def plot_dendrogram(data, labels=None, ax=None):
"""Plot a dendrogram of the given data.
Parameters
----------
data : 2d array
Data to plot in a dendrogram.
labels : list of str, optional
Labels for the dendrogram.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
Notes
-----
This function is a wrapper of the scipy `dendrogram` plot function.
Examples
--------
See the example for the :meth:`~.compute_score` method of the :class:`~.Counts` class.
"""
linkage_data = hier.linkage(data, method='complete', metric='cosine')
hier.dendrogram(linkage_data,
orientation='left',
color_threshold=0.25,
leaf_font_size=12,
ax=check_ax(ax),
**check_args(['labels'], labels))
plt.tight_layout()
def plot_years(years, year_range=None, ax=None):
"""Plot a histogram of the number publications across years.
Parameters
----------
years : collections.Counter
Data on the number of publications per year.
year_range : list of [int, int], optional
The range of years to plot on the x-axis.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
"""
ax = check_ax(ax, (10, 5))
# Extract x & y data to plot
x_dat = list(years.keys())
y_dat = list(years.values())
# Add line and points to plot
plt.plot(x_dat, y_dat)
plt.plot(x_dat, y_dat, '.', markersize=16)
# Set plot limits
if year_range:
plt.xlim([year_range[0], year_range[1]])
plt.ylim([0, max(y_dat) + 3])
# Add title & labels
plt.title('Publication History', fontsize=24, fontweight='bold')
plt.xlabel('Year of Publication', fontsize=18)
plt.ylabel('Number of Articles', fontsize=18)
def plot_matrix(data,
x_labels=None,
y_labels=None,
attribute='score',
transpose=False,
cmap='purple',
square=False,
ax=None,
**kwargs):
"""Plot a matrix as a heatmap.
Parameters
----------
data : Counts or 2d array
Data to plot in matrix format.
x_labels : list of str, optional
Labels for the x-axis.
y_labels : list of str, optional
Labels for the y-axis.
attribute : {'score', 'counts'}, optional
Which data attribute from the counts object to plot the data for.
Only used if the `data` input is a Counts object.
transpose : bool, optional, default: False
Whether to transpose the data before plotting.
cmap : {'purple', 'blue'} or matplotlib.cmap
Colormap to use for the plot.
If string, uses a sequential palette of the specified color.
square : bool
Whether to plot all the cells as equally sized squares.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
**kwargs
Additional keyword arguments to pass through to seaborn.heatmap.
Notes
-----
This function is a wrapper of the seaborn `heatmap` plot function.
Examples
--------
See the example for the :meth:`~.compute_score` method of the :class:`~.Counts` class.
"""
if isinstance(cmap, str):
cmap = get_cmap(cmap)
data, x_labels, y_labels = counts_data_helper(data, x_labels, y_labels,
attribute, transpose)
with sns.plotting_context("notebook",
font_scale=kwargs.pop('font_scale', 1.0)):
sns.heatmap(data,
square=square,
ax=check_ax(ax, kwargs.pop('figsize', None)),
cmap=cmap,
**check_args(['xticklabels', 'yticklabels'], x_labels,
y_labels),
**kwargs)
plt.tight_layout()
def plot_dendrogram(data,
labels=None,
attribute='score',
transpose=False,
method='complete',
metric='cosine',
ax=None,
**kwargs):
"""Plot a dendrogram of the given data based on hierarchical clustering.
Parameters
----------
data : Counts or 2d array
Data to plot in a dendrogram.
labels : list of str, optional
Labels for the dendrogram.
attribute : {'score', 'counts'}, optional
Which data attribute from the counts object to plot the data for.
Only used if the `data` input is a Counts object.
transpose : bool, optional, default: False
Whether to transpose the data before plotting.
method : str, optional, default: 'complete'
The linkage algorithm to use. See `scipy.cluster.hierarchy.linkage` for options.
metric : str or function, optional, default: 'cosine'
The distance metric to use. See `scipy.spatial.distance.pdist` for options.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
**kwargs
Additional keyword arguments to pass through to scipy.cluster.hierarchy.dendrogram.
Notes
-----
This function is a wrapper of the `scipy.cluster.hierarchy.dendrogram' plot function.
Examples
--------
See the example for the :meth:`~.compute_score` method of the :class:`~.Counts` class.
"""
if isinstance(data, Counts):
labels = data.terms['A' if not transpose else 'B'].labels
data = getattr(data, attribute).T if transpose else getattr(
data, attribute)
linkage_data = hier.linkage(data, method=method, metric=metric)
with sns.plotting_context("notebook",
font_scale=kwargs.pop('font_scale', 1.0)):
hier.dendrogram(linkage_data,
orientation=kwargs.pop('orientation', 'left'),
color_threshold=kwargs.pop('color_threshold', 0.25),
leaf_font_size=kwargs.pop('leaf_font_size', 12),
ax=check_ax(ax, kwargs.pop('figsize', None)),
**check_args(['labels'], labels),
**kwargs)
plt.tight_layout()
def plot_vector(data,
dim='A',
transpose=False,
cmap='purple',
ax=None,
**kwargs):
"""Plot a vector as an annotated heatmap.
Parameters
----------
data : Counts or 1d array
Data to plot as a heatmap.
dim : {'A', 'B'}, optional
Which set of terms to plot.
