Module streamgen.visualizations
🖼️parameter visualization functions.
View Source
"""🖼️parameter visualization functions."""
from collections.abc import Callable
from copy import deepcopy
from itertools import islice
from typing import Any
import IPython
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from IPython.utils import io
from ipywidgets import widgets
from matplotlib import animation
from streamgen.parameter import Parameter
from streamgen.parameter.store import ParameterStore
from streamgen.samplers.tree import SamplingTree
def plot(values: list[int | float | list | np.ndarray], ax: plt.Axes | None = None, title: str | None = None) -> plt.Axes:
"""📈 plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Args:
values (list[int | float | list | np.ndarray]): list of values to plot.
ax (plt.Axes | None, optional): matplotlib Axes to plot to. Defaults to None.
title (str | None, optional): title of the plot. Defaults to None
Raises:
NotImplementedError: when the type of the parameter values are not yet supported by this function
Returns:
plt.Axes: parameter plot
"""
match values[0]:
case int() | float():
ax = sns.lineplot(values, ax=ax, marker="o")
case np.ndarray() | list():
if ax is None:
fig, ax = plt.subplots()
sns.set_theme()
indices = list(range(len(values)))
values = np.array(values).T
assert len(values.shape) == 2, "only arrays with two dimensions can be visualized here" # noqa: S101, PLR2004
ax.stackplot(indices, values)
case _:
raise NotImplementedError
ax.set_title(title)
return ax
def plot_parameter(param: Parameter, num_values: int | None = None, ax: plt.Axes | None = None) -> plt.Axes:
"""⚙️ plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Args:
param (Parameter): parameter to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
ax (plt.Axes | None, optional): matplotlib Axes to plot to. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported by this function
Returns:
plt.Axes: parameter plot
"""
match num_values:
case int():
assert num_values > 1, "at least two value are needed for the plot." # noqa: S101
values = [param.value, *list(islice(deepcopy(param.schedule), num_values - 1))]
case None:
values = [param.value, *list(islice(deepcopy(param.schedule), None))]
return plot(values, ax, title=param.name)
def plot_parameter_store(store: ParameterStore, num_values: int | None = None) -> mpl.figure.Figure:
"""🗄️ plots every parameter in a `ParameterStore` in one figure.
Args:
store (ParameterStore): parameter store to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported
Returns:
mpl.figure.Figure: matplotlib figure object
"""
num_columns: int = len(store.parameter_names)
sns.set_theme()
fig = plt.figure()
for idx, param in enumerate(store.parameter_names):
ax = fig.add_subplot(num_columns, 1, idx + 1)
plot_parameter(store[param], num_values=num_values, ax=ax)
fig.set_figheight(num_columns * 3.0)
plt.tight_layout()
return fig
def plot_parameter_store_widget(store: ParameterStore, num_values: int | None = None) -> widgets.Tab:
"""📂 plots every parameter of each scope in a `ParameterStore` in a separate `ipywidgets.widgets.Tab`.
Args:
store (ParameterStore): parameter store to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported
Returns:
widgets.Tab: ipywidgets tab widget
"""
scopes = list(store.scopes)
tabs = [widgets.Output() for _ in scopes]
widget = widgets.Tab(children=tabs)
for idx, scope in enumerate(scopes):
widget.set_title(idx, scope)
with tabs[idx]:
params = store.get_scope(scope)
fig = plot_parameter_store(params, num_values=num_values)
plt.show(fig)
return widget
def plot_labeled_samples_grid(
tree: SamplingTree,
plotting_func: Callable[[Any, plt.Axes], plt.Axes],
columns: int = 4,
) -> mpl.figure.Figure:
"""📟 plots a `columns`x`columns` grid of labeled samples generated from a `SamplingTree` with `ClassLabelNode`s.
Args:
tree (SamplingTree): tree to generate samples from
plotting_func (Callable[[Any, plt.Axes], plt.Axes]): function to visualize a single sample.
