text

Text

Augmentations

Try running some AugLy text augmentations in Colab! For a full list of available augmentations, see here.

Our text augmentations use nlpaug as their backbone. All functions accept a list of original texts to be augmented as input and return the list of augmented texts.

Function-based

You can call the functional augmentations like so:

import augly.text as txtaugs

texts = ["hello world", "bye planet"]

augmented_synonyms = txtaugs.insert_punctuation_chars(
    texts,
    granularity="all",
    cadence=5.0,
    vary_chars=True,
)

Class-based

You can also call any augmentation as a Transform class with a given probability:

import augly.text as txtaugs

texts = ["hello world", "bye planet"]
transform = InsertPunctuationChars(granularity="all", p=0.5)
aug_texts = transform(texts)

Unit Tests

You can run our text unit tests if you have cloned augly (see here) by running the following:

python -m unittest discover -s augly/tests/text_tests/ -p "*"