New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Pairwise transformers tests #1096
Pairwise transformers tests #1096
Conversation
I think this is still failing because the automated tests in the |
Yea I was going to ask you about this since that was what I was about to fix and realised I didn't know what you wanted. Because this is inheriting BaseEstimator there are standard test expectations such as _is_fittedm fit(), predict() etc. Was just wondering why it is inheriting BaseEstimator if these aren't being used? |
The useful functionality we want is I think the distances set the precedent for objects that do not need fitting, so it may require careful thinking what to do about the tests. An easy way may be to just manually specify the methods that the distance is expected to have - or, if that is not easy, just skip some of the tests if you see a distance/kernel. |
quick question: do you know what to do here, or should I have a look? |
Sorry yea should be all working now had to write some extra tests for the panel data |
I've made some changes to BasePairwiseTransformerPanel that you may not like btw. The comment for what BasePairwiseTransformerPanel supports was numpy or List of dataframes, but it was only supporting List of dataframes so I just added a method to allow it support List of dataframes, Numpy of dataframes, and 3d Numpy - maybe wasn't your intention can remove very easily if you don't want it. |
Yes, I'm fine with the changes to the input conversion - this is something we may anyway refactor in the near future, but good for now as it covers a potential problem. Also happy with the changes to the Regarding the Maybe being a bit overcautious here, but I'd consider some adherence with "defensive programming" principles as good practice. Feel free to merge either way though. |
Reference Issues/PRs
Tests to go along side the PR #1071 @fkiraly
What does this implement/fix? Explain your changes.
Implements basic tests for general transformers and specific tests for the scipy_dist pairwise transformer.
Does your contribution introduce a new dependency? If yes, which one?
What should a reviewer concentrate their feedback on?
Is this enough I can add more but felt for now keep it simple until we see where this goes fully.
Any other comments?
PR checklist
For all contributions
For new estimators