
VerticaPy
Python API for Vertica Data Science at Scale
What's New in VerticaPy v0.4.1
New features:
User-Defined Functions
- Automatically generate your own customized UDF using the create_lib_udf function.
Geospatial
- New Function: use create_index to create geospatial indexes.
Machine Learning & Statistical Tools
- VerticaPy Pipelines are now available. These allow you to combine multiple VerticaPy models.
- OneHotEncoder has more parameters.
- The Regressors method regression_report now more options (Anova Table, Metrics...) and computes more metrics (R2 adjusted, RMSE).
- You can now compute the ANOVA table using the function anova_table.
- New statistical tests to verify Homoscedasticity: het_arch, het_breuschpagan, het_goldfeldquandt, het_white.
- New statistical tests to verify variables Normality: normaltest, kurtosistest, skewtest.
vDataFrame
- Removed: DataFrame.
to_vdf andread_vdf - New method: vDataFrame.to_pickle allows pickling vDataFrames.
GitHub & Unit Tests
- GitHub Action now allows for on-site testing.
- More unit tests are available.