spaCy Matcher and PhraseMatcher

Natural Language Processing with spaCy

Azadeh Mobasher

Principal Data Scientist

Matcher in spaCy

RegEx patterns can be complex, difficult to read and debug.
spaCy provides a readable and production-level alternative, the Matcher class.

import spacy
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
doc = nlp("Good morning, this is our first day on campus.")

matcher = Matcher(nlp.vocab)

Matcher in spaCy

Matching output include start and end token indices of the matched pattern.

pattern = [{"LOWER": "good"}, {"LOWER": "morning"}]

matcher.add("morning_greeting", [pattern])

matches = matcher(doc)
for match_id, start, end in matches:
    print("Start token: ", start, " | End token: ", end, 
          "| Matched text: ", doc[start:end].text)

>>> Start token:  0  | End token:  2 | Matched text:  Good morning

Matcher extended syntax support

Allows operators in defining the matching patterns.
Similar operators to Python's in, not in and comparison operators

Attribute	Value type	Description
`IN`	any type	Attribute value is a member of a list
`NOT_IN`	any type	Attribute value is *not* a member of a list
`==`, `>=`, `<=`, `>`, `<`	int, float	Comparison operators for equality or inequality checks

Matcher extended syntax support

Using IN operator to match both good morning and good evening

doc = nlp("Good morning and good evening.")
matcher = Matcher(nlp.vocab)
pattern = [{"LOWER": "good"}, {"LOWER": {"IN": ["morning", "evening"]}}]
matcher.add("morning_greeting", [pattern])
matches = matcher(doc)

The output of matching using IN operator

for match_id, start, end in matches:
    print("Start token: ", start, " | End token: ", end,
          "| Matched text: ", doc[start:end].text)

>>> Start token:  0  | End token:  2 | Matched text:  Good morning
Start token:  3  | End token:  5 | Matched text:  good evening

PhraseMatcher in spaCy

PhraseMatcher class matches a long list of phrases in a given text.

from spacy.matcher import PhraseMatcher
nlp = spacy.load("en_core_web_sm")
matcher = PhraseMatcher(nlp.vocab)
terms = ["Bill Gates", "John Smith"]

PhraseMatcher in spaCy

PhraseMatcher outputs include start and end token indices of the matched pattern

patterns = [nlp.make_doc(term) for term in terms]
matcher.add("PeopleOfInterest", patterns)

doc = nlp("Bill Gates met John Smith for an important discussion regarding 
           importance of AI.")

matches = matcher(doc)
for match_id, start, end in matches:
    print("Start token: ", start, " | End token: ", end, 
          "| Matched text: ", doc[start:end].text)

>>> Start token:  0  | End token:  2 | Matched text:  Bill Gates
Start token:  3  | End token:  5 | Matched text:  John Smith

PhraseMatcher in spaCy

We can use attr argument of the PhraseMatcher class

matcher = PhraseMatcher(nlp.vocab, attr = "LOWER")

terms = ["Government", "Investment"]
patterns = [nlp.make_doc(term) for term in terms]
matcher.add("InvestmentTerms", patterns)
doc = nlp("It was interesting to the investment division of the government.")

matcher = PhraseMatcher(nlp.vocab, attr = "SHAPE")

terms = ["110.0.0.0", "101.243.0.0"]
patterns = [nlp.make_doc(term) for term in terms]
matcher.add("IPAddresses", patterns)
doc = nlp("The tracked IP address was 234.135.0.0.")

Let's practice!

Natural Language Processing with spaCy