Word tree

ruts.visualizers.wordtree()

Description

Building a word tree (Word Tree) that shows the context of a given keyword in a text.

Note

The word tree is described in detail in this paper.

Parameters

Parameter	Type	Default	Description
texts	List[List[str]]	-	List of word lists
keyword	str	-	Keyword whose context is searched
max_n	int	5	Maximum context size
max_per_n	int	8	Maximum number of examples for each context size
**kwargs	-	-	Drawing parameters: max_font_size (default 30), min_font_size (12), font_interp - a function interpolating the font size from frequency

Usage example

Let us look at the visualizer on 100 texts of the StalinWorks dataset.

Example

Code:

# Import the libraries
import tempfile
from ruts import SentsExtractor, WordsExtractor
from ruts.datasets import StalinWorks
from ruts.visualizers import wordtree

# Prepare the data
sw = StalinWorks()
se = SentsExtractor()
we = WordsExtractor(min_len=3)
texts = [text for text in sw.get_texts(limit=50)]
text = "\n".join(texts)

# Build the list of word lists
words = []
for text in texts:
    sents = se.extract(text)
    for sent in sents:
        words.append(we.extract(sent))

# Build the tree
g = wordtree(words, "рабочий", max_n=6)

# Save the visualization to disk
g.view(tempfile.mktemp(".gv"))

Result:

Warning

Viewing the visualization requires the Graphviz tool to be installed.