Exceeded LFS Budget

[donutbrew]

We're having problems installing the assignment cache from the command line--running pangolin --add-assignment-cache gives a pretty opaque python exception.

Runnig pip install -vvv --upgrade "git+https://github.com/cov-lineages/pangolin-assignment.git@v1.35" reveals:

  Traceback (most recent call last):
    File "/private/var/folders/xk/0lc34xqs77bfj_p0shkxhz4m0000gp/T/pip-req-build-ji81d3xc/setup.py", line 20, in pull_first
      g.execute(['git', 'lfs', 'pull'])
    File "/Users/xxx/miniconda3/envs/pangolin-dev/lib/python3.12/site-packages/git/cmd.py", line 1406, in execute
      raise GitCommandError(redacted_command, status, stderr_value, stdout_value)
  git.exc.GitCommandError: Cmd('git') failed due to: exit code(2)
    cmdline: git lfs pull
    stderr: 'batch response: This repository exceeded its LFS budget. The account responsible for the budget should increase it to restore access.
  Failed to fetch some objects from 'https://github.com/cov-lineages/pangolin-assignment.git/info/lfs''

Can you say what the thaw time might be on this?

Interestingly, I can still download the assignments cache via the github web interface, but that doesn't really help with automated buids 😺

[AngieHinrichs]

Sorry about this. It's a resource issue (TL;DR GitHub started charging money for something and we're not paying enough yet) and we're working on it. [It's also breaking a CI test in the pangolin repo where I'm working on your other issue from today.]

[donutbrew]

No worries. I suspect it was giving me some problems earlier, and I didn’t recognize the issue. Just looked at the GitHub documentation to see the default bandwidth cap is 10 GiB/mo, which at 300MB a download may be pretty limiting.

In order to be more conservative with the download count, can you direct me to where I could copy an exiting assignment cache file? Or maybe add in an optional uri arg for adding the assignment cache from a mirror?

[AngieHinrichs]

It sounds like you were able to download the current assignments cache file via the github web interface, so perhaps you can copy that file into your docker image, and then run pangolin with --assignment-cache /path/to/usher_assignments.cache.csv.gz:

  --assignment-cache ASSIGNMENT_CACHE
                        Cached precomputed assignment file to use instead of default from pangolin-assignment
                        repository. Does not require installation of pangolin-assignment.

That means making a new docker image for each pangolin data release, sorry if you were hoping to avoid that, but it would mean a lot less downloading of the big assignments file.

The LFS budget is exceeded for the rest of October at the moment, but I'll try to get the new payment set up ASAP.

[AngieHinrichs]

Speaking of data-release-specific Docker images, there is staphb/pangolin (:4.3.1-pdata-1.35 from https://github.com/StaPH-B/docker-builds/blob/master/build-files/pangolin/4.3.1-pdata-1.35/Dockerfile).