Currently, the job wsi-artifacts in CI is taking too long. Here below I report a couple of job runs stats.

Size

liver-1.svs size: 362M
cmu-3.svs size: 242M

Context 1:

https://github.com/histolab/histolab/runs/4069757248?check_suite_focus=true
Total time: approx. 32 min.

time liver-1.svs: 30 min (419 KB/s)
time cmu-3.svs: 1 min (3.25 MB/s)

Context 2:

https://github.com/histolab/histolab/runs/3951659424?check_suite_focus=true
Total time: approx. 38 min.
time liver-1.svs: 36 min (289 KB/s)
time cmu-3.svs: 2 min (2.47 MB/s)

Now, it seems clear that the problem is not the size of the WSI but rather the tggates server speed, or the way we are downloading from there.

Possible follow-ups:

1. We change server and slide - and update tests expectations
2. We try to find the same slide on a different server - unlikely
3. We play with ftp parameters of WGET to see if we can speed things up
4. We don't care and we close this issue

@nicolebussola and @ernestoarbitrio I need inputs from the both of you 🙏