add endpoint for variable transform/add new column

[raprasad]

This is a substitute for the R transform on the Dallas version of TwoRavens

For example, if I want to transform a variable, eg log(x), I’ll need to compute log(x), get back the new variable, preprocess it individually, and embed that metadata in the json

example input:

• preprocess id
• existing variable name
• type of transform
  • have predefined transforms. e.g. users passes name of operation

example output:

• full preprocess file with the new variable included

[raprasad]

example scenario: square values in the at_bats column

• open at_bats column in a dataframe. e.g. open source file but only 1 column
• use pandas for transform; to square the column
• send dataframe for processing
  • this is a new PreprocessJob!
• combine the metadata from the new PreprocessJob and original PreprocessJob metadata file
  • update:
    • dataset: number of variables
    • add new variable to variables section
    • add new variable to variable_display section
• delete the PreprocessJob for one column