Repository navigation
Why do we copy Flow fields instead of using retrieved Flow? #437
Description
When automatically performing a Flow on a Task, we need to get information about the Flow from the server, if it exists there already. If it exists, we retrieve this information from the server and copy it to our local flow object here.
Why do we not use the Flow we get from get_flow, but rather copy its information?
Activity
mfeurer commentedon Apr 9, 2018
That's the PR which introduced this behavior: https://github.com/openml/openml-python/pull/253/files
And that's part of the reasoning behind it: #253 (comment)
If you could do a PR which explains the reasoning here in form of a comment, that would be appreciated.
PGijsbers commentedon Apr 9, 2018
Do you mean that when I commit the PR for Regression, you'd like me to have added a comment to the source code explaining why it's done?
Or should I set up a PR with just this minor code documentation change?
mfeurer commentedon Apr 9, 2018
For me it's fine if you add it to the regression PR.
PGijsbers commentedon Apr 9, 2018
All right, I'll do that.