1. Processing A Line of Text

Import the English language class
from spacy.lang.en import English

# Create the nlp object
nlp = English()

# Process a text
doc = nlp("This is a sentence.")

# Print the document text, only 3 to 5
print(doc[2:5].text)

# Print the document text
print(doc.text)

Tokens

When you call nlp on a string, spaCy first tokenizes the text and creates a document object. In this exercise, you’ll learn more about the Doc, as well as its views Token and Span.

# Import the English language class and create the nlp object
from spacy.lang.en import English

nlp = English()

# Process the text
doc = nlp("I like tree kangaroos and narwhals.")

# Select the first token
first_token = doc[2]

# Print the first token's text
print(first_token.text)

3. Lexical Attributes

Use spaCy’s Doc and Token objects, and lexical attributes to find percentages in a text. We'll be looking for two subsequent tokens: a number and a percent sign.

from spacy.lang.en import English

nlp = English()

# Process the text
doc = nlp(
    "In 1990, more than 60% of people in East Asia were in extreme poverty. "
    "Now less than 4% are."
)

# Iterate over the tokens in the doc
for token in doc:
    # Check if the token resembles a number
    if token.like_num:
        # Get the next token in the document
        next_token = doc[token.i + 1]
        # Check if the next token's text equals '%'
        if next_token.text == "%":
            print("Percentage found:", token.text)

Loading Models

import spacy

# Load the small English model – spaCy is already imported
nlp = spacy.load("en_core_web_sm")

text = "It’s official: Apple is the first U.S. public company to reach a $1 trillion market value"

# Process the text
doc = nlp(text)

# Print the document text
print(doc.text)

Predicting Linguistic Annotations

1 For each token, print the token text, the token’s .pos_ (part-of-speech tag) and the token’s .dep_ (dependency label).

import spacy

nlp = spacy.load("en_core_web_sm")

text = "It’s official: Apple is the first U.S. public company to reach a $1 trillion market value"

# Process the text
doc = nlp(text)

for token in doc:
    # Get the token text, part-of-speech tag and dependency label
    token_text = token.text
    token_pos = token.pos_
    token_dep = token.dep_
    # This is for formatting only
    print("{:<12}{:<10}{:<10}".format(token_text, token_pos, token_dep))

2. Iterate over the doc.ents and print the entity text and label_ attribute.

import spacy

nlp = spacy.load("en_core_web_sm")

text = "It’s official: Apple is the first U.S. public company to reach a $1 trillion market value"

# Process the text
doc = nlp(text)

# Iterate over the predicted entities
for ent in doc.ents:
  # Print the entity text and its label
  print(ent.text, ent.label_)

Predicting named entities in context

import spacy

nlp = spacy.load("en_core_web_sm")

text = "New iPhone X release date leaked as Apple reveals pre-orders by mistake"

# Process the text
doc = nlp(text)

# Iterate over the entities
for ent in doc.ents:
    # Print the entity text and label
    print(ent.text, ent.label_)

# Get the span for "iPhone X"
iphone_x = doc[1:3]

# Print the span text
print("Missing entity:", iphone_x.text)

Note: Models are statistical and not always right. Whether their predictions are correct depends on the training data and the text you’re processing.

Rule-based matching

Using the matcher

import spacy

# Import the Matcher
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
doc = nlp("New iPhone X release date leaked as Apple reveals pre-orders by mistake")

# Initialize the Matcher with the shared vocabulary
matcher = Matcher(nlp.vocab)

# Create a pattern matching two tokens: "iPhone" and "X"
pattern = [{"TEXT": "iPhone"}, {"TEXT": "X"}]

# Add the pattern to the matcher
matcher.add("IPHONE_X_PATTERN", None, pattern)

# Use the matcher on the doc
matches = matcher(doc)
print("Matches:", [doc[start:end].text for match_id, start, end in matches])

Writing more complex match patterns using different token attributes and operators.

import spacy
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
matcher = Matcher(nlp.vocab)

doc = nlp(
    "After making the iOS update you won't notice a radical system-wide "
    "redesign: nothing like the aesthetic upheaval we got with iOS 7. Most of "
    "iOS 11's furniture remains the same as in iOS 10. But you will discover "
    "some tweaks once you delve a little deeper."
)

# Write a pattern for full iOS versions ("iOS 7", "iOS 11", "iOS 10")
pattern = [{"TEXT": "iOS"}, {"IS_DIGIT": True}]

# Add the pattern to the matcher and apply the matcher to the doc
matcher.add("IOS_VERSION_PATTERN", None, pattern)
matches = matcher(doc)
print("Total matches found:", len(matches))

# Iterate over the matches and print the span text
for match_id, start, end in matches:
    print("Match found:", doc[start:end].text)

Write one pattern that only matches forms of “download” (tokens with the lemma “download”), followed by a token with the part-of-speech tag 'PROPN' (proper noun).

import spacy
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
matcher = Matcher(nlp.vocab)

doc = nlp(
    "i downloaded Fortnite on my laptop and can't open the game at all. Help? "
    "so when I was downloading Minecraft, I got the Windows version where it "
    "is the '.zip' folder and I used the default program to unpack it... do "
    "I also need to download Winzip?"
)

# Write a pattern that matches a form of "download" plus proper noun
pattern = [{"LEMMA": "download"}, {"POS": "PROPN"}]

# Add the pattern to the matcher and apply the matcher to the doc
matcher.add("DOWNLOAD_THINGS_PATTERN", None, pattern)
matches = matcher(doc)
print("Total matches found:", len(matches))

# Iterate over the matches and print the span text
for match_id, start, end in matches:
    print("Match found:", doc[start:end].text)

Write one pattern that matches adjectives ('ADJ') followed by one or two 'NOUN's (one noun and one optional noun).

import spacy
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
matcher = Matcher(nlp.vocab)

doc = nlp(
    "Features of the app include a beautiful design, smart search, automatic "
    "labels and optional voice responses."
)

# Write a pattern for adjective plus one or two nouns
pattern = [{"POS": "ADJ"}, {"POS": "NOUN"}, {"POS": "NOUN", "OP": "?"}]

# Add the pattern to the matcher and apply the matcher to the doc
matcher.add("ADJ_NOUN_PATTERN", None, pattern)
matches = matcher(doc)
print("Total matches found:", len(matches))

# Iterate over the matches and print the span text
for match_id, start, end in matches:
    print("Match found:", doc[start:end].text)

Data Structures

Strings to hashes

Look up the string “cat” in nlp.vocab.strings to get the hash. Look up the hash to get back the string.

import spacy

nlp = spacy.load("en_core_web_sm")
doc = nlp("I have a cat")

# Look up the hash for the word "cat"
cat_hash = nlp.vocab.strings["cat"]
print(cat_hash)

# Look up the cat_hash to get the string
cat_string = nlp.vocab.strings[cat_hash]
print(cat_string)

Look up the string label “PERSON” in nlp.vocab.strings to get the hash.

import spacy

nlp = spacy.load("en_core_web_sm")
doc = nlp("David Bowie is a PERSON")

# Look up the hash for the string label "PERSON"
person_hash = nlp.vocab.strings["PERSON"]
print(person_hash)

# Look up the person_hash to get the string
person_string = nlp.vocab.strings[person_hash]
print(person_string)

mepsrajput/SpaCy.md