Skip to content
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.

Sign up for GitHub

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

Jump to bottom

[ENH] New detector base class #11

Merged
merged 75 commits into from
Aug 23, 2024
Merged

[ENH] New detector base class #11

merged 75 commits into from
Aug 23, 2024

Conversation

Tveten
Copy link
Collaborator

@Tveten Tveten commented Aug 22, 2024 •
edited
Loading

  • Update dependencies:

    • sktime >= 0.30. The sktime update is an API break to BaseSeriesAnnotator, which among other things removes the fmt and labels arguments. Since BaseSeriesAnnotator is still experimental, this PR refactors a lot of basic functionality.
    • No upper restriction on numpy and pandas.

  • Introduce BaseDetector, which inherits from BaseTransformer in sktime. See here sktime time series anomalies/changepoint interface upgrade, moving skchange to 2nd or 1st party? #8 for reasoning. This class is an alternative to BaseSeriesAnnotator as long as it is experimental, but aims to not stray too far away. It also removes the fmt and labels arguments.

  • Add base classes per detection type:

    • ChangeDetector
    • PointAnomalyDetector
    • CollectiveAnomalyDetector
    • SubsetCollectiveAnomalyDetector

  • Refactor all change and anomaly detectors to depend on a suitable subclass of BaseDetector.

  • Convert from black to ruff for formatting.

Tveten added 30 commits August 19, 2024 16:14
@Tveten
Update dependencies
0415c4f
@Tveten
Add first draft of BaseDetector
6987a38
@Tveten
Allow index names in dataframe inputs
fe87c36
@Tveten
Remove old commented code
e78b766
@Tveten
Rewrite BaseDetector such that subclasses define detection types
11a6a91 
Separates functionality regarding different detectors more clearly. Less clutter in BaseDetector.
@Tveten
Rephrasing module description
99cb1a3
@Tveten
Use separate integer labels per collective anomaly as dense output
56e5f2b
@Tveten
Fix bug
1c9e707
@Tveten
Add expected output typing in subclasses
25413ba
@Tveten
Improve documentation and remove old comments
666f05c
@Tveten
Improve module description
d0afbff
@Tveten
Move anomaly and change detector base classes to respective modules
afdd42a
@Tveten
Improve docstring: Mark optional methods
42cc9c7
@Tveten
Fix invalid type hints for pd.Series
3c8603e
@Tveten
Switch class dependence to internal BaseDetector
5abd326
@Tveten
Add utility functions for sparse output
430033f 
Can be reused when implementing _predict in subclasses
@Tveten
Set object_type tag to estimator
18c737c 
Checks for sktime compatibility with estimators, not subclasses.
@Tveten
Use parametrize_with_checks from sktime
00d23c5
@Tveten
Standardise predict output
b214b1d
@Tveten
Conform Pelt to BaseDetector
5da1cbd
@Tveten
Conform SeededBinarySegmentation to BaseDetector
a1c2b86
@Tveten
Inherit from BaseTransformer
7da353e 
To pass sktime check_estimator tests. It requires inheritance from an sktime.BaseEstimator subclass. Could be a good idea, since it passes as a transformer immediately.
@Tveten
Inherit from BaseTransformer rather than BaseEstimator
862b420 
Mainly to pass sktime conformance tests, which requires inheritance of a subclass of BaseEstimator. But aso useful to be a fully functional transformer.
@Tveten
Add base class tests
c6e7a4c
@Tveten
Add more tests
1c366e5
@Tveten
Remove duplicate test
96d164a 
Exists in test_detectors.py
@Tveten
Fix changepoint tests for new interface
0fb4992
@Tveten
Fix tests for new interface
b0e24bd
@Tveten
Fix tests for new interface
7387cc8
@Tveten
Delete change_detectors.utils module
2cd54d1 
None of the functions are in use anymore.
Tveten added 28 commits August 23, 2024 08:46
@Tveten
Fix bug in test
5322a76
@Tveten
Remove old commented fit method
cc0248a
@Tveten
Add tests for NotImplementedError
dd4ffc6
@Tveten
Add sparse_to_dense error test
d22f54d
@Tveten
Use ignore_point_anomalies in test
2ec9c1a
@Tveten
Add columns argument to sparse_to_dense
c9ebae5 
To allow for subset detectors
@Tveten
Fix CAPA docstring
760153b
@Tveten
Add SubsetCollectiveAnomalyDetector base class
e202a42
@Tveten
Conform Mvcapa to BaseDetector
78f50d2
@Tveten
Remove merge_anomalies
306cb40 
Unnecessarily complicated function
@Tveten
Remove old commented code
edd64a2
@Tveten
Allow sparse_to_dense and transform to mismatch on Series vs DataFrame
78c940c
@Tveten
Fix doctrings and typings
ac780ae
@Tveten
Fix tests
ac910b7
@Tveten
Clean up files
788064d
@Tveten
Tidy up tags
b4a5da2
@Tveten
Update readme
a922a90
@Tveten
Fix bug: Input columns being overwritten
6f6427f
@Tveten
Update benchmarking script
4ded197
@Tveten
Update the capa interactive scripts
2b49a1c
@Tveten
Fix docstring
7e2eeca
@Tveten
Update the circ bin seg interactive script
1b81f86
@Tveten
Update the moscore anomaly interactives cript
ab48424
@Tveten
Update the moscore interactive script
301183f
@Tveten
Update the PELT interactive script
953fb92
@Tveten
Update the seeded binary segmentation interactive script
0f9f1ce
@Tveten
Update the stat threshold anomaliser interactive script
b1cbb21
@Tveten
Remove reference to the old fmt argument
8777d95
@Tveten Tveten merged commit 6cd6796 into main Aug 23, 2024
6 checks passed
@Tveten Tveten deleted the new_detector_base_class branch September 3, 2024 13:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@Tveten