How do I pass arguments to methods wrapped in panel.param.ParamMethod

Panel
1

I’d like to make an element (or multiple elements) of a dashboard by wrapping a parametrized method with arguments in panel.param.ParamMethod, but can’t figure out how to pass the arguments.
This snippet might get across what I am up to:

plot_settings = {"line": {"x": "x", "y": ["y"]}, "hist": {"kind": "hist", "y": ["y"]}}
# really to be loaded from json

class Foo(param.Parameterized):
    mean_y = param.Number(default=0)
    _df = param.Parameter(
        default=pd.DataFrame({"x": np.arange(10), "y": np.arange(10)})
    )

    @param.depends("mean_y", watch=True)
    def make_df(self):
        x = np.arange(10)
        y = np.random.rand(*x.shape) - 0.5 + self.mean_y
        self._df = pd.DataFrame({"x": x, "y": y})

    @param.depends("_df")
    def make_hvplot(self, settings={}):
        return self._df.hvplot(**settings)


foo = Foo()

myplots = [pn.param.ParamMethod(foo.make_hvplot, method_args=plot_settings[k]) for k in plot_settings]

row = pn.Column(
    pn.Param(foo, parameters=["mean_y"]),
    **myplots,
)
row

except, of course, ParamMethod doesn’t have the method_args argument that I’m looking for.
Any suggestions?

2

Could you move your plot_settings into your class?

Then it’s calling hvplot(self.settings)?

3

If I did that, how would I make the different elements in the dashboard that should be made by calling the method with different settings values simultaneously?

4 
class Foo(param.Parameterized):
    mean_y = param.Number(default=0)
    settings = param.Dict()
    _df = param.Parameter(
        default=pd.DataFrame({"x": np.arange(10), "y": np.arange(10)})
    )

    @param.depends("mean_y", watch=True)
    def make_df(self):
        x = np.arange(10)
        y = np.random.rand(*x.shape) - 0.5 + self.mean_y
        self._df = pd.DataFrame({"x": x, "y": y})

    @param.depends("_df")
    def make_hvplot(self, settings={}):
        return self._df.hvplot(**settings)

foo1 = Foo(settings=plot_settings)
foo2 = Foo(settings=other_plot_settings)

5

Hmm, so then I’d need to make the method in foo2 depend on the _df parameter of foo1, because I need all plots to update when the data changes. This might be doable, but I’ll need to figure out how.

6

If it gets too complicated, perhaps this is not the right way forward. Instead, there might be a simpler approach like using hv.DynamicMap with a layout of your plots, but I don’t know the whole context.

Some examples:
https://pydeas.readthedocs.io/en/latest/holoviz_interactions/tips_and_tricks.html#Wrap-DynamicMap-around-Panel-methods-to-maintain-extents-and-improve-runtime

1 Like