Extract YTM playback history from Google Takeout JSON as CSV

ytmhist2csv.py

import json, re, sys

items = json.load(open('watch-history.json'))
music = (list(x for x in items if x['header'] == 'YouTube Music'))

known_music_channels = {
	'Reol Official': 'Reol'
}

counters = {
	'Total': len(music),
	'Native (GPM/YTM)': 0,
	'Known': 0,
	'Guessed': 0,
	'Fallback': 0,
	'Removed': 0,
	'Without metadata': 0,
}


for track in music:
	if not track['title'].startswith('Watched '):
		print(track)
		exit(1)

	fullname = track['title'][8:]

	parts = re.split('\\s+[–—-]\\s+', fullname)
	artist = 'unknown'
	title = 'unknown'

	if len(parts) < 2:
		if 'subtitles' in track and track['subtitles'][0]['name'].endswith(' - Topic'):

			# Native YTM/GPM artist
			artist = track['subtitles'][0]['name'][0:-8]
			title = fullname
			counters['Native (GPM/YTM)'] += 1
		elif 'watch?' in track['title']:


			# no metadata at all :shrug: 
			counters['Without metadata'] += 1
			continue 

		elif 'a video that has been removed' in track['title']:
			
			# :shrug: 
			counters['Removed'] += 1
			continue
		elif 'subtitles' in track and track['subtitles'][0]['name'] in known_music_channels:
			
			# known authors on youtube
			title = fullname
			artist = known_music_channels[track['subtitles'][0]['name']]
			counters['Known'] += 1
		else:
			
			# 3rd-party upload without ' - ' sign 
			# trying to guess

			# 1. Asian-style
			match_quotes_a = re.search('「.*?」', fullname)
			match_quotes_b = re.search('【.*?】', fullname)
			match_quotes_c = re.search('『.*?』', fullname)
			

			if match_quotes_a:
				title = match_quotes_a[0][1:-1]
			if match_quotes_b:
				artist = match_quotes_b[0][1:-1]

			if match_quotes_c:
				artist = match_quotes_c[0][1:-1]	
				title = fullname[match_quotes_c.end():].strip()

			# 2. artist: title style
			n_parts = re.split(':\\s+', fullname)
			if len(n_parts) == 2:
				artist = n_parts[0]
				title = n_parts[1]

			# 3. title by artist style
			n_parts = re.split('\\s+by\\s+', fullname)
			if len(n_parts) == 2:
				artist = n_parts[1]
				title = n_parts[0]

			if (title != 'unknown') and (artist != 'unknown'):
				counters['Guessed'] += 1	
				
	else:
		title = parts[1]
		artist = parts[0]
		counters['Guessed'] += 1

	# fallback
	if artist == 'unknown' or title == 'unknown':
		counters['Fallback'] += 1
	if artist == 'unknown' and 'subtitles' in track:
		artist = track['subtitles'][0]['name']
	if title == 'unknown':
		title = fullname

	# clean mess
	artist = re.sub('【.*】', '', artist)	
	artist = artist.strip()
	title = title.strip()

	#print('"{}","{}","","{}","",""'.format(artist,title,re.sub('\\.\\d{3}$', '', track['time'].replace('T', ' ').replace('Z', ''))))
	print(artist, '\t', title)

print('-----------------', file=sys.stderr)
for (counter, value) in counters.items():
	print('{}: {}'.format(counter, value), file=sys.stderr)