KevinMidboe/spotify-downloader
1
0
Fork 0
mirror of https://github.com/KevinMidboe/spotify-downloader.git synced 2025-10-29 18:00:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

Improves accuracy of selected youtube video

Browse Source 
Finds a balance between viewcount (Youtube's relevancy) and proper song
duration based on Spotify duration. Thus, a Youtube video that is 30
seconds longer than the Spotify track will not be considered

Takes the first result from Youtube (using Youtube's original ordering
based on relevance) that has a similar duration to the Spotify song

Fixes a bug where if there were no suitable videos for a song, the
program would infinitely loop. The program will now retry to find a song
up to 5 times before moving on (this is necessary because occasionally the
song isn't properly fetched or parsed from Youtube)

Fixes bug where songs that are retried were appended to the playlist
file without being separated by a newline
This commit is contained in:
Luke Garrison
2017-08-09 21:42:31 -04:00
parent 7e0fdfbce3
commit 9d25197a5f
2 changed files with 58 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
core/misc.py
View File

@@ -103,9 +103,10 @@ def generate_search_url(song, viewsort=False):
# urllib.request.quote() encodes URL with special characters
song = quote(song)
if viewsort:
url = u"https://www.youtube.com/results?q={0}&sp=CAMSAhABUBQ%253D".format(song)
url = u"https://www.youtube.com/results?q={0}".format(song)
else:
url = u"https://www.youtube.com/results?sp=EgIQAQ%253D%253D&q={0}".format(song)
return url

73
spotdl.py
View File

@@ -53,8 +53,12 @@ def generate_metadata(raw_song):
return meta_tags
def generate_youtube_url(raw_song):
def generate_youtube_url(raw_song, tries_remaining=5):
"""Search for the song on YouTube and generate a URL to its video."""
# prevents an infinite loop but allows for a few retries
if tries_remaining == 0:
return
meta_tags = generate_metadata(raw_song)
if meta_tags is None:
song = raw_song
@@ -69,19 +73,30 @@ def generate_youtube_url(raw_song):
videos = []
for x in items_parse.find_all('div', {'class': 'yt-lockup-dismissable yt-uix-tile'}):
# ensure result is not a channel
if x.find('channel') is not None or 'yt-lockup-channel' in x.parent.attrs['class'] or 'yt-lockup-channel' in x.attrs['class']:
continue
# ensure result is not a mix/playlist
if 'yt-lockup-playlist' in x.parent.attrs['class']:
continue
# confirm the video result is not an advertisement
if x.find('channel') is None and x.find('googleads') is None:
y = x.find('div', class_='yt-lockup-content')
link = y.find('a')['href']
title = y.find('a')['title']
try:
videotime = x.find('span', class_="video-time").get_text()
except AttributeError:
return generate_youtube_url(raw_song)
youtubedetails = {'link': link, 'title': title, 'videotime': videotime, 'seconds':misc.get_sec(videotime)}
videos.append(youtubedetails)
if meta_tags is None:
break
if x.find('googleads') is not None:
continue
y = x.find('div', class_='yt-lockup-content')
link = y.find('a')['href']
title = y.find('a')['title']
try:
videotime = x.find('span', class_="video-time").get_text()
except AttributeError:
return generate_youtube_url(raw_song, tries_remaining - 1)
youtubedetails = {'link': link, 'title': title, 'videotime': videotime, 'seconds':misc.get_sec(videotime)}
videos.append(youtubedetails)
if meta_tags is None:
break
if not videos:
return None
@@ -100,10 +115,32 @@ def generate_youtube_url(raw_song):
return None
else:
if meta_tags is not None:
videos.sort(key=lambda x: abs(x['seconds'] - (int(meta_tags['duration_ms'])/1000)))
result = videos[0];
# filter out videos that do not have a similar length to the Spotify song
duration_tolerance = 10
max_duration_tolerance = 20
possible_videos_by_duration = list()
'''
start with a reasonable duration_tolerance, and increment duration_tolerance
until one of the Youtube results falls within the correct duration or
the duration_tolerance has reached the max_duration_tolerance
'''
while len(possible_videos_by_duration) == 0:
possible_videos_by_duration = list(filter(lambda x: abs(x['seconds'] - (int(meta_tags['duration_ms'])/1000)) <= duration_tolerance, videos))
duration_tolerance += 1
if duration_tolerance > max_duration_tolerance:
print(meta_tags['name'], 'by', meta_tags['artists'][0]['name'], 'was not found')
return None
result = possible_videos_by_duration[0]
else:
# if the metadata could not be acquired, take the first result from Youtube because the proper song length is unknown
result = videos[0]
full_link = None
if result:
full_link = u'youtube.com{0}'.format(result['link'])
full_link = u'youtube.com{0}'.format(result['link'])
return full_link
@@ -214,6 +251,7 @@ def check_exists(music_file, raw_song, islist=True):
# do not prompt and skip the current song
# if already downloaded when using list
if islist:
print('Song already exists')
return True
# if downloading only single song, prompt to re-download
else:
@@ -257,7 +295,7 @@ def grab_list(text_file):
misc.trim_song(text_file)
# and append it to the last line in .txt
with open(text_file, 'a') as myfile:
myfile.write(raw_song)
myfile.write(raw_song + '\n')
print('Failed to download song. Will retry after other songs.')
continue
except KeyboardInterrupt:
@@ -303,6 +341,7 @@ def grab_single(raw_song, number=None):
content = go_pafy(raw_song)
if content is None:
return
# print '[number]. [artist] - [song]' if downloading from list
# otherwise print '[artist] - [song]'
print(get_youtube_title(content, number))