Support for python's dataclass members #126
Description
Hi, I am currently documenting a python project with neogen and I am liking it very much so far!
One thing that I am currently missing is the support for dataclass attributes.
@dataclass
class RunSettings():
"""A collection of settings for an algorithm run.
Attributes:
algorithm: String resembling the type of algorithm used.
Currently, only 'nsga2' and 'ga' are supported.
population: Population size of the run.
generations: Number of generations of the run.
objectives: Dict with load values for each objective.
E.g. { "torque": 2e10, "compression": 4.6e4 }
architecture: Decides the used substructure architecture for the run.
Currently 'simple' or 'complex'.
constraints: SubstructureConstraints object that sets design
constraints for the algorithm run.
combine_stabilities: Bool deciding if the static stability test values
should be combined into one value.
probabilities: Dict with probabilities for mutation and crossover.
tracking: Bool deciding, if genealogical tracking information is saved
in design dicts throughout the run.
"""
algorithm: str = "nsga2"
population: int = 20
generations: int = 5
objectives: dict = field(default_factory=lambda: REAL_LIFE_LOADS)
architecture: str = "simple"
constraints: SubstructureConstraints = None
combine_stabilities: bool = True
probabilities: dict = field(default_factory=lambda: DEFAULT_PROBABILITIES)
tracking: bool = False
def __post_init__(self):
do_stuff()Currently, when generating the docstring for this class, only Attributes that are set in some of its functions are included in the list.
Do you think, this is complicated to implement? If I find time, I could also try to understand the source code of neogen and try to create a PR :)