Call a Custom Function

You can even use a custom function that operates on each element of the parent directory to add the outputs as classifiers. Do this my adding the names of the classifier columns, defining the function call, and adding any needed arguments in the form of a dictionary. For example, if the function is:

def function_handle(directory, args_dict):

    use_directory = directory
    output1 = random.randint(0, args_dict['par1'])
    output2 = random.randint(args_dict['par1'], args_dict['par2'])

    return [output1, output2]

Create the object,

import lsframe as ls

lsobject = ls.start(path,
                    patterns={'sample': '\d\d'},
                    skip='sample7',
                    date_format='any'
                    classifiers=['output1', 'output2'],
                    function=function_handle,
                    function_args={'par1': 1,
                                   'par2': 2}
                    )

Call the drive() method

lsobject.drive()

and two new columns would be added called "output1" and "output2" with the values corresponding to the function outputs. Make sure to have the function accept a path and a single dictionary that contains any additional parameters needed. Also make sure the function returns the outputs in a list that is equal in length to the given list of classifiers. Use the above example function as a template.

You may also define the classifiers, function, and function_args in the call to .drive(). This will add to the start object if these attributes do not already exist, or will re-define them. For example, the following is equivalent to above,

import lsframe as ls

lsobject = ls.start(path,
                    patterns={'sample': '\d\d'},
                    skip='sample7',
                    date_format='any'
                    )

lsobject.drive(classifiers=['output1', 'output2'],
               function=function_handle,
               function_args={'par1': 1, 'par2': 2}
               )

Handling Errors

If the custom function gives an error, for example if it tries to operate on a file/folder in the path that is not compatible, it will return the string "null" for the classifier. This can be useful for avoiding tedious reorganizing of directories. Simply run lsobject.drive(), collect a frame full of successful runs or "null", then use something similar to,

lsobject.frame = lsobject.frame[lsobject.frame['output1'] != 'null']

to reduce the frame to only compatible files/folders.