1]}-{int(parts[2]):0{width}d}"


def fmt_provenance(provenance):
    """Make lowercase, replace spaces by underscores, remove funny chars."""
    table = str.maketrans("", "", string.punctuation)
    return (
        provenance
        .lower()
        .replace("-", " ")  # hide "-" from translate
        .translate(table)
        .replace(" ", "_")
    )


if __name__ == "__main__":
    parser = argparse.ArgumentParser(
        formatter_class=argparse.RawDescriptionHelpFormatter,
        description="Copy dipsExport files, giving them nice names",
        epilog=textwrap.dedent("""\
            examples:

            $ renamevzs c845750e-9296-11ee-8821-005056871b7c.xml
            dips-2023-012_ministerium_für_zauberei.dips.xml

            $ renamevzs -w 5 c845750e-9296-11ee-8821-005056871b7c.xml
            dips-2023-00012_ministerium_für_zauberei.dips.xml
            """))
    parser.add_argument(
        "files",
        nargs="*",
        help="dipsExport XML files")
    parser.add_argument(
        "-d", "--dest",
        default=None,
        help="destination directory (default: same as input file)")
    parser.add_argument(
        "-m", "--move",
        action="store_true",
        help="move, don't copy")
    parser.add_argument(
        "-w", "--width",
        type=int,
        default=3,
        help="width of the last part of the transaction number (default: 3)")
    args = parser.parse_args()
    for f in args.files:
        rename(f, dest=args.dest, move=args.move, width=args.width)