Parsing a sentence

Transforming a sentence in natural language into an hyperedge is the most fundamental and quintessential task one can perform with Graphbrain.

We start by creating a parser, in this case for the English language:

from graphbrain.parsers import *
parser = create_parser(name='en')

Initializing the parser requires loading potentially large language models. This can take from a few seconds to a minute. Let’s assign some text to a variable, in this case a simple sentence:

text = "The Turing test, developed by Alan Turing in 1950, is a test of machine intelligence."

Finally, let us parse the text and print the result:

parses = parser.parse(text)
for parse in parses:
    edge = parse['main_edge']
    print(edge.to_str())

Calling the parse() method on a parser object returns a collection of parses – one per sentence. Each parse object is a dictionary, where ‘main_edge’ contains the hyperedge that directly corresponds to the sentence. Hyperedge objects have a to_str() method that can be used to produce a string representation. The code above should cause a single hyperedge to be printed to the screen.

Experiment with changing the text that is passed to the parser object and see what happens.

Working with notebooks

Jupyter notebooks are a particularly handy way to perform exploratory computation with Python, and very popular for scientific applications. Graphbrain is no exception. The notebook corresponding to this tutorial can be found here:

https://github.com/graphbrain/graphbrain/blob/master/notebooks/parser.ipynb

Notice how to import the utility functions that exist specifically for working with notebooks:

from graphbrain.notebook import *

The show() function allows one to render hyperedges in a nicer way. In the example above, we could replace the print() call with show(edge), and obtain something like this:

(is/pd.sc.|f--3s-/en
[:/b/.
[the/md/en
[+/b.mm/. turing/cp.s/en test/cc.s/en]]
(developed/pc.ax.<pf----/en
(by/x/en
[+/b.am/. alan/cp.s/en turing/cp.s/en])
{in/tt/en 1950/c#/en})]
[a/md/en
[of/br.ma/en
test/cc.s/en
[+/b.am/. machine/cc.s/en intelligence/cc.s/en]]])


The show() function provides several visualization styles, and also the possibility of reducing visual clutter by only displaying the roots of the atoms. Refer to the function signature for all the details.