"""Auxiliary functions for plotting the results of onion-clustering.
* Author: Becchi Matteo <bechmath@gmail.com>
* Date: November 28, 2024
"""
import os
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
import plotly.graph_objects as go
from matplotlib.colors import rgb2hex
from matplotlib.patches import Ellipse
from matplotlib.ticker import MaxNLocator
from numpy.typing import NDArray
from tropea_clustering._internal.onion_smooth.functions import gaussian
from tropea_clustering._internal.onion_smooth.main import StateUni
from tropea_clustering._internal.onion_smooth.main_2d import StateMulti
COLORMAP = "viridis"
[docs]
def plot_output_uni(
title: Path,
input_data: NDArray[np.float64],
state_list: list[StateUni],
):
"""Plots clustering output with Gaussians and thresholds.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
input_data : ndarray of shape (n_particles, n_frames)
The input data array.
state_list : list[StateUni]
The list of the cluster states.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig1.png
:alt: Example Image
:width: 600px
The left planel shows the input time-series data, with the backgound
colored according to the thresholds between the clusters. The left panel
shows the cumulative data distribution, and the Gaussians fitted to the
data, corresponding to the identified clusters.
"""
n_particles, t_steps = input_data.shape
flat_m = input_data.flatten()
counts, bins = np.histogram(flat_m, bins=100, density=True)
bins -= (bins[1] - bins[0]) / 2
counts *= flat_m.size
fig, axes = plt.subplots(
1,
2,
sharey=True,
gridspec_kw={"width_ratios": [3, 1]},
figsize=(9, 4.8),
)
axes[1].stairs(
counts, bins, fill=True, orientation="horizontal", alpha=0.5
)
palette = []
n_states = len(state_list)
cmap = plt.get_cmap(COLORMAP, n_states + 1)
for i in range(1, cmap.N):
rgba = cmap(i)
palette.append(rgb2hex(rgba))
time = np.linspace(0, t_steps - 1, t_steps)
step = 1
if input_data.size > 1e6:
step = 10
for mol in input_data[::step]:
axes[0].plot(
time,
mol,
c="xkcd:black",
ms=0.1,
lw=0.1,
alpha=0.5,
rasterized=True,
)
for state_id, state in enumerate(state_list):
attr = state.get_attributes()
popt = [attr["mean"], attr["sigma"], attr["area"]]
axes[1].plot(
gaussian(np.linspace(bins[0], bins[-1], 1000), *popt),
np.linspace(bins[0], bins[-1], 1000),
color=palette[state_id],
)
style_color_map = {
0: ("--", "xkcd:black"),
1: ("--", "xkcd:blue"),
2: ("--", "xkcd:red"),
}
time2 = np.linspace(
time[0] - 0.05 * (time[-1] - time[0]),
time[-1] + 0.05 * (time[-1] - time[0]),
100,
)
for state_id, state in enumerate(state_list):
th_inf = state.get_attributes()["th_inf"]
th_sup = state.get_attributes()["th_sup"]
linestyle, color = style_color_map.get(th_inf[1], ("-", "xkcd:black"))
axes[1].hlines(
th_inf[0],
xmin=0.0,
xmax=np.amax(counts),
linestyle=linestyle,
color=color,
)
axes[0].fill_between(
time2,
th_inf[0],
th_sup[0],
color=palette[state_id],
alpha=0.25,
)
axes[1].hlines(
state_list[-1].get_attributes()["th_sup"][0],
xmin=0.0,
xmax=np.amax(counts),
linestyle=linestyle,
color="black",
)
# Set plot titles and axis labels
axes[0].set_ylabel("Signal")
axes[0].set_xlabel(r"Time [frame]")
axes[1].set_xticklabels([])
fig.savefig(title, dpi=600)
[docs]
def plot_one_trj_uni(
title: Path,
example_id: int,
input_data: NDArray[np.float64],
labels: NDArray[np.int64],
):
"""Plots the colored trajectory of one example particle.
Unclassified data points are colored with the darkest color.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
example_id : int
The ID of the selected particle.
input_data : ndarray of shape (n_particles, n_frames)
The input data array.
labels : ndarray of shape (n_particles, n_frames)
The output of Onion Clustering.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig2.png
:alt: Example Image
:width: 600px
The datapoints are colored according to the cluster they have been
assigned.
