Skip to main content

bnaug is a text augmentation tool for Bangla text.

Project description

bnaug (Bangla Text Augmentation)

bnaug is a text augmentation tool for Bangla text.

Installation

pip install bnaug
  • Dependencies
    • pytorch >=1.7.0

Demo Notebook

Necessary Model Links

Sentence Augmentation

Token Replacement

  • Mask generation based augmentation

    from bnaug.sentence import TokenReplacement
    
    tokr = TokenReplacement()
    text = "আমি ঢাকায় বাস করি।"
    output = tokr.masking_based(text, sen_n=5)
    
  • Word2Vec based augmentation

    from bnaug.sentence import TokenReplacement
    
    tokr = TokenReplacement()
    text = