treeverse / treeverse/dvc

Allow customization of the temporary directory for `dvc exp --queue`

Open
#8,980 4 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

A: experiments
Dominant language
Python
Stars
15.9k
Forks
1.3k
PR merge metrics
No merged PRs in 30d

Description

Hi team 👋

Thanks for the wonderful tool!

I am exploring a use-case for DVC along with other tools which I rely upon in my projects and was wondering whether there's a way to customize the temporary directory for the experiments queue mechanism.

I have tried adding the --temp command but the results seems to be the same and each experiment still runs in a separate environment which downloads the large genomic files into each of those experiments.

$ dvc exp run -S "annotation_tool=dfast,prokka" -S "outdir=results" --queue [[--temp]]

The feature request here is the (optional) ability to run the experiments in the same directory as opposed to individual directories (sensible default).

Happy to provide other details if needed.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start at the dvc exp run --queue entry point and trace how the experiment queue creates separate temporary environments. Reproduce the genomic-file download behavior with the command shown in the issue, then determine where an optional shared temporary-directory setting belongs. Done means queued experiments can use the same directory when requested while retaining separate directories by default.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
cli
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.