"""
n_particles, n_frames = input_data.shape
time = np.linspace(0, n_frames - 1, n_frames)
fig, axes = plt.subplots()
unique_labels = np.unique(labels)
# If there are no assigned window, we still need the "-1" state
# for consistency:
if -1 not in unique_labels:
unique_labels = np.insert(unique_labels, 0, -1)
cmap = plt.get_cmap(COLORMAP, unique_labels.size)
color = labels[example_id] + 1
axes.plot(time, input_data[example_id], c="black", lw=0.1)
axes.scatter(
time,
input_data[example_id],
c=color,
cmap=cmap,
vmin=0,
vmax=unique_labels.size - 1,
s=1.0,
)
# Add title and labels to the axes
fig.suptitle(f"Example particle: ID = {example_id}")
axes.set_xlabel("Time [frame]")
axes.set_ylabel("Signal")
fig.savefig(title, dpi=600)
[docs]
def plot_state_populations(
title: Path,
labels: NDArray[np.int64],
):
"""
Plot the populations of clusters over time.
For each trajectory frame, plots the fraction of the population of each
cluster. In the legend, "ENV0" refers to the unclassified data.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
labels : ndarray of shape (n_particles, n_frames)
The output of Onion Clustering.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig4.png
:alt: Example Image
:width: 600px
"""
n_particles, n_frames = labels.shape
unique_labels = np.unique(labels)
if -1 not in unique_labels:
unique_labels = np.insert(unique_labels, 0, -1)
list_of_populations = []
for label in unique_labels:
population = np.sum(labels == label, axis=0)
list_of_populations.append(population / n_particles)
palette = []
cmap = plt.get_cmap(COLORMAP, unique_labels.size)
for i in range(cmap.N):
rgba = cmap(i)
palette.append(rgb2hex(rgba))
fig, axes = plt.subplots()
time = range(n_frames)
for label, pop in enumerate(list_of_populations):
axes.plot(time, pop, label=f"ENV{label}", color=palette[label])
axes.set_xlabel(r"Time [frame]")
axes.set_ylabel(r"Population fraction")
axes.legend()
fig.savefig(title, dpi=600)
[docs]
def plot_sankey(
title: Path,
labels: NDArray[np.int64],
tmp_frame_list: list[int] | NDArray[np.int64],
):
"""
Plots the Sankey diagram at the desired frames.
.. warning::
This function requires the python package Kaleido, and uses plotly
instead of matplotlib.pyplot. For this reason is deprecated and not
supported since tropea-clustering 2.0.0.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
labels : ndarray of shape (n_particles, n_frames)
The output of the clustering algorithm.
tmp_frame_list : list[int] | NDArray[np.int64]
The list of frames at which we want to plot the Sankey.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig5.png
:alt: Example Image
:width: 600px
For each of the selected frames, the colored bars width is proportional
to each cluster population. The gray bands' witdh are proportional to
the number of data points moving from one cluster to the other between the
selected frames. State "-1" refers to the unclassified data.
"""
frame_list = np.array(tmp_frame_list)
unique_labels = np.unique(labels)
if -1 not in unique_labels:
unique_labels = np.insert(unique_labels, 0, -1)
n_states = unique_labels.size
source = np.empty((frame_list.size - 1) * n_states**2)
target = np.empty((frame_list.size - 1) * n_states**2)
value = np.empty((frame_list.size - 1) * n_states**2)
count = 0
tmp_label = []
# Loop through the frame_list and calculate the transition matrix
# for each time window.
for i, t_0 in enumerate(frame_list[:-1]):
# Calculate the time jump for the current time window.
t_jump = frame_list[i + 1] - frame_list[i]
trans_mat = np.zeros((n_states, n_states))
# Iterate through the current time window and increment
# the transition counts in trans_mat
for label in labels:
trans_mat[label[t_0] + 1][label[t_0 + t_jump] + 1] += 1
# Store the source, target, and value for the Sankey diagram
# based on trans_mat
for j, row in enumerate(trans_mat):
for k, elem in enumerate(row):
source[count] = j + i * n_states
target[count] = k + (i + 1) * n_states
value[count] = elem
count += 1
# Create node labels
for j in unique_labels:
tmp_label.append(f"State {j}")
state_label = np.array(tmp_label).flatten()
