Scraping and grabbing Now! albums
Recently a collegue at work came to me to download them an album from online, unfortunately as it was a compilation album and the individual tracks had been released a million times already this wasn’t to be released through the usual channels.
No matter though, vague scripting to the rescue! The tracklist that I was after was available on the now website which had no issues being scraped.
source=$(wget https://www.nowmusic.com/album/now-rock-n-roll/ -qO-)
artists=$(printf "$source" | grep artist | sed 's/^.*>\([^<]*\)<.*$/\1/')
titles=$(printf "$source" | grep \"title\" | sed 's/^.*>\([^<]*\)<.*$/\1/')
paste <(printf "$artists") <(printf "$titles") | sed -e 's/\t/ - /g' > parse_list.txt
Now we have all 73 tracks in a single text file, no fuss, no muss.
All of these tracks are incredibly likely to be uploaded to youtube, so we can grab them using the ever-excellent youtube-dl
To manage this, we’ll run a youtube search on every entry, and grab the resulting output, converting it to mp3
along the way.
while read line; do youtube-dl -x --audio-format=mp3 ytsearch:"$line lyrics"; done < parse_list.txt
Please note, I append a " lyrics" in the search string to avoid too obvious music videos that sometimes have
With this, we have 73 mp3
files dumped into our working directory with messy filenames. I usually throw these into beets
in singleton mode via docker to improve the quality of the filenames/tags.
docker run -it -v $(pwd):/music linuxserver/beets bash
beet im -s /music
This will take some time, and will need a lot of nannying as there are no existing tags to work with initially. After the process however you’ll be rewarded with tagged files ready to (rock ’n) roll.