# yellowbrick.text.dispersion
# Implementations of lexical dispersions for text visualization.
#
# Author: Larry Gray
# Created: Fri Jun 22 15:40:49 2018 -0400
#
# Copyright (C) 2018 The scikit-yb developers
# For license information, see LICENSE.txt
#
# ID: dispersion.py [3822dd6] lwgray@gmail.com $
"""
Implementation of lexical dispersion for text visualization
"""
##########################################################################
## Imports
##########################################################################
import itertools
from collections import defaultdict
import numpy as np
from yellowbrick.text.base import TextVisualizer
from yellowbrick.style.colors import resolve_colors
from yellowbrick.exceptions import YellowbrickValueError
##########################################################################
## Dispersion Plot Visualizer
##########################################################################
[docs]class DispersionPlot(TextVisualizer):
"""
Lexical dispersion illustrates the homogeneity of a word (or set of words) across
the documents of a corpus.
DispersionPlot allows for visualization of the lexical dispersion of words in a
corpus. This plot illustrates with vertical lines the occurrences of one or more
search terms throughout the corpus, noting how many words relative to the beginning
of the corpus it appears. If the target vector of the corpus documents is provided,
the points will be colored with respect to their document category, which allows for
additional analysis of relationships in search term homogeneity within and across
document categories. If annotation is requested, document boundaries will be
displayed as vertical lines in the plot.
Parameters
----------
search_terms : list
A list of search terms whose dispersion across a corpus passed at fit
should be visualized.
ax : matplotlib axes, default: None
The axes to plot the figure on.
colors : list or tuple of colors
Specify the colors for each individual class. Will override colormap if both are
provided.
colormap : string or matplotlib cmap
Qualitative colormap for discrete target
ignore_case : boolean, default: False
Specify whether input will be case-sensitive.
annotate_docs : boolean, default: False
Specify whether document boundaries will be displayed. Vertical lines
are positioned at the end of each document.
labels : list of strings
The names of the classes in the target, used to create a legend.
Labels must match names of classes in sorted order.
kwargs : dict
Pass any additional keyword arguments to the super class.
Attributes
----------
self.classes_ : list
A list of strings representing the unique classes in the target in sorted order.
If ``y`` is provided, these are extracted from ``y``, unless a list of class
labels is provided by the user on instantiation.
self.boundaries_ : list
A list of integers indicating the document boundaries with respect to
word offsets.
self.indexed_words_ : list
A list of integers indicating the y position for each occurrence of each of
the search terms.
self.word_categories_ : list
A list of strings indicating the corresponding document category of each search
term occurrence.
"""
# NOTE: cannot be np.nan
NULL_CLASS = None
def __init__(
self,
search_terms,
ax=None,
colors=None,
colormap=None,
ignore_case=False,
annotate_docs=False,
labels=None,
**kwargs
):
super(DispersionPlot, self).__init__(ax=ax, **kwargs)
self.labels = labels
self.colors = colors
self.colormap = colormap
self.ignore_case = ignore_case
self.search_terms = search_terms
self.annotate_docs = annotate_docs
def _compute_dispersion(self, X, y):
"""
Produces a generator containing the offset word count, y_coordinate, and
label/category for each occurrance of the search terms.
Attributes
----------
self.boundaries_ : list
A list of integers indicating the document boundaries with respect to
word offsets.
"""
self.boundaries_ = []
offset = 0
if y is None:
y = itertools.repeat(None)
for doc, category in zip(X, y):
for word in doc:
if self.ignore_case:
word = word.lower()
# NOTE: this will find all indices if duplicate words are supplied
# In the case that word is not in target words, any empty list is
# returned and no data will be yielded
offset += 1
for y_coord in (self.indexed_words_ == word).nonzero()[0]:
y_coord = int(y_coord)
yield (offset, y_coord, category)
if self.annotate_docs:
self.boundaries_.append(offset)
self.boundaries_ = np.array(self.boundaries_, dtype=int)
def _check_missing_words(self, points):
"""
Helper method to raise an error if any of the requested search
terms do not appear in the corpus.
"""
for index in range(len(self.indexed_words_)):
if index in points[:, 1]:
pass
else:
raise YellowbrickValueError(
("The search term '{}' is not found in " "this corpus").format(
self.indexed_words_[index]
)
)
[docs] def fit(self, X, y=None, **kwargs):
"""
The fit method is the primary drawing input for the dispersion
visualization.
Parameters
----------
X : list or generator
Should be provided as a list of documents or a generator
that yields a list of documents that contain a list of
words in the order they appear in the document.
y : ndarray or Series of length n
An optional array or series of target or class values for
instances. If this is specified, then the points will be colored
according to their class.
Returns
-------
self : instance
Returns the instance of the transformer/visualizer
Attributes
----------
self.classes_ : list
A list of strings representing the unique classes in the target in sorted order.
If ``y`` is provided, these are extracted from ``y``, unless a list of class
labels is provided by the user on instantiation.
self.indexed_words_ : list
A list of integers indicating the y position for each occurrence of each of
the search terms.
self.word_categories_ : list
A list of strings indicating the corresponding document category of each search
term occurrence.