# Generate a color palette for the Sankey diagram.
palette = []
cmap = plt.get_cmap(COLORMAP, n_states)
for i in range(cmap.N):
rgba = cmap(i)
palette.append(rgb2hex(rgba))
# Tile the color palette to match the number of frames.
color = np.tile(palette, frame_list.size)
# Create dictionaries to define the Sankey diagram nodes and links.
node = {"label": state_label, "pad": 30, "thickness": 20, "color": color}
link = {"source": source, "target": target, "value": value}
# Create the Sankey diagram using Plotly.
sankey_data = go.Sankey(link=link, node=node, arrangement="perpendicular")
fig = go.Figure(sankey_data)
# Add the title with the time information.
fig.update_layout(title=f"Frames: {frame_list}")
fig.write_image(title, scale=5.0)
[docs]
def plot_time_res_analysis(
title: Path,
tra: NDArray[np.float64],
):
"""
Plots the results of clustering at different time resolutions.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
tra : ndarray of shape (delta_t_values, 3)
tra[j][0] must contain the j-th value used as delta_t;
tra[j][1] must contain the corresponding number of states;
tra[j][2] must contain the corresponding unclassified fraction.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig6.png
:alt: Example Image
:width: 600px
For each of the analyzed time resolutions, the blue curve shows the number
of identified clusters (not including the unclassified data); the orange
line shows the fraction of unclassififed data.
"""
fig, ax = plt.subplots()
ax.plot(tra[:, 0], tra[:, 1], marker="o")
ax.set_xlabel(r"Time resolution $\Delta t$ [frame]")
ax.set_ylabel(r"# environments", weight="bold", c="#1f77b4")
ax.set_xscale("log")
ax.set_ylim(-0.2, np.max(tra[:, 1]) + 0.2)
ax.yaxis.set_major_locator(MaxNLocator(integer=True))
ax_r = ax.twinx()
ax_r.plot(tra[:, 0], tra[:, 2], marker="o", c="#ff7f0e")
ax_r.set_ylabel("Unclassified fraction", weight="bold", c="#ff7f0e")
ax_r.set_ylim(-0.02, 1.02)
fig.savefig(title, dpi=600)
[docs]
def plot_pop_fractions(
title: Path,
list_of_pop: list[list[float]],
tra: NDArray[np.float64],
):
"""
Plot, for every time resolution, the populations of the clusters.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
list_of_pop : list[list[float]]
For every delta_t, this is the list of the populations of all the
states (the first one is the unclassified data points).
tra : ndarray of shape (delta_t_values, 3)
tra[j][0] must contain the j-th value used as delta_t;
tra[j][1] must contain the corresponding number of states;
tra[j][2] must contain the corresponding unclassified fraction.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/uni_Fig7.png
:alt: Example Image
:width: 600px
For each time resolution analysed, the bars show the fraction of data
points classified in each cluster. Clusters are ordered according to the
value of their Gaussian's mean; the bottom cluster is always the
unclassified data points.
"""
# Pad the lists in list_of_pop to ensure they all have the same length
max_num_of_states = np.max([len(pop_list) for pop_list in list_of_pop])
for pop_list in list_of_pop:
while len(pop_list) < max_num_of_states:
pop_list.append(0.0)
pop_array = np.array(list_of_pop)
fig, axes = plt.subplots()
time = tra[:, 0]
bottom = np.zeros(len(pop_array))
width = time / 2 * 0.5
for _, state in enumerate(pop_array.T):
_ = axes.bar(time, state, width, bottom=bottom, edgecolor="black")
bottom += state
axes.set_xlabel(r"Time resolution $\Delta t$ [frames]")
axes.set_ylabel(r"Populations fractions")
axes.set_xscale("log")
fig.savefig(title, dpi=600)
[docs]
def plot_output_multi(
title: Path,
input_data: NDArray[np.float64],
state_list: list[StateMulti],
labels: NDArray[np.int64],
):
"""
Plot a cumulative figure showing trajectories and identified states.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
input_data : ndarray of shape (n_particles, n_frames, n_features)
The input data array.
state_list : list[StateMulti]
The list of the cluster states.
labels : ndarray of shape (n_particles, n_frames)
The output of the clustering algorithm.
