normalize_unicode.py

#!/bin/python
from mutagen.flac import FLAC
from pathlib import Path
import sys

REPLACEMENTS = {
    '‘': "'",
    '’': "'",
    '“': '"',
    '”': '"',
    '…': '...',
    '–': '-',
    '—': '-',
    '\xa0': ' ',# non-breaking space
}

rename_queue = []
tag_edit_queue = []

for p in Path('..').glob('**/*'):
    if not p.is_file() or p.suffix != '.flac':
        continue

    curname = p.name
    newname = curname
    for orig, new in REPLACEMENTS.items():
        newname = newname.replace(orig, new)

    if curname != newname:
        print('Renaming file:')
        print('<', curname)
        print('>', newname)
        print()
        rename_queue.append((p, p.with_name(newname)))

    flac = FLAC(p)
    updated = False
    new_tags = {}

    for tag in ['artist', 'title', 'album']:
        if tag in flac:
            orig_val = flac[tag][0]
            new_val = orig_val
            for orig, new in REPLACEMENTS.items():
                new_val = new_val.replace(orig, new)
            if orig_val != new_val:
                print(f"Updating '{tag}' in {p.name}:")
                print('<', orig_val)
                print('>', new_val)
                print()
                new_tags[tag] = new_val
                updated = True

    if updated:
        tag_edit_queue.append((p, new_tags))

if not rename_queue and not tag_edit_queue:
    print('Nothing to do.')
    sys.exit(0)

response = input("Continue? (y/N): ").strip().lower()
if response != 'y':
    print('Aborted.')
    sys.exit(1)

for path, new_tags in tag_edit_queue:
    flac = FLAC(path)
    for tag, val in new_tags.items():
        flac[tag] = val
    flac.save()
    print(f"Updated tags in: {path.name}")

for old_path, new_path in rename_queue:
    old_path.rename(new_path)
    print(f"Renamed: {old_path.name} -> {new_path.name}")