Skip to content
This repository has been archived by the owner on Jun 22, 2024. It is now read-only.

pyWATTS 0.3.1

Pre-release
Pre-release
Compare
Choose a tag to compare
@benHeid benHeid released this 31 Mar 12:16
· 40 commits to master since this release
v0.3.1
06908b2
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

0.3.0 - 2022-31-03

Added

  • Add average module for pre-processing (#166)
  • Metric can be calculated on cutouts (#149)
  • Add custom scaler module for pre-processing (#167)
  • Add a MASE Summary (#148)
  • Add ensemble module for post-processing (#165)
  • Online Start Feature (#189)
  • Enable Multiple RefitConditions (#192)
  • Periodic Refit Condition (#192)
  • Add AnomalyGeneration modules (#195)
  • Add get_min_data method in modules to enable them to specify how much data they need. (#147)

Changed

  • Retraining is triggered after all steps are transformed (#117)
  • The sklean-wrapper includes now methods based on the SelectorMixin (#185)

Fixed

  • Try to handle non matching shapes in metrics. If not possible raise an exception. (#154)
  • Fixed Sampler-module, samples now in forwarding direction (#174)
  • Fixed wrong check of if summary should be returned in _run method of pipeline (#157)
  • Fixed Pipeline.to_folder() for modules that store parameters in numpy format by adding a special JSON-encoder for types that are not json-serializable by default (#175)
  • Added missing summaries to pywatts.summaries.init.py (#183)
  • Naming in the Metric Summaries (#193)
  • Fixed wrong time dimension name in ComputationMode.Transform (#199)
  • Fixed online execution of subpipeline (#203)
  • Enable to store a Keras Model with custom objects as custom loss functions (#17)
  • Fixed online execution of stacked pipelines (#213)

Deprecated

Removed

  • MAECalculator and RMSECalculator
Assets 2
Loading