Example
-------
.. image:: ../_static/images/multi_Fig1.png
:alt: Example Image
:width: 600px
All the data are plotted, colored according to the cluster thay have been
assigned to. The clusters are shown as black ellipses, whose orizontal and
vertical axis length is given by the standard deviation of the Gaussians
corresponding to the cluster. Unclassififed data points are colored in
purple.
"""
n_states = len(state_list) + 1
tmp = plt.get_cmap(COLORMAP, n_states)
colors_from_cmap = tmp(np.arange(0, 1, 1 / n_states))
colors_from_cmap[-1] = tmp(1.0)
if input_data.shape[2] == 3:
fig, ax = plt.subplots(2, 2, figsize=(6, 6))
dir0 = [0, 0, 1]
dir1 = [1, 2, 2]
ax0 = [0, 0, 1]
ax1 = [0, 1, 0]
id_max, id_min = 0, 0
for idx, mol in enumerate(input_data):
if np.max(mol) == np.max(input_data):
id_max = idx
if np.min(mol) == np.min(input_data):
id_min = idx
for k in range(3):
d_0 = dir0[k]
d_1 = dir1[k]
a_0 = ax0[k]
a_1 = ax1[k]
# Plot the individual trajectories
line_w = 0.05
max_t = labels.shape[1]
step = 5 if input_data.size > 1000000 else 1
for i, mol in enumerate(input_data[::step]):
ax[a_0][a_1].plot(
mol[:, d_0], # type: ignore # mol is 2D, mypy can't see it
mol[:, d_1], # type: ignore # mol is 2D, mypy can't see it
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
color_list = labels[i * step] + 1
ax[a_0][a_1].scatter(
mol.T[d_0], # type: ignore # mol is 2D, mypy can't see it
mol.T[d_1], # type: ignore # mol is 2D, mypy can't see it
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
color_list = labels[id_min] + 1
ax[a_0][a_1].plot(
input_data[id_min].T[d_0],
input_data[id_min].T[d_1],
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
ax[a_0][a_1].scatter(
input_data[id_min].T[d_0],
input_data[id_min].T[d_1],
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
color_list = labels[id_max] + 1
ax[a_0][a_1].plot(
input_data[id_max].T[d_0],
input_data[id_max].T[d_1],
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
ax[a_0][a_1].scatter(
input_data[id_max].T[d_0],
input_data[id_max].T[d_1],
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
# Set plot titles and axis labels
ax[a_0][a_1].set_xlabel(f"Signal {d_0}")
ax[a_0][a_1].set_ylabel(f"Signal {d_1}")
ax[1][1].axis("off")
fig.savefig(title, dpi=600)
plt.close(fig)
elif input_data.shape[2] == 2:
fig, ax = plt.subplots(figsize=(6, 6))
# Plot the individual trajectories
id_max, id_min = 0, 0
for idx, mol in enumerate(input_data):
if np.max(mol) == np.max(input_data):
id_max = idx
if np.min(mol) == np.min(input_data):
id_min = idx
line_w = 0.05
max_t = labels.shape[1]
m_resized = input_data[:, :max_t:, :]
step = 5 if m_resized.size > 1000000 else 1
for i, mol in enumerate(m_resized[::step]):
ax.plot(
mol.T[0], # type: ignore # mol is 2D, mypy can't see it
mol.T[1], # type: ignore # mol is 2D, mypy can't see it
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
color_list = labels[i * step] + 1
ax.scatter(
mol.T[0], # type: ignore # mol is 2D, mypy can't see it
mol.T[1], # type: ignore # mol is 2D, mypy can't see it
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
color_list = labels[id_min] + 1
ax.plot(
m_resized[id_min].T[0],
m_resized[id_min].T[1],
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
ax.scatter(
m_resized[id_min].T[0],
m_resized[id_min].T[1],
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
color_list = labels[id_max] + 1
ax.plot(
m_resized[id_max].T[0],
m_resized[id_max].T[1],
c="black",
lw=line_w,
rasterized=True,
zorder=0,
)
ax.scatter(
m_resized[id_max].T[0],
m_resized[id_max].T[1],
c=color_list,
cmap=COLORMAP,
vmin=0,
vmax=n_states - 1,
s=0.5,
rasterized=True,
)
# Plot the Gaussian distributions of states
for state in state_list:
att = state.get_attributes()
width, height, angle = state.get_boundaries()
ellipse = Ellipse(
xy=att["mean"],
width=width,
height=height,
angle=angle,
color="black",
fill=False,
)
ax.add_patch(ellipse)
# Set plot titles and axis labels
ax.set_xlabel(r"$x$")
ax.set_ylabel(r"$y$")
fig.savefig(title, dpi=600)
[docs]
def plot_one_trj_multi(
title: Path,
example_id: int,
input_data: NDArray[np.float64],
labels: NDArray[np.int64],
):
"""Plots the colored trajectory of an example particle.
