v2.1.3

Include bugfix for hyperparameter optimization.
Removed teams dependency from optimize_hyperparameters.

v2.1.2

Incorporate bugfix basecase

v2.1.1

includes:

• check model age before retrieving input data
• github repo renamed

v2.1.0

Changelog

• Always sort the index after adding feature for multiple horizons
• Use the forecast tzinfo format when add the tAhead column
• Use pytz.utc for the create_basecase_forecast pipeline
• Add documentation on the expected/required keys in prediction jobs for multiple functions/methods
• Add option to set model_type when calling main() of
  • create_forecast task
  • train_model task

v2.0.3

• changed ktpbase dependency to openstf-dbc

v2.0.2

Fixes

• Fix wrong signature ModelCreator.create_model (used for optimizing hyperparameters)

v2.0.1

Fixes

• Fix Python 3.7, 3.8 compatibility
• Bugfix algorithm type tag

Pre-release-bugfix-algtype

v2.0.1a1

Update setup.py

v2.0.0

Changes:

• Pipelines based on sklearn (general and most important change)
• Logging in sklearn pipelines #72
• Integrate monitoring in sklearn pipelines #456
• Sklearn based hyperparameter optimization pipeline #83
• Sklearn compatible quantile model #73
• Modularlized sklearn pipeline #92
• New basecase forecast pipeline #88
• New fallback forecast #70
• Moved DB calls out of pipeline #74
• New model serializer #71
• Added backtest pipeline

Fixes:

• Bug in calculate completeness
• Updated unittests

Pre-release-sklearn-based-pipelines

Update setup.py

bump version number