Specifying result objects and webpanelsΒΆ

The asr.core.ASRResult object is a way for you to document the what might be contained in the results that you recipe-step returns.

A typical realistic example of a Result object declaration can be seen below

import typing
from asr.core import ASRResult, prepare_result
from asr.database.browser import WebPanel, describe_entry


def webpanel(result, context):
    parameter_description = context.parameter_description(
        'asr.c2db.gs@calculate')

    energy = describe_entry(result.energy, parameter_description)
    prop_table = {'type': 'table',
                  'header': ['Property', 'value'],
                  'rows': [['energy', energy]]}

    return [
        WebPanel(title='Title of my webpanel',
                 columns=[[prop_table], []])
    ]


@prepare_result
class Result(ASRResult):
    """My ground state results object.

    These results are generated using...
    """

    energy: float
    forces: typing.List[typing.Tuple[float, float, float]]

    key_descriptions = dict(
        energy='The energy of the material.',
        forces='The forces on the atoms of the material.',
    )

    formats = {'webpanel2': webpanel}


# # Call like this
# myresult = Result.fromdata(energy=1.0, forces=[[1.0, 1.0, 1.0]])
# webpanels = myresult.format_as('ase_webpanel', row=..., key_descriptions=...)
# webpanel = webpanels[0]

There are several points to be made here

  • The Result class takes some type hints like energy: float to indicate the type of the data contained in the object.

  • The class also takes key_descriptions which eventually are used in the docstring for the class.

  • The formats attribute can be set to include additional formatting functions and in this case we want to include an ase_webpanel. The name is arbitrary but we are using this name in the ASR web application so it is important that you get it correct.

  • The @prepare_result decorator sets up the result class by automatically settings the docstring and more.

The webpanel formatter takes a result object an ase.db.AtomsRow object and a key_descriptions dictionary.