remove unnecessary tests #175
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
| @@ -340,16 +340,16 @@ def test_fractional_diff(pd_X): | |||
|
|
|||
|
|
|||
| ### Temporarily commented out. Uncomment when benchmarking is ready. ### | |||
There was a problem hiding this comment.
I guess you can remove this comment too
|
Hey there, thanks for the PR. The proposed solution is to test less regressors for the moment? It can work right now to speed up the feedback loop. I am not super expert on our internals but I think in the long run we should add to our roadmap to refactor the tests again to run forecasters on smaller datasets (eg the commodities data? or a fake one with 10 series 100 observations) to make sure they actually fit. What do you think? @topher-lo too I'd love to hear your opinion on this. As we mentioned in the discord too, tests run too slow and fail (after 3/4 hours). We could run the benchmarks before every release or in a separate repo like polars does. |
|
We can disable most of the expensive datasets |
|
Also would recommend splitting up feature extraction and forecasting tests into two separate workflows. |
|
Do you think we should do this now? Can we change the test dataset to commodities? How do we know what set of tests we can run on each PR? |
Ignoring a set of slow tests by default and some other small cleanup