Only used if `data` is a `Counts` object.
transpose : bool, optional, default: False
Whether to transpose the data before plotting.
cmap : {'purple', 'blue'} or matplotlib.cmap
Colormap to use for the plot.
If string, uses a sequential palette of the specified color.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
**kwargs
Additional keyword arguments to pass through to seaborn.heatmap.
"""
if isinstance(cmap, str):
cmap = get_cmap(cmap)
if isinstance(data, Counts):
data = data.terms[dim].counts
if data.ndim == 1:
data = np.expand_dims(data, 1)
if transpose:
data = data.T
sns.heatmap(data,
cmap=cmap,
square=kwargs.pop('square', True),
annot=kwargs.pop('annot', True),
fmt=kwargs.pop('fmt', 'd'),
annot_kws={"size": 18},
cbar=kwargs.pop('cbar', False),
xticklabels=kwargs.pop('xticklabels', []),
yticklabels=kwargs.pop('yticklabels', []),
ax=check_ax(ax, kwargs.pop('figsize', None)),
**kwargs)
def plot_wordcloud(freq_dist, n_words, ax=None):
"""Plot a wordcloud.
Parameters
----------
freq_dist : nltk.FreqDist
Frequency distribution of words to plot.
n_words : int
Number of top words to include in the wordcloud.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
"""
cloud = create_wordcloud(conv_freqs(freq_dist, n_words))
ax = check_ax(ax, (8, 8))
ax.imshow(cloud)
ax.axis("off")
def plot_matrix(data,
x_labels=None,
y_labels=None,
cmap='purple',
square=False,
ax=None):
"""Plot a matrix representation of given data.
Parameters
----------
data : 2d array
Data to plot in matrix format.
x_labels : list of str
Labels for the x-axis.
y_labels : list of str
Labels for the y-axis.
cmap : {'purple', 'blue'} or matplotlib.cmap
Colormap to use for the plot.
If string, uses a sequential palette of the specified color.
square : bool
Whether to plot all the cells as equally sized squares.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
Notes
-----
This function is a wrapper of the seaborn `heatmap` plot function.
Examples
--------
See the example for the :meth:`~.compute_score` method of the :class:`~.Counts` class.
"""
if isinstance(cmap, str):
cmap = get_cmap(cmap)
sns.heatmap(data,
square=square,
ax=check_ax(ax),
cmap=cmap,
**check_args(['xticklabels', 'yticklabels'], x_labels,
y_labels))
plt.tight_layout()
def plot_wordcloud(freq_dist, n_words, ax=None, **plt_kwargs):
"""Plot a wordcloud.
Parameters
----------
freq_dist : collections.Counter
Frequency distribution of words to plot.
n_words : int
Number of top words to include in the wordcloud.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
plt_kwargs
Additional keyword arguments for the plot.
Examples
--------
See the :meth:`~.create_freq_dist` method of the :class:`~.ArticlesAll` object.
"""
cloud = create_wordcloud(conv_freqs(freq_dist, n_words))
ax = check_ax(ax, plt_kwargs.pop('figsize', (8, 8)))
ax.imshow(cloud, **plt_kwargs)
ax.axis("off")
def plot_years(years, year_range=None, ax=None, **plt_kwargs):
"""Plot a histogram of the number publications across years.
Parameters
----------
years : collections.Counter
Data on the number of publications per year.
year_range : list of [int, int], optional
The range of years to plot on the x-axis, inclusive.
ax : matplotlib.Axes, optional
Figure axes upon which to plot.
plt_kwargs
Additional keyword arguments for the plot.
Examples
--------
Plot a histogram of publication years:
>>> from collections import Counter
>>> plot_years(years=Counter({'2018': 25, '2019': 50, '2020':75}))
Notes
-----
Publication years are collected together in the :class:`~.ArticlesAll` class.
"""
ax = check_ax(ax, plt_kwargs.pop('figsize', (10, 5)))
# Get the plot data, making sure it is sorted
sort_inds = np.argsort(list(years.keys()))
x_data = np.array(list(years.keys()))[sort_inds]
y_data = np.array(list(years.values()))[sort_inds]
# Restrict the data to the desired plot range
if year_range:
range_inds = np.logical_and(
x_data >= (year_range[0] if year_range[0] else -np.inf), x_data <=
(year_range[1] if year_range[1] else np.inf))
x_data = x_data[range_inds]
y_data = y_data[range_inds]
# Grab any plot inputs for labels
fontsize = plt_kwargs.pop('fontsize', 18)
xlabel = plt_kwargs.pop('xlabel', 'Year of Publication')
ylabel = plt_kwargs.pop('ylabel', 'Number of Articles')
# Add line and points to plot
ax.plot(x_data,
y_data,
linewidth=check_aliases(plt_kwargs, ['linewidth', 'lw'], 3),
marker=plt_kwargs.pop('marker', '.'),
markersize=check_aliases(plt_kwargs, ['markersize', 'ms'], 10),
markerfacecolor=plt_kwargs.pop('markerfacecolor', 'white'),
**plt_kwargs)
# Set plot limits
ax.set_ylim(
[0, max(y_data) + int(np.ceil(0.03 * (max(y_data) - min(y_data))))])
# Add title & labels
ax.set_xlabel(xlabel, fontsize=fontsize)
ax.set_ylabel(ylabel, fontsize=fontsize)