Plotting

  1. ProgressPlotter

  2. SelectionPlotter

Plotting is done by Plotter objects. Plotters are calculators which produce graphs. They generally do this by decorating other calculators to hook into them, collect data from them and then plot the results. To see examples of how to use plotters work, look at the documentation of the individual plotters. For example SelectionPlotter and ProgressPlotter.

Making New Plotters

New plotters should inherit Plotter and define a plot() method. Note that plotters do no necessarily have to make a plot when plot() is made. For example, SelectionPlotter plots every time the Selector it is used with becomes exhausted and it’s plot() does nothing.

class Plotter

Bases: object

A base class for plotters.

Methods

plot(self, progress)

Plots a graph.
__init__(self, /, *args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

plot(self, progress)

Plots a graph.

Parameters

progress (Population) – A Population where each generation of the GA is a subpopulation.

Returns

None

Return type

NoneType

class ProgressPlotter(filename, property_fn, y_label, progress_fn=None, filter=<function ProgressPlotter.<lambda>>)

Bases: stk.calculators.ea.plotters.Plotter

Plots how a property changes during a GA run.

The produced plot will show the GA generations on the x axis and the min, mean and max values of an attribute on the y axis.

Examples

Plot how fitness value changes with GA generations

import stk

# progress has subpopulations where each subpopulation is a
# generation of a GA.
progress = stk.Population(...)

# Make the plotter which plots the fitness change across
# generations.
plotter = stk.ProgressPlotter(
    filename='fitness_plot',
    property_fn=lambda mol: mol.fitness,
    y_label='Fitness'
)
plotter.plot(progress)

Plot how the number of atoms changes with GA generations

plotter = ProgressPlotter(
    filename='atom_plot',
    property_fn=lambda mol: len(mol.atoms),
    y_label='Number of Atoms'
)
plotter.plot(progress)

Methods

plot(self, progress)

Plot a progress plot.
__init__(self, filename, property_fn, y_label, progress_fn=None, filter=<function ProgressPlotter.<lambda> at 0x7fcf9a683b00>)

Initialize a ProgressPlotter instance.

Parameters

  • filename (str) – The basename of the files. This means it should not include file extensions.

  • property_fn (callable) – A callable which takes a EAPopulation and a Molecule object and returns a property value of that molecule, which is used for the plot. The callable must return a valid value for each Molecule in the population.

  • y_label (str) – The y label for the produced graph.

  • progress_fn (callable, optional) – Takes the population passed to plot() and excutes a computation on it. This may be useful if you want to apply a normalization to the fitness values in the progress population, for example.

  • filter (callable, optional) – Takes an EAPopulation and a Molecule as input and returns True or False. Only molecules which return True will be plotted. Default is for all molecules to be plotted.

plot(self, progress)

Plot a progress plot.

Parameters

progress (Population) – A Population where each generation of the GA is a subpopulation.

Returns

None

Return type

NoneType

class SelectionPlotter(filename, selector, x_label='Molecule: name - fitness value', molecule_label=<function SelectionPlotter.<lambda>>, heat_map_value=<function SelectionPlotter.<lambda>>, heat_map_label='Fitness', order_by=<function SelectionPlotter.<lambda>>)

Bases: stk.calculators.ea.plotters.Plotter

Plots which molecules a Selector selects.

Examples

import stk

# Make a population of molecules.
pop = stk.Population(...)

# Make a selector.
roulette = stk.Roulette(num=10)

# Make a plotter.
stk.SelectionPlotter('roulette_counter', roulette)

# Select the molecules.
selected = list(roulette.select(pop))

# There should now be a file called "roulette_counter_1.png"
# which shows a graph of all the selected molecules.

# Make another population.
pop2 = stk.Population(...)

# Select molecules from this other population.
selected2 = list(roulette.select(pop2))

# There should now be a file called "roulette_counter_2.png"
# which shows a graph of all the selected molecules.

# Select from the original population again.
selected3 = list(roulette.select(pop))

# There should now be a file called "roulette_counter_3.png"
# which shows a graph of all the selected molecules.

# And so on every time you use "roulette.select()".

Methods

plot(self, progress)

Does nothing.
__init__(self, filename, selector, x_label='Molecule: name - fitness value', molecule_label=<function SelectionPlotter.<lambda> at 0x7fcf96cb7830>, heat_map_value=<function SelectionPlotter.<lambda> at 0x7fcf96cb78c0>, heat_map_label='Fitness', order_by=<function SelectionPlotter.<lambda> at 0x7fcf96cb7950>)

Initialize a SelectionPlotter instance.

Parameters

  • filename (str) – The basename of the files. This means it should not include file extensions.

  • selector (Selector) – The Selector whose selection of molecules is plotted.

  • x_label (str, optional) – The label use for the x axis.

  • molecule_label (callable, optional) – A callable which takes a EAPopulation and a Molecule, for each molecule which is to be included on the x-axis of the counter plot. It should return a string, which is the label used for the Molecule on the plot.

  • heat_map_value (callable, optional) – A callable, which takes a EAPopulation and a Molecule, for each molecule which is to be included on the x-axis, and returns a value. The value is used for coloring the heat map used in the plot.

  • heat_map_label (str, optional) – The label used for the heat map key.

  • order_by (callable, optional) – A callable, which takes a EAPopulation and a Molecule, for each molecule which is to be included on the x-axis, and returns a value. The value is used to sort the plotted molecules along the x-axis in descending order.

plot(self, progress)

Does nothing.

This Plotter creates a plot each time selector is finished selecting molecules.

Parameters

progress (Population) – A Population where each generation of the GA is a subpopulation.

Returns

None

Return type

NoneType