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] interfacing pyts estimators #5850

Open
fkiraly opened this issue Jan 27, 2024 · 1 comment
Open

[ENH] interfacing pyts estimators #5850

fkiraly opened this issue Jan 27, 2024 · 1 comment
Labels
enhancement Adding new functionality interfacing algorithms Interfacing existing algorithms/estimators from third party packages module:classification classification module: time series classification module:clustering clustering module: time series clustering module:distances&kernels dists_kernels and distances modules: time series distances, kernels, pairwise transforms module:transformations transformations module: time series transformation, feature extraction, pre-/post-processing

Comments

@fkiraly
Copy link
Collaborator

fkiraly commented Jan 27, 2024 •
edited

pyts is a nice package with classifiers, distances, transformations, etc
https://github.com/johannfaouzi/pyts

We should consider interfacing it.

It is quite similar to tslearn in terms of interfaces, so copy-pasting the tslearn adapter and tinkering until it works might be a good approach.

Related post: #5849, re pyts ROCKET transformer
@fkiraly fkiraly added interfacing algorithms Interfacing existing algorithms/estimators from third party packages module:clustering clustering module: time series clustering module:classification classification module: time series classification module:transformations transformations module: time series transformation, feature extraction, pre-/post-processing module:distances&kernels dists_kernels and distances modules: time series distances, kernels, pairwise transforms enhancement Adding new functionality labels Jan 27, 2024
@fkiraly fkiraly mentioned this issue Jan 27, 2024
Merged
@fkiraly
Copy link
Collaborator Author

fkiraly commented Jan 27, 2024 •
edited

From #5851: indeed using a similar adapter works. pyts uses the numpyflat mtype internally, so the tslearn adapter mostly translates (only diff is mtype, tslearn uses numpy3D). Mid-term, one could think about refactoring both adapters to increase DRY-ness, if it should work for pyts.

fkiraly added a commit that referenced this issue Feb 15, 2024
@fkiraly
[ENH] pyts adapter and interface to pyts ROCKET (#5851)
7bde5e4 
Towards #5850, this PR adds a general interface towards `pyts`, and an
interface to `pyts` `ROCKET` (as implicitly suggested in
#5849).

`pyts` uses the `numpyflat` mtype internally, so the `tslearn` adapter
should mostly translate (only diff is mtype, `tslearn` uses `numpy3D`).
Mid-term, one could think about refactoring both adapters to incrase
DRY-ness, if it should work for `pyts`.

`pyts` is added to the `all_extras` dependency set. Notably, `pyts`
depends on `numba`, which means 3.11 or lower.

For now, I have avoided adding it to the other dependency sets, as it
might cause restrictions.
@fkiraly fkiraly mentioned this issue Mar 19, 2024
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement Adding new functionality interfacing algorithms Interfacing existing algorithms/estimators from third party packages module:classification classification module: time series classification module:clustering clustering module: time series clustering module:distances&kernels dists_kernels and distances modules: time series distances, kernels, pairwise transforms module:transformations transformations module: time series transformation, feature extraction, pre-/post-processing
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@fkiraly