search

kube-HPC / hkube

🐟 High Performance Computing over Kubernetes - Core Repo 🎣
http://hkube.io
MIT License
305 stars 20 forks source link

UX Enhancments - some nitpicks, some more important #974

Closed brainoom closed 2 years ago

brainoom commented 3 years ago

  1. the double click for opening stuff in the table is non intuitive, to me.

  2. hkubectl is for k8s experts not algorithm developers. the names for stuff are hard to understand, and hkubectl is long to writhe make it hkc algo create , hk algo run . in the yaml file - use convention over configuration: yaml should be in the root, code path is current folder, and entry point should be hkube.py or something similar.

  3. .hkingore or treat .gitignore or .*ignore for ignoring files not to be zipped / uploaded as part of the algorithm.

  4. Run algorithm "add input" - missing the "gray text" that appear in other locations and say "this should be josn"

  5. The experiment UI is confusing. Should be a "pane"

  6. There should be a tab for single algorithm information -- all the runs of an algorithm. Mabye the experiment pane would be enough.

  7. For code/zip algorithms - let me download the code back to see what it is.

  8. Urls printed in api can prepend the current hkube url so the will be clickable.

  9. When updating algorithm to current version set current looks like a text box, and when pressing there is no indication for a few seconds that something is happening.

  10. The panes that open for details of stuff should have an x or arrow to close.

  11. Add a tooltip for the sidebar expansion button. Took me some time to understand that it is related.

  12. The search in the spec just highlight, but does not hide locations with no result - it's pretty useless this way.

  13. Tensorboard per algorithm, and not per pipeline. Compare between algorithm runs.

  14. Need to be able to see the command line run, and the environment variables in the run information.

  15. "User centric view" - I as a developer want to see what's relevant to me -- the algorithms I work on, the data I work on etc. Some concept of a "project" that has it's algorithms, pipelines, data, runs etc.

  16. Documentation order: The order is a bit confusing. starting with a pipeline before covering a single algorithm. Start with an algorithm a trivial example of creating an algorithm running it "add hoc", adding it, setting requirements, running with input etc. and only after you covered the algorithm move to a pipeline.

stale[bot] commented 3 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

stale[bot] commented 3 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

stale[bot] commented 2 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.