The function should take a sample and a `plt.Axes` as arguments.
columns (int, optional): number of samples in the columns (and rows). Defaults to 4.
Returns:
mpl.figure.Figure: matplotlib figure object
"""
num_samples = columns * columns
with io.capture_output() as _captured:
samples = tree.collect(num_samples)
sns.set_theme()
fig = plt.figure()
for idx, (sample, target) in enumerate(samples):
ax = fig.add_subplot(columns, columns, idx + 1)
plotting_func(sample, ax)
ax.set_title(target)
fig.set_figheight(columns * 3.0)
fig.set_figwidth(columns * 3.0)
plt.tight_layout()
return fig
def plot_labeled_samples_animation(
tree: SamplingTree,
plotting_func: Callable[[Any, plt.Axes], plt.Axes],
num_samples: int = 8,
interval: int = 200,
) -> IPython.display.HTML:
"""🎞️ plots several labeled samples generated from a `SamplingTree` with `ClassLabelNode`s as an animation.
Args:
tree (SamplingTree): tree to generate samples from
plotting_func (Callable[[Any, plt.Axes], plt.Axes]): function to visualize a single sample.
The function should take a sample and a `plt.Axes` as arguments.
num_samples (int, optional): number of samples to include in the animation. Defaults to 8.
interval (int, optional): delay between frames in milliseconds. Defaults to 200.
Returns:
IPython.display.HTML: animation player as an ipython HTML output
"""
with io.capture_output() as _captured:
samples = tree.collect(num_samples)
sns.set_theme()
fig, ax = plt.subplots()
def _plotting_func(idx: int, ax: plt.Axes) -> None:
sample, target = samples[idx]
ax.clear()
plotting_func(sample, ax)
ax.set_title(target)
anim = animation.FuncAnimation(fig, _plotting_func, frames=num_samples, fargs=(ax,), interval=interval)
return IPython.display.HTML(anim.to_jshtml())
Functions
plot
def plot(
values: list[int | float | list | numpy.ndarray],
ax: matplotlib.axes._axes.Axes | None = None,
title: str | None = None
) -> matplotlib.axes._axes.Axes
📈 plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| values | list[int | float | list |
| ax | plt.Axes | None | matplotlib Axes to plot to. Defaults to None. |
| title | str | None | title of the plot. Defaults to None |
Returns:
| Type | Description |
|---|---|
| plt.Axes | parameter plot |
Raises:
| Type | Description |
|---|---|
| NotImplementedError | when the type of the parameter values are not yet supported by this function |
View Source
def plot(values: list[int | float | list | np.ndarray], ax: plt.Axes | None = None, title: str | None = None) -> plt.Axes:
"""📈 plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Args:
values (list[int | float | list | np.ndarray]): list of values to plot.
ax (plt.Axes | None, optional): matplotlib Axes to plot to. Defaults to None.
title (str | None, optional): title of the plot. Defaults to None
Raises:
NotImplementedError: when the type of the parameter values are not yet supported by this function
Returns:
plt.Axes: parameter plot
"""
match values[0]:
case int() | float():
ax = sns.lineplot(values, ax=ax, marker="o")
case np.ndarray() | list():
if ax is None:
fig, ax = plt.subplots()
sns.set_theme()
indices = list(range(len(values)))
values = np.array(values).T
assert len(values.shape) == 2, "only arrays with two dimensions can be visualized here" # noqa: S101, PLR2004
ax.stackplot(indices, values)
case _:
raise NotImplementedError
ax.set_title(title)
return ax
plot_labeled_samples_animation
def plot_labeled_samples_animation(
tree: streamgen.samplers.tree.SamplingTree,
plotting_func: collections.abc.Callable[[typing.Any, matplotlib.axes._axes.Axes], matplotlib.axes._axes.Axes],
num_samples: int = 8,
interval: int = 200
) -> IPython.core.display.HTML
🎞️ plots several labeled samples generated from a SamplingTree with ClassLabelNodes as an animation.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| tree | SamplingTree | tree to generate samples from | None |
| plotting_func | Callable[[Any, plt.Axes], plt.Axes] | function to visualize a single sample. The function should take a sample and a plt.Axes as arguments. |
None |
| num_samples | int | number of samples to include in the animation. Defaults to 8. | 8 |
| interval | int | delay between frames in milliseconds. Defaults to 200. | 200 |
Returns:
| Type | Description |
|---|---|
| IPython.display.HTML | animation player as an ipython HTML output |
View Source
def plot_labeled_samples_animation(
tree: SamplingTree,
plotting_func: Callable[[Any, plt.Axes], plt.Axes],
num_samples: int = 8,
interval: int = 200,
) -> IPython.display.HTML:
"""🎞️ plots several labeled samples generated from a `SamplingTree` with `ClassLabelNode`s as an animation.
