Hi Holoviz Community.
I would like to post my app to the community in exchange for requesting code reviews and some help with troubleshooting 
You may also find the content interesting.
The work is inspired by this paper:
The app so far looks like this:
import hvplot.pandas
import numbergen as ng
import numpy as np
import pandas as pd
import panel as pn
import param as pm
class Society(pm.Parameterized):
N = pm.Integer(default=30, bounds=(1, None))
C = pm.Integer(default=20, bounds=(1, None))
P = pm.Integer(default=10, bounds=(1, None))
society = pm.List(precedence=1)
community = pm.List(precedence=1)
public_goods = pm.List(precedence=1)
sample = pm.Action(lambda self: self._sample())
def __init__(self, **params):
super().__init__(**params)
self._sample()
@pm.depends('N', 'C', 'P', watch=True)
def _sample(self):
self.society = list(range(self.N))
self.community = list(
np.random.choice(a=list(self.society), size=self.C, replace=False, p=None)
)
self.public_goods = list(
enumerate(
np.random.choice(
a=list(self.community),
size=self.P,
replace=True,
p=None,
)
)
)
def text_view(self):
text = f"""
Number of Citizens in the Society: **{self.N}**\n
Community Size: **{self.C}**\n
Number of Public Goods: **{self.P}**\n
Society: **{self.society}**\n
Community: **{self.community}**\n
Public Goods: **{self.public_goods}**"""
return pn.pane.Markdown(text)
def view(self):
"""š¶ļø"""
return pn.Row(self, self.text_view)
society = Society()
class ConcaveFunctionGenerator(pm.Parameterized):
f0 = pm.Number(default=0.2, bounds=(0, 1), doc='Value of f(0)')
f1 = pm.Number(default=0.8, bounds=(0, 1), softbounds=(0, 1), doc='Value of f(1)')
steepness = pm.Number(default=5, bounds=(1, 20), doc='Steepness of the curve')
def __init__(self, **params):
super().__init__(**params)
self._update_f1_bounds()
@pm.depends('f0', watch=True)
def _update_f1_bounds(self):
# Clip the value of f1 if it's below f0
self.f1 = max(self.f0, self.f1)
# Update the lower bound of f1 to be the value of f0
self.param['f1'].bounds = (self.f0, 1)
def x(self):
return np.linspace(0, 1, 400)
def f(self, x):
# Using the negative exponential function as a base
y = 1 - np.exp(-self.steepness * x)
# Adjusting the function to start at f0 and end at f1
y = self.f0 + (self.f1 - self.f0) * (y - y.min()) / (y.max() - y.min())
return y
def chart_view(self):
x = self.x()
y = self.f(x)
df = pd.DataFrame({'x': x, 'y': y})
return df.hvplot.line(x='x', y='y', ylim=(0, 1.01))
def view(self):
return pn.Row(self, self.chart_view)
concave_gen = ConcaveFunctionGenerator()
app = pn.Column(society.view(), concave_gen.view())
class PopulationFunctionGenerator(pm.Parameterized):
f0_mu = pm.Number(default=0.1, bounds=(0, 1), doc='Mean Value of f(0)')
f0_sigma = pm.Number(default=0.3, bounds=(0, 1), doc='Standard Deviationn of f(0)')
f1_mu = pm.Number(default=0.5, bounds=(0, 1), doc='Mean Value of f(1)')
f1_sigma = pm.Number(default=0.4, bounds=(0, 1), doc='Standard Deviationn of f(1)')
steepness_lbound = pm.Integer(
default=1, bounds=(1, 20), doc='Lower Bound of Steepness'
)
steepness_rbound = pm.Integer(
default=20, bounds=(1, 20), doc='Upper Bound of Steepness'
)
n = pm.Integer(constant=True)
society = pm.Selector(default=society, objects=[society], instantiate=True)
value_function_generator = pm.Selector(
default=ConcaveFunctionGenerator,
objects=[ConcaveFunctionGenerator],
instantiate=True,
)
value_functions = pm.List()
sample = pm.Action(lambda self: self._sample(), instantiate=True)
i = pm.Integer(0)
def __init__(self, **params):
super().__init__(**params)
self._update()
self._sample()
@pm.depends('society.public_goods', 'society.society', watch=True)
def _update(self):
with pm.edit_constant(self):
self.n = len(self.society.public_goods) * len(self.society.society)
self.param['i'].bounds = (0, self.n - 1)
@pm.depends('steepness_lbound', watch=True)
def _update_steepness_rbound_bounds(self):
# Clip the value of f1 if it's below f0
self.steepness_rbound = max(self.steepness_lbound, self.steepness_rbound)
# Update the lower bound of f1 to be the value of f0
self.param['steepness_rbound'].bounds = (self.steepness_lbound, 20)
# fix this so the parameters are from this class
def concave_function_parameters_generator(self):
return dict(
f0=ng.BoundedNumber(
generator=ng.NormalRandom(mu=self.f0_mu, sigma=self.f0_sigma),
bounds=(0, 1),
)(),
f1=ng.BoundedNumber(
generator=ng.NormalRandom(mu=self.f1_mu, sigma=self.f1_sigma),
bounds=(0, 1),
)(),
steepness=ng.UniformRandom(
lbound=self.steepness_lbound, ubound=self.steepness_rbound
)(),
)
@pm.depends(
'steepness_lbound',
'steepness_rbound',
'f0_mu',
'f0_sigma',
'f1_mu',
'f1_sigma',
watch=True,
)
def _sample(self):
value_functions = [
self.value_function_generator(
**self.concave_function_parameters_generator()
)
for p_i in range(self.n)
]
self.value_functions = value_functions
# @pm.depends('i')
def samples_view(self):
return self.value_functions[self.i].chart_view()
@pm.depends('i', 'steepness_lbound')
def view(self):
return pn.Row(self, self.samples_view())
population_function_generator = PopulationFunctionGenerator(
society=society,
value_function_generator=ConcaveFunctionGenerator,
)
app = population_function_generator.view()
if __name__ == '__main__':
print(society)
print(concave_gen)
print(population_function_generator)
print('here')
Iām aiming to work my way through the paper in the form of a panel app. The following is an approach to sampling concave, smooth, increasing functions.
The app is deployed with:
panel serve app/app.py --autoreload
Problem #1: the above is not re-drawing at
population_function_generator.view()
ie. the chart is not redrawing when I slide i at the bottom.
Problem #2: The toggle buttons are not working for the Selector Drop Downs. The expand functionality.
I read through the following resource:
https://panel.holoviz.org/how_to/param/subobjects.html
Which I posted a minimum example here: Panel Selector Expand Not Working After Restarting Server - #3 by LinuxIsCool
In general, you can follow the developments here:
Holoviz community rocks 