The albums are stored at /var/tmp/albums/data
on node05
.
/var/tmp/albums/data/metadata.csv
lists all 100'000+ albums with their
metadata, such as ID, artist, genres, etc.
The data directory itself then contains a folder for each album which contains a folder for each artist of that album, which then contain the album covers. Note that we call the albums "masters" in the metadata CSV, a naming we adopted from the Discogs API. Should be the other way around but I messed it up in the download script.
(/var/tmp/group5) [studi9@node05 albums]$ ./tree
data/
metadata.csv # The metadata csv with all albums
1095565/ # Album ID because I messed up the order
81131/ # Artist ID
primary.jpg # Front Cover
478444/
2886362/
primary.jpg
176048/
469595/
secondary.jpg # Optional Back Cover
primary.jpg
1019012/
90235/
secondary.jpg
primary.jpg
1198129/
301147/
primary.jpg
673419/
304779/
secondary.jpg
primary.jpg
First, make sure that you are in the Unibe network. Use a VPN if necessary.
-
ssh into the cluster (provide the password from the screenshot in the whatsapp group)
$ ssh studi9@cluster.inf.unibe.ch
-
ssh into node05
$ ssh node05
-
Load the correct python environment
$ module load anaconda/3 $ source activate /var/tmp/group5/
Done!