Generators

Currently, there is only one generator DatasetGenerator which allows you to define schema for new datasets. This generator requires use of the `faker library <http://faker.readthedocs.io/>`_.

Generator Initialization Requirements

To initialize a generator you need the following:

parser:

a parsers.SchemaYAMLParser or parsers.SchemaCLIParser or a dictionary with the appropriate keys (see Required Parser Arguments below)

output:

a string specifying output (see: I/O Options)

Required Parser Arguments

The SchemaYAMLParser and SchemaCLIParser as well as any dictionary you use in leiu of a parser object need to have certain keys in order to generate the data. The required arguments are as follows:

schema:

a dictionary of column names and values to use. Optional values are any of the faker providers defined like 'faker.name', as well as lists or iterators of options or range and arange objects.

num_rows:

an integer for the number of rows to generate

The parser objects (or a dictionary you use) may also have the following optional arguments which are used to create timeseries:

start_time:

isoformat start time which will be used to generate a timeseries. Currently, the timeseries column defaults to ‘timestamp’.

end_time:

if desired, set an end_time in isoformat. Note: this may or may not be reached given the number of rows requirement.

increments:

choice from: ‘days’, ‘hours’, ‘seconds’ and ‘random’ for timestamp increments. Default is random which is a mix of days, hours and seconds.

For more examples on how to utilize these generators, check the Usage documentation.