orphaned_files

#!/usr/bin/env python3
# Copyright 2019 Alex K (wtwf.com)
# PYTHON_ARGCOMPLETE_OK

"""Program to do the thing.

"""

__author__ = "wtwf.com (Alex K)"

import argparse
import collections
import html
import logging
import os
import urllib.parse

import argcomplete


def main():
    """Parse args and do the thing."""
    logging.basicConfig()

    parser = argparse.ArgumentParser(description="Program to do the thing.")
    parser.add_argument(
        "--music-folder", help="Location of iTunes media", default="/Volumes/d8/iTunes"
    )
    parser.add_argument("-v", "--verbose", help="Log verbosely", action="store_true")
    parser.add_argument("-d", "--debug", help="Log debug messages", action="store_true")
    parser.add_argument(
        "--no-get-data",
        action="store_false",
        default=True,
        dest="get_data",
        help="Collect the data first",
    )

    argcomplete.autocomplete(parser)
    args = parser.parse_args()

    if args.verbose:
        logging.getLogger().setLevel(logging.INFO)
    if args.debug:
        logging.getLogger().setLevel(logging.DEBUG)

    logging.info("Log")

    if args.get_data:
        get_data(args)

    find_orphaned()


def get_data(args):
    print("Getting Data from iTunes Music Library.xml")
    library = os.path.expanduser("~/Music/iTunes/iTunes Music Library.xml")
    os.system(
        f"egrep -o 'file:///[^<]+' '{library}' "
        "> /tmp/itunesmusiclibrary.locations.urlencoded.txt"
    )
    print("Finding Files")
    os.system(f"(cd {args.music_folder}; find . -type f > /tmp/itunes_real_files.txt)")


def load_itdb_files():
    items = []
    prefix_length = len("file:///Volumes/D8/iTunes/")
    with open("/tmp/itunesmusiclibrary.locations.urlencoded.txt") as infile:
        for line in infile:
            item = html.unescape(
                urllib.parse.unquote(line[prefix_length:].strip())
            ).lower()

            if item:
                items.append("./" + item)
    return items


def read_real_files():
    items = []
    with open("/tmp/itunes_real_files.txt") as infile:
        for line in infile:
            items.append(line.strip().lower())
    return items


def find_orphaned():
    real_files = read_real_files()
    itdb_files = load_itdb_files()
    real_files = [
        fil
        for fil in real_files
        if fil[-10:] != "/.ds_store"
        and fil[:39] != "./automatically add to itunes.localized"
        and fil[:8] != "./tones/"
        and fil[:22] != "./mobile applications/"
        and fil[:27] != "./.itunes preferences.plist"
    ]

    real_files_dict = dict.fromkeys(real_files, True)

    missing = []
    for fil in itdb_files:
        if fil == "./":
            print("FOUND", fil)

        if fil in real_files_dict:
            real_files_dict[fil] = False
        else:
            missing.append(fil)

    print(f"Missing count: {len(missing)}")
    orphans = [key for key, value in real_files_dict.items() if value]
    orphan_dirs = collections.defaultdict(int)
    for orphan in orphans:
        orphan_dirs[os.path.dirname(orphan)] += 1

    print(f"Orphans count: {len(orphans)}")
    write_file("/tmp/orphans.txt", orphans)
    write_file("/tmp/missing.txt", missing)

    orphan_dirs = {
        k: v for k, v in sorted(orphan_dirs.items(), key=lambda item: -1 * item[1])
    }

    for dirname, count in orphan_dirs.items():
        if count > 3:
            print(f"{count:4d} {dirname}")


def write_file(filename, items):
    print(f"writing: {filename}")
    with open(filename, "w") as outfile:
        for item in sorted(items):
            print(item, file=outfile)


if __name__ == "__main__":
    main()