Allow file input in the command line for skip/only
Make the manhattan plot function less restrictive of the data passed into it
Use skip/only in the transform function
Categorization would silently fail if there was only one variable of a given type
Improvements to the CLI and printed log messages.
The functions from the ‘Process’ module were put into the ‘Modify’ module.
Datasets are no longer split apart when categorizing.
Extensive changes in organization, but limited new functionality (not counting the CLI).
Added a function to recode values - https://github.com/HallLab/clarite-python/issues/4
Added a function to filter outlier values - https://github.com/HallLab/clarite-python/issues/5
Added a function to generate manhattan plots for multiple datasets together - https://github.com/HallLab/clarite-python/issues/9
Add some validation of input DataFrames to prevent some errors in calculations
Added an initial batch of tests
Support EWAS with binary outcomes. Additional handling of NA values in covariates and the phenotype. Add a ‘min_n’ parameter to the ewas function to require a minimum number of observations after removing incomplete cases. Add additional functions including ‘plot_distributions’, ‘merge_variables’, ‘get_correlations’, ‘get_freq_table’, and ‘get_percent_na’
Add support for complex survey designs
Added documentation for existing functions
First functional version. Mutliple methods are available under a ‘clarite’ Pandas accessor.