tom/soundcloud-scrape
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Check if the user exists through their username element

Browse Source
This commit is contained in:
Thomas Wade 2018-10-01 10:03:37 +09:30
parent 61df852e63
commit ba9be2ee3f
1 changed files with 14 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
scrape.py
View File

@ -138,9 +138,14 @@ def Get_Basic_Info(user):
# Grab their username # Grab their username
# Use stripped_strings generator as workaround for users with premium badge # Use stripped_strings generator as workaround for users with premium badge
username = next(soup.find(class_='profileHeaderInfo__userName').stripped_strings) username_elem = soup.find(class_='profileHeaderInfo__userName')
user.username = username if username_elem:
print(' Username: ' + user.username) username = next(username_elem.stripped_strings)
user.username = username
print(' Username: ' + user.username)
else:
print(' No username, cannot be a valid user')
return False
# Grab their track count # Grab their track count
track_count_elem = soup.find('a', href='/' + user.url_username + '/tracks', class_='infoStats__statLink') track_count_elem = soup.find('a', href='/' + user.url_username + '/tracks', class_='infoStats__statLink')
@ -166,6 +171,8 @@ def Get_Basic_Info(user):
user.like_count = Human_Str_To_Int(str(like_count_elem.string.split(' ')[0])) user.like_count = Human_Str_To_Int(str(like_count_elem.string.split(' ')[0]))
print(' Likes: ' + str(user.like_count)) print(' Likes: ' + str(user.like_count))
return True
# Gets a list of fully-populated Track objects # Gets a list of fully-populated Track objects
def Get_Tracks_Info(url, limit = 100): def Get_Tracks_Info(url, limit = 100):
# Get the page and soup it # Get the page and soup it
@ -384,8 +391,10 @@ if not args.dump:
user.url_username = user_str user.url_username = user_str
print(' URL Username: ' + user.url_username) print(' URL Username: ' + user.url_username)
# Get their basic info # Get their basic info and check if it was successful
Get_Basic_Info(user) if not Get_Basic_Info(user):
print(' Failed to get basic info, skipping')
continue
# Get their tracks # Get their tracks
for track in Get_Tracks_Info('https://soundcloud.com/' + user.url_username + '/tracks', USER_MAX_TRACKS): for track in Get_Tracks_Info('https://soundcloud.com/' + user.url_username + '/tracks', USER_MAX_TRACKS):