Rename MP3 Files

rename_mp3s.py

#!/usr/bin/env python3

import os
import mutagen.id3
import unicodedata
import re
import errno
from shutil import copyfile

OUTPUT_FILENAMES = set()
DUPLICATES = set()


def slugify(filename):
    """
    Heavily inspired by:
    https://stackoverflow.com/questions/295135/turn-a-string-into-a-valid-filename/46801075
    """
    filename = str(filename)

    # Convert to ASCII
    filename = (
        unicodedata.normalize("NFKD", filename)
        .encode("ascii", "ignore")
        .decode("ascii")
    )

    # Replace ampersands with 'and'
    filename = re.sub(r"&", "and", filename)

    # Remove characters that aren't alphanumerics, underscores, or hyphens
    filename = re.sub(r"[^\w\s-]", "", filename)

    # Convert spaces or repeated dashes to single dashes
    filename = re.sub(r"[-\s]+", "_", filename)

    # Also strip leading and trailing whitespace, dashes, and underscores
    return filename.strip("-_")


def is_mp3(filename):
    return filename.lower().endswith(".mp3")


def mkdir(path: str) -> None:
    """Make a directory recursively
    Functionality should be equivalent to `mkdir -p`.
    """
    try:
        os.makedirs(path)
    except OSError as e:
        if e.errno != errno.EEXIST or not os.path.isdir(path):
            raise


def process_file(directory, filename):
    fullpath = os.path.join(directory, filename)

    if not is_mp3(filename):
        print("Not an MP3, skipping: {}".format(fullpath))
        return

    print("Processing file: {}".format(fullpath))

    metadata = mutagen.id3.Open(fullpath)
    artists = metadata["TPE1"]
    title = metadata["TIT2"]

    new_filename = "{} {}".format(title, artists)
    new_filename = slugify(new_filename)
    new_filename = "{}.mp3".format(new_filename)

    print("New filename: {}".format(new_filename))

    new_directory = "{}.renamed".format(directory)
    mkdir(new_directory)

    new_fullpath = os.path.join(new_directory, new_filename)

    # Detect duplicates so we can warn at the end
    global OUTPUT_FILENAMES
    global DUPLICATES
    if new_fullpath in OUTPUT_FILENAMES:
        DUPLICATES.add(new_fullpath)
    else:
        OUTPUT_FILENAMES.add(new_fullpath)

    copyfile(fullpath, new_fullpath)


def process_files(directory):
    for root, _, files in os.walk(directory):
        for f in files:
            process_file(root, f)


def main():
    # Download files from beatport and unzip them into a directory structure such as:
    # Temp/
    # |_ Beatport/
    #    |_ Song 1.mp3
    #    |_ Song 2.mp3
    #
    # The Temp directory should not contain any other mp3 files, either
    # directly or in a child directory. Run the script, which will copy the
    # folder structure of Beatport into Beatport.renamed, renaming all mp3
    # files based on metadata in the ID3 tags.
    process_files(".")

    global DUPLICATES
    for dupe in DUPLICATES:
        print(
            'WARNING: Duplicate output filename "{}", possible loss of data.  This can occur if two files have the same ID3 metadata for title and artist.'.format(
                dupe
            )
        )


if __name__ == "__main__":
    main()