Args:
tree (SamplingTree): tree to generate samples from
plotting_func (Callable[[Any, plt.Axes], plt.Axes]): function to visualize a single sample.
The function should take a sample and a `plt.Axes` as arguments.
num_samples (int, optional): number of samples to include in the animation. Defaults to 8.
interval (int, optional): delay between frames in milliseconds. Defaults to 200.
Returns:
IPython.display.HTML: animation player as an ipython HTML output
"""
with io.capture_output() as _captured:
samples = tree.collect(num_samples)
sns.set_theme()
fig, ax = plt.subplots()
def _plotting_func(idx: int, ax: plt.Axes) -> None:
sample, target = samples[idx]
ax.clear()
plotting_func(sample, ax)
ax.set_title(target)
anim = animation.FuncAnimation(fig, _plotting_func, frames=num_samples, fargs=(ax,), interval=interval)
return IPython.display.HTML(anim.to_jshtml())
plot_labeled_samples_grid
def plot_labeled_samples_grid(
tree: streamgen.samplers.tree.SamplingTree,
plotting_func: collections.abc.Callable[[typing.Any, matplotlib.axes._axes.Axes], matplotlib.axes._axes.Axes],
columns: int = 4
) -> matplotlib.figure.Figure
📟 plots a columnsxcolumns grid of labeled samples generated from a SamplingTree with ClassLabelNodes.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| tree | SamplingTree | tree to generate samples from | None |
| plotting_func | Callable[[Any, plt.Axes], plt.Axes] | function to visualize a single sample. The function should take a sample and a plt.Axes as arguments. |
None |
| columns | int | number of samples in the columns (and rows). Defaults to 4. | 4 |
Returns:
| Type | Description |
|---|---|
| mpl.figure.Figure | matplotlib figure object |
View Source
def plot_labeled_samples_grid(
tree: SamplingTree,
plotting_func: Callable[[Any, plt.Axes], plt.Axes],
columns: int = 4,
) -> mpl.figure.Figure:
"""📟 plots a `columns`x`columns` grid of labeled samples generated from a `SamplingTree` with `ClassLabelNode`s.
Args:
tree (SamplingTree): tree to generate samples from
plotting_func (Callable[[Any, plt.Axes], plt.Axes]): function to visualize a single sample.
The function should take a sample and a `plt.Axes` as arguments.
columns (int, optional): number of samples in the columns (and rows). Defaults to 4.
Returns:
mpl.figure.Figure: matplotlib figure object
"""
num_samples = columns * columns
with io.capture_output() as _captured:
samples = tree.collect(num_samples)
sns.set_theme()
fig = plt.figure()
for idx, (sample, target) in enumerate(samples):
ax = fig.add_subplot(columns, columns, idx + 1)
plotting_func(sample, ax)
ax.set_title(target)
fig.set_figheight(columns * 3.0)
fig.set_figwidth(columns * 3.0)
plt.tight_layout()
return fig
plot_parameter
def plot_parameter(
param: streamgen.parameter.Parameter,
num_values: int | None = None,
ax: matplotlib.axes._axes.Axes | None = None
) -> matplotlib.axes._axes.Axes
⚙️ plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| param | Parameter | parameter to plot | None |
| num_values | int | None | number of values to plot. If None, collect all values from the schedule. Defaults to None. |
| ax | plt.Axes | None | matplotlib Axes to plot to. Defaults to None. |
Returns:
| Type | Description |
|---|---|
| plt.Axes | parameter plot |
Raises:
| Type | Description |
|---|---|
| NotImplementedError | when the type of the parameter values are not yet supported by this function |
View Source
def plot_parameter(param: Parameter, num_values: int | None = None, ax: plt.Axes | None = None) -> plt.Axes:
"""⚙️ plots the scheduled values of a parameter.