Parameters
----------
title : pathlib.Path
The path of the .png file the figure will be saved as.
example_id : int
The ID of the selected particle.
input_data : ndarray of shape (n_particles, n_frames, n_features)
The input data array.
labels : ndarray of shape (n_particles, n_frames)
The output of the clustering algorithm.
Example
-------
Here's an example of the output:
.. image:: ../_static/images/multi_Fig2.png
:alt: Example Image
:width: 600px
The datapoints are colored according to the cluster they have been
assigned to.
"""
# Get the signal of the example particle
sig_x = input_data[example_id].T[0]
sig_y = input_data[example_id].T[1]
fig, ax = plt.subplots(figsize=(6, 6))
# Create a colormap to map colors to the labels
cmap = plt.get_cmap(
COLORMAP,
int(np.max(np.unique(labels)) - np.min(np.unique(labels)) + 1),
)
color = labels[example_id]
ax.plot(sig_x, sig_y, c="black", lw=0.1)
ax.scatter(
sig_x,
sig_y,
c=color,
cmap=cmap,
vmin=float(np.min(np.unique(labels))),
vmax=float(np.max(np.unique(labels))),
s=1.0,
zorder=10,
)
# Set plot titles and axis labels
fig.suptitle(f"Example particle: ID = {example_id}")
ax.set_xlabel(r"$x$")
ax.set_ylabel(r"$y$")
fig.savefig(title, dpi=600)
[docs]
def color_trj_from_xyz(
trj_path: str,
labels: np.ndarray,
n_particles: int,
tau_window: int,
):
"""
Saves a colored .xyz file ('colored_trj.xyz') in the working directory.
Warning
-------
This function is WIP.
Parameters
----------
trj_path : str
The path to the input .xyz trajectory.
labels : np.ndarray (n_particles * n_windows,)
The output of the clustering algorithm.
n_particles : int
The number of particles in the system.
tau_window : int
The length of the signal windows.
Notes
-----
In the input file, the (x, y, z) coordinates of the particles need to be
stored in the second, third and fourth column respectively.
"""
if os.path.exists(trj_path):
with open(trj_path, "r", encoding="utf-8") as in_file:
tmp = [line.strip().split() for line in in_file]
tmp_labels = labels.reshape((n_particles, -1))
all_the_labels = np.repeat(tmp_labels, tau_window, axis=1) + 1
total_time = int(labels.shape[0] / n_particles) * tau_window
nlines = (n_particles + 2) * total_time
tmp = tmp[:nlines]
with open("colored_trj.xyz", "w+", encoding="utf-8") as out_file:
i = 0
for j in range(total_time):
print(tmp[i][0], file=out_file)
print("Properties=species:S:1:pos:R:3", file=out_file)
for k in range(n_particles):
print(
all_the_labels[k][j],
tmp[i + 2 + k][1],
tmp[i + 2 + k][2],
tmp[i + 2 + k][3],
file=out_file,
)
i += n_particles + 2
else:
raise ValueError(f"ValueError: {trj_path} not found.")