Python API

[BauerLab]

A python API needs to be added

[piotrszul]

Python API for generic variant-spark interface.

Should be implemented in variants module in python directory.

The main access point should be VariantsContext which can be used to load features and labels.
The API should be based on the 'non-depricated' scala api (from au.csiro.variantspark.api)

The general use pattern should he like this

from variants import VariantsContext
vc = VariantsContext(spark)
features = vc.import_vcf(input_path = "path-to-a-vcf-file")
label = vs.load_label(input_path = "path-to-a-csv_file", col_name = "column-name)
imp_analysis = features.importance_analysis(label, n_trees = 1000, m_try_fraction = 0.1, ...)
oob_error = imp_analysis.oob_error
imp_vars_df = imp_analysis.important_variables()
# imp_vars_df is a spark.sql.DataFrame

Some guidance on the implementation can be found in the variants.hail module.

[piotrszul]

API implemented.
Some technical debts reported as new issues.