Data Export

PyRolL includes a class capable of converting the simulation results to a pandas DataFrame and save this to different file formats. The feature can be accessed by use of the CLI through the export command.

The data included in the frame can be modified by hooks. The available file formats can be extended by the use of hooks.

To read about the basics of hooks and plugins, see here.

Specifying data to include

There is a hook columns(unit : Unit) that can be used to specify the columns included in the data frame. One can use the pyroll.utils.hookutils.applies_to_unit_types(types) decorator to specify the unit types the hook implementation should apply to (currently only Unit, RollPass, Transport).

Each implementation must return a mapping of column names (string) to values (any type that can be data in a DataFrame). The list of hook results will be combined to the final set of columns. Later registered implementations will override earlier ones.

Define new implementations of this hook to include more data in the export. Commonly you would return a dict mapping from str to a numeric type or string.

Adding new file formats

For exporting to a file a hook is defined to handle the formatting:

export(data: pandas.DataFrame, export_format: str)

It takes the generated DataFrame and a string specifying the format as arguments. Depending on the value of export_format an implementation can decide whether it is able to handle the format or not. If it can, it should return the binary data that will be saved to file. If it can not, it should return None. The first implementation not returning None will be used for the file content (firstresult).

Current basic implementations support CSV and XML formats by use of the methods provided by DataFrame.

Class Documentation

class Exporter

Class able to export simulation results to several data formats.

export(units: List[Unit], export_format: str) → bytes

Call get_dataframe and export its results to a specified format.

Parameters

• units – list of units to take the data from

• export_format – a string key identifying the export format, valid values depend on the loaded implementations of the ‘export’ hook

Returns

the exported data as binary stream

get_dataframe(units: List[Unit]) → DataFrame

Generate a pandas DataFrame by use of the unit_columns, roll_pass_columns and transport_columns hooks.

Parameters

units – list of units to take the data from

Returns

a pandas data frame filled with the exported data

Hooks

columns(unit) → Mapping[str, Any]

Take a unit object and extract some data to be listed in the CSV output. Return a mapping of column names to values. All hookimpls will be joined in order of definition.

export(data: DataFrame, export_format: str) → bytes

Export the data to a specified format. Return binary data that can be saved to a file. First hookimpl that does not return None is taken. Return None to signal, that the impl does not support the export_format