This function currently supports plotting numeric parameters and probabilities.
Args:
param (Parameter): parameter to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
ax (plt.Axes | None, optional): matplotlib Axes to plot to. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported by this function
Returns:
plt.Axes: parameter plot
"""
match num_values:
case int():
assert num_values > 1, "at least two value are needed for the plot." # noqa: S101
values = [param.value, *list(islice(deepcopy(param.schedule), num_values - 1))]
case None:
values = [param.value, *list(islice(deepcopy(param.schedule), None))]
return plot(values, ax, title=param.name)
plot_parameter_store
def plot_parameter_store(
store: streamgen.parameter.store.ParameterStore,
num_values: int | None = None
) -> matplotlib.figure.Figure
🗄️ plots every parameter in a ParameterStore in one figure.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| store | ParameterStore | parameter store to plot | None |
| num_values | int | None | number of values to plot. If None, collect all values from the schedule. Defaults to None. |
Returns:
| Type | Description |
|---|---|
| mpl.figure.Figure | matplotlib figure object |
Raises:
| Type | Description |
|---|---|
| NotImplementedError | when the type of the parameter values are not yet supported |
View Source
def plot_parameter_store(store: ParameterStore, num_values: int | None = None) -> mpl.figure.Figure:
"""🗄️ plots every parameter in a `ParameterStore` in one figure.
Args:
store (ParameterStore): parameter store to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported
Returns:
mpl.figure.Figure: matplotlib figure object
"""
num_columns: int = len(store.parameter_names)
sns.set_theme()
fig = plt.figure()
for idx, param in enumerate(store.parameter_names):
ax = fig.add_subplot(num_columns, 1, idx + 1)
plot_parameter(store[param], num_values=num_values, ax=ax)
fig.set_figheight(num_columns * 3.0)
plt.tight_layout()
return fig
plot_parameter_store_widget
def plot_parameter_store_widget(
store: streamgen.parameter.store.ParameterStore,
num_values: int | None = None
) -> ipywidgets.widgets.widget_selectioncontainer.Tab
📂 plots every parameter of each scope in a ParameterStore in a separate ipywidgets.widgets.Tab.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
| store | ParameterStore | parameter store to plot | None |
| num_values | int | None | number of values to plot. If None, collect all values from the schedule. Defaults to None. |
Returns:
| Type | Description |
|---|---|
| widgets.Tab | ipywidgets tab widget |
Raises:
| Type | Description |
|---|---|
| NotImplementedError | when the type of the parameter values are not yet supported |
View Source
def plot_parameter_store_widget(store: ParameterStore, num_values: int | None = None) -> widgets.Tab:
"""📂 plots every parameter of each scope in a `ParameterStore` in a separate `ipywidgets.widgets.Tab`.
Args:
store (ParameterStore): parameter store to plot
num_values (int | None, optional): number of values to plot.
If None, collect all values from the schedule. Defaults to None.
Raises:
NotImplementedError: when the type of the parameter values are not yet supported
Returns:
widgets.Tab: ipywidgets tab widget
"""
scopes = list(store.scopes)
tabs = [widgets.Output() for _ in scopes]
widget = widgets.Tab(children=tabs)
for idx, scope in enumerate(scopes):
widget.set_title(idx, scope)
with tabs[idx]:
params = store.get_scope(scope)
fig = plot_parameter_store(params, num_values=num_values)
plt.show(fig)
return widget