Step/Event values in runner about pycopancore HOT 9 CLOSED

Things that are not saved properly are:
location of cell, cell of individual and so on, so attributes that are not part of a process yet. Maybe this is solving itself if they become part of processes!

from pycopancore.

the question is now more: is this a bug? Do we need iterating through all variables? If they don't change, do we need their information? Or is it enough to get the information once, not every time a discontinuity is called?

from pycopancore.

Code proposal:

for (v, oc) in  self.ODE_variables,  self.explicit_variables,  self.step_variables,  self.event_variables:
        entities = oc.entities
        values = v.get_value_list(entities)
        for i in range(len(entities)):
            entity = entities[i]
            value = np.array([values[i]])
            try:
                trajectory_dict[v][entity] = np.concatenate((
                    trajectory_dict[v][entity], value))
            except KeyError:
                trajectory_dict[v][entity] = value
    t_np = np.array([t])
    trajectory_dict['t'] = np.concatenate((trajectory_dict['t'], t_np))

So the first line has changed from self.model.variables to the variables that change. Now, after finishing the while loop, do following:

for (v,oc) in self.model.variables and not in self.ODE_variables, self.explicit_variables, self.step_variables, self.event_variables:
    entities = oc.entities
                values = v.get_value_list(entities)
                for i in range(len(entities)):
                    entity = entities[i]
                    value = np.array([values[i]])
                    try:
                        trajectory_dict[v][entity] = np.concatenate((
                            trajectory_dict[v][entity], value))
                    except KeyError:
                        trajectory_dict[v][entity] = value
return trajectory_dict

Maybe there are some errors, but this is the idea

from pycopancore.

I agree with you.The code proposal also make sense since the performance will be better.

from pycopancore.

To make it easier to implement further processes in a later stage it might does make sense to reintroduce the process_variables in model.py.

from pycopancore.

Just a small proposal:
Instead of

for i in range(len(things)):
    thing = things[i]
    ...

you could use

for i, thing in enumerate(things):
    ...

from pycopancore.

Thank you! Just changed it.

from pycopancore.

While we are at that, two more comments (:
(1) maybe simply use zip
(2) Usually it's good to avoid try except staements in Python, following the Zen of Python's (PEP20) "Explicit is better than implicit."

maybe simple use:

        for entity, value in zip(entities, values):
            if entity in trajectory[v]:
                trajectory_dict[v][entity] = np.concatenate((
                    trajectory_dict[v][entity], value))
            else:
                trajectory_dict[v][entity] = np.array([value])

It shortens the whole thing and makes it directly readable.

from pycopancore.

Yes, I think it makes sense to change this, I put that suggestions into a separate issue #52 that only concerns code enhancement/improvement .
Since the main problem of this issue is already solved I closed it.

from pycopancore.