Upstreaming 2024 Nyrkiö patches #27
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Ah. I think I know what this is: We use hunter together with Nyrkiö, and the latter has pytest-benchmark, so it successfully went through all our own testing. I'll addpytest-benchmark to poetry files and try again. (But not today.) |
|
All the changes look good! We just need to remove |
|
Ok actually I think you need to restart the test. I'm not a committer yet :-D |
This can be used to compare whether an AnalyzedSeries object is more recent than the set of change points it was computed from. (Think cache invalidation, even if the AnalyzedSeries isn't necessarily a cache.)
Adds dependency pytest-benchmark
Hunter modified e-divisive such that it first does a pass using a higher p-value, then filters out all change points that have a higher p-value than the actual max_pvalue specified by the user. The initial higher pvalue is max_pvalue * 10. This means that for values higher than 0.1, the first pvalue is > 1.0. This doesn't make sense. In fact 1.0 also doesn't make sense because now every point is a weak change point. This patch modifies the call to split() such that: max_pvalue: first pass pvalue 0.0 - 0.05: max_pvalue * 10 (unchanged) 0.0 - 0.5 : max_pvalue * 2 0.5 - 1.0 : max_pvalue Since values above 0.1 didn't really make sense before this patch, the area where this could cause changes is for users using max_pvalue between 0.05 and 0.1. The merge() phase should however eventually produce approximately the same set of change points anyway, but this isn't guaranteed.
This is the default in AnalysisOptions, which most users would use, since it is required in the typical code path.
The common case is to add new data points to the end of the series. In this case we don't need to recompute all change points, we can just compute window_len points from the end. We do roughly 2 * window_len for good measure.
henrikingo
force-pushed
the
to-asf-upstream2
branch
from
03f22e9 to
e616f64
Compare
Never mind. It was just so fast I didn't realize it had ran already. |
Since it is random by design, can't really unit testas usual.
Gerrrr
approved these changes
Jan 10, 2025
smccarthy788
approved these changes
Jan 10, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.