1.1.0 release notes
This release further solidifies our product footprint with a more efficiently managed code base, a demo environment for prospects to test our product, and new functionality to enable users to understand and prepare data for machine learning.
High-level updates include:
- Code modularization, user experience improvements, and extended ML platform support
- New lab environment to test AI-Link, contact ailink@atscale.com for access
- Exploratory Data Analysis (Beta) functions for users to analyze key trends in data
- Principal Component Analysis (Beta) method to reduce the dimensionality of datasets and enable selection of the most important features for ML training
It should be noted that there are some syntax changes to method calls in
AI-Link that are different from version 1.0.0. Please refer to the
/api/index
for the latest method syntax that should be used. Some key
changes as examples include (but are not limited to):
-
One hot encoding syntax has an 's' at the end of the string:
create_one_hot_encoding_features
-
Operations related to Semantic Model Features now live within the features_utils submodule. For example, create_calculated_column now lives in feature_utils and takes new parameters, including an instance of the new DataModel class:
create_calculated_column(data_model: DataModel, dataset_name: str, column_name: str, expression: str, publish: bool = True)