"""
if y is not None:
self.classes_ = np.unique(y)
else:
self.classes_ = np.array([self.NULL_CLASS])
# Create an index (e.g. the y position) for the target words
self.indexed_words_ = np.flip(self.search_terms, axis=0)
if self.ignore_case:
self.indexed_words_ = np.array([w.lower() for w in self.indexed_words_])
# Stack is used to create a 2D array from the generator
try:
offsets_positions_categories = np.stack(self._compute_dispersion(X, y))
except ValueError:
raise YellowbrickValueError(("No search terms were found in the corpus"))
word_positions = np.stack(
zip(
offsets_positions_categories[:, 0].astype(int),
offsets_positions_categories[:, 1].astype(int),
)
)
self.word_categories_ = offsets_positions_categories[:, 2]
self._check_missing_words(word_positions)
self.draw(word_positions, **kwargs)
return self
[docs] def draw(self, points, **kwargs):
"""
Called from the fit method, this method creates the canvas and
draws the plot on it.
Parameters
----------
kwargs: generic keyword arguments.
"""
# Resolve the labels with the classes
labels = self.labels if self.labels is not None else self.classes_
if len(labels) != len(self.classes_):
raise YellowbrickValueError(
(
"number of supplied labels ({}) does not "
"match the number of classes ({})"
).format(len(labels), len(self.classes_))
)
# Create the color mapping for the labels.
color_values = resolve_colors(
n_colors=len(labels), colormap=self.colormap, colors=self.colors
)
colors = dict(zip(labels, color_values))
# Transform labels into a map of class to label
labels = dict(zip(self.classes_, labels))
# Define boundaries with a vertical line
if self.annotate_docs:
for xcoords in self.boundaries_:
self.ax.axvline(x=xcoords, color="lightgray", linestyle="dashed")
series = defaultdict(lambda: {"x": [], "y": []})
if self.word_categories_ is not None:
for point, t in zip(points, self.word_categories_):
label = labels[t]
series[label]["x"].append(point[0])
series[label]["y"].append(point[1])
else:
label = self.classes_[0]
for x, y in points:
series[label]["x"].append(x)
series[label]["y"].append(y)
for label, points in series.items():
self.ax.scatter(
points["x"],
points["y"],
marker="|",
c=colors[label],
zorder=100,
label=label,
)
self.ax.set_yticks(list(range(len(self.indexed_words_))))
self.ax.set_yticklabels(self.indexed_words_)
return self.ax
[docs] def finalize(self, **kwargs):
"""
Prepares the figure for rendering by adding a title, axis labels, and
managing the limits of the text labels. Adds a legend outside of the plot.
Parameters
----------
kwargs: generic keyword arguments.
Notes
-----
Generally this method is called from show and not directly by the user.
"""
self.set_title("Lexical Dispersion Plot")
self.ax.set_ylim(-1, len(self.indexed_words_))
self.ax.set_xlabel("Word Offset")
self.ax.grid(False)
# Add the legend outside of the figure box.
if not all(self.classes_ == np.array([self.NULL_CLASS])):
box = self.ax.get_position()
self.ax.set_position([box.x0, box.y0, box.width * 0.8, box.height])
self.ax.legend(loc="center left", bbox_to_anchor=(1, 0.5))
##########################################################################
## Quick Method
##########################################################################
[docs]def dispersion(
search_terms,
corpus,
y=None,
ax=None,
colors=None,
colormap=None,
annotate_docs=False,
ignore_case=False,
labels=None,
show=True,
**kwargs
):
""" Displays lexical dispersion plot for words in a corpus
This helper function is a quick wrapper to utilize the DispersionPlot
Visualizer for one-off analysis
Parameters
----------
search_terms : list
A list of words whose dispersion will be examined within a corpus
corpus : list
Should be provided as a list of documents that contain
a list of words in the order they appear in the document.
y : ndarray or Series of length n
An optional array or series of target or class values for
instances. If this is specified, then the points will be colored
according to their class.
ax : matplotlib axes, default: None
The axes to plot the figure on.
colors : list or tuple of colors
Specify the colors for each individual class. Will override colormap if both are
provided.
colormap : string or matplotlib cmap
Qualitative colormap for discrete target
annotate_docs : boolean, default: False
Specify whether document boundaries will be displayed. Vertical lines
are positioned at the end of each document.
ignore_case : boolean, default: False
Specify whether input will be case-sensitive.
labels : list of strings
The names of the classes in the target, used to create a legend.
Labels must match names of classes in sorted order.
show : bool, default: True
If True, calls ``show()``, which in turn calls ``plt.show()`` however
you cannot call ``plt.savefig`` from this signature, nor
``clear_figure``. If False, simply calls ``finalize()``
kwargs : dict
Pass any additional keyword arguments to the super class.
Returns
-------
viz: DispersionPlot
Returns the fitted, finalized visualizer
"""
# Instantiate the visualizer
visualizer = DispersionPlot(
search_terms,
ax=ax,
colors=colors,
colormap=colormap,
ignore_case=ignore_case,
labels=labels,
annotate_docs=annotate_docs,
**kwargs
)
visualizer.fit(corpus, y, **kwargs)
if show:
visualizer.show()
else:
visualizer.finalize()
return visualizer