Skip to content

Compute worker reference data folder - #2524

Open
ObadaS wants to merge 3 commits into
developfrom
compute_worker_reference_data
Open

ObadaS wants to merge 3 commits into
developfrom
compute_worker_reference_data

Conversation

@ObadaS

@ObadaS ObadaS commented Aug 28, 2026 •

Copy link
Copy Markdown
Collaborator

Description

Add a folder to put reference data in on the compute worker, similar to /codabench/data. This folder will only be mounted during the scoring phase.

Checklist

  • Code review by me
  • Hand tested by me
  • I'm proud of my work
  • Code review by reviewer
  • Hand tested by reviewer
  • CircleCi tests are passing
  • Ready to merge

@ObadaS
ObadaS marked this pull request as ready for review August 28, 2026 07:22
@ObadaS
ObadaS force-pushed the compute_worker_reference_data branch from 27c3515 to a31df34 Compare September 1, 2026 12:54
@Didayolo Didayolo self-assigned this Sep 8, 2026
@Didayolo
Didayolo self-requested a review September 8, 2026 12:25
@Didayolo

Didayolo commented Sep 8, 2026

Copy link
Copy Markdown
Member

Note: test that the folder is not accessible during ingestion

@ObadaS
ObadaS force-pushed the compute_worker_reference_data branch 4 times, most recently from 332aee4 to b992bef Compare September 14, 2026 13:51
@ObadaS
ObadaS force-pushed the compute_worker_reference_data branch 2 times, most recently from e77d1b3 to 0e2a05c Compare September 29, 2026 09:26
@ObadaS
ObadaS force-pushed the compute_worker_reference_data branch 2 times, most recently from 78b4cfc to 07d452a Compare October 8, 2026 07:17
@Didayolo
Didayolo force-pushed the compute_worker_reference_data branch from 07d452a to c952847 Compare October 9, 2026 13:40
@Didayolo

Didayolo commented Oct 9, 2026 •

Copy link
Copy Markdown
Member

Change (fix I made)

Line 526

        self.data_dir = os.path.join(Settings.HOST_DIRECTORY, "data")  # absolute path to data in the host
        self.reference_dir = os.path.join(Settings.BASE_DIR, "reference")  # path inside the worker container

Instead of:

        self.data_dir = os.path.join(Settings.HOST_DIRECTORY, "data")  # absolute path to data in the host
        self.reference_dir = os.path.join(Settings.HOST_DIRECTORY, "reference")

Why

self.reference_dir is used with the _get_host_path convertion:

Line 1283

            # Reference data, only for the scoring phase.
            volumes_host.extend([self._get_host_path(self.reference_dir)])

While self.data_dir is not:

Line 1262

        volumes_host = [
            self._get_host_path(program_dir),
            self._get_host_path(self.output_dir),
            self.data_dir,
            self._get_host_path(self.root_dir, "submission")
        ]

By the way, we could update the code here (line 1262) to convert self.data_dir in the same way we convert the other paths.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants