`dvc exp list` (and other commands): always show full unique experiment hash
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 15.9k
- Forks
- 1.3k
- PR merge metrics
- No merged PRs in 30d
Description
I hit this issue with duplicated experiments (don't know yet why, any ideas? did we change the logic behind the naming recently @daavoo @dberenbaum ?):
@shcheklein ➜ /workspaces/ultralytics (dvclive) $ dvc exp list -A
21122ee:
armed-fuzz
testy-leno
d0f99f9:
level-tugs
mothy-jeer
weepy-haet
armed-fuzz
testy-leno
nosed-mold
dvclive:
testy-leno
withy-keek
5a2d27b:
nosed-mold
armed-fuzz
testy-leno
withy-keek
fd8435a:
testy-leno
@shcheklein ➜ /workspaces/ultralytics (dvclive) $ dvc exp remove armed-fuzz
ERROR: Ambiguous name 'armed-fuzz' refers to multiple experiments. Use one of the following full refnames instead:
refs/exps/5a/2d27b1aeb5575916e3480b1b77900b155002a1/armed-fuzz
refs/exps/d0/f99f94555e5488a2f64cd7680b2419d921eae6/armed-fuzz
refs/exps/21/122eefcbe711de67eea087c4f93e0566ce3dbd/armed-fuzz
dvc exp show -A doesn't show all commits in this case (can be orphaned commits and experiments in this case - I'm doing a lot of rebases, etc).
There is not easy way to disambiguate those from the listing (I don't know how to remove a specific one under a specific commit).
We should always everywhere show a Git hash for an experiment. Name should serve as tag, or as a human readable name, etc. Collisions should be fine.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the dvc exp list, dvc exp show -A, and dvc exp remove entry points described in the issue, reproducing the duplicate-name case. Trace how experiment references are formatted and resolved; done means listings expose a full unique Git hash wherever needed so a specific experiment can be selected unambiguously.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- git, python
- Domain
- cli, developer-experience
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100