Dataclass parsing

[samwaseda]

At the latest pyiron meeting I had claimed that data classes can be parsed automatically, but afterwards I realized that it was only partly true. Currently the plan is for the user to be able to write something like:

@semantikon_class
@dataclass
class Input:
    temperature: u(float, units="kelvin")
    number_of_steps: int

@semantikon_class
@dataclass
class Output:
    energy: u(float, units="electron_volt")
    volume: u(float, units="angstrom**3")

@Workflow.wrap.as_function_node
def run_md(inp: Input) -> Output:
    out = Output(energy=1., volume=2.)
    return out

And temperature and number_of_steps should be stored as subclasses of Input. More in the PR (to come)

[samwaseda]

This is my note for me to see what should come out. No need for anyone to understand

class Class1:
    var_2: float

class_1 = Class1(var_2=1.0)

flowchart TB;
    node_1 --> node_1.outputs.var_1;
    node_1.outputs.var_1 --> node_2.inputs.var_1;
    node_2.inputs.var_1 --> node_2;
    node_2.inputs.var_1 -- hasValue --> node_1.outputs.var_1.value;
    node_1.outputs.var_1 -- hasValue --> node_1.outputs.var_1.value;
    node_1 --> node_1.outputs.class_1;
    node_1.outputs.class_1 --> node_2.inputs.class_1;
    node_2.inputs.class_1--> node_2;
    node_2.inputs.class_1 -- hasValue --> node_1.outputs.class_1.value;
    node_1.outputs.class_1 -- hasValue --> node_1.outputs.class_1.value;
    node_1.outputs.class_1.var_2.value -- subClassOf --> node_1.outputs.class_1.value;
    node_2.inputs.class_1 -- hasValue --> node_1.outputs.class_1.var_2.value;
    node_1.outputs.class_1 -- hasValue --> node_1.outputs.class_1.var_2.value;

Loading

[samwaseda]

EDIT: With this PR I realized that what I had claimed was not wrong in the end, and you don't need dependence on semantikon_class in contrast to what's presented above (but you can also use it)