param.List with item_type=param.XYCoordinate

Param
1

I would like to have a list of points (x,y) as a parameter on my class so that I can construct a numpy array from it of the proper share whenever it is modified. I have two questions

  1. This doesn’t appear to be the correct approach
class T(param.Parameterized):
    vertices = param.List(item_type=param.XYCoordinates)

    @param.depends("vertices", watch=True)
    def _update_points(self):
        self.coords = np.array(self.vertices).T

and allows me to append (2,2,2) to the list

  1. The @param.depends decorator never gets called when I append or extend the list but does when I replace it all together. Is there a way to trigger on adding a subitem to the list?
2

for list you will need to trigger manually the change of the parameter because appending an object to a list does not modify the list reference

t = T()
t.vertices.append((1,1))
t.param.trigger("vertices")
print(t.coords)
t.vertices.append((1,2))
t.param.trigger("vertices")
print(t.coords)
3

That seems to be the case for issue 2 after digging into it more.

Any thoughts on issue 1 where

item_type=param.XYCoordinates

seems to be ignored and allows (2,2,2)? I still can’t seem to find a reason for that.

4

same problem, param does not see the list modified when appending so th validation routine is not triggered.
However when we trigger the change the error is raised:

image
image693×122 4.38 KB

and appending (2,2) will throw the same error since it’s a tuple and not a param.XYCoordinate

1 Like
5

I don’t think you can use Parameters (from Param) to set item_type in your code. Instead you could just use tuple or create your own Point class outside of T and use it to set item_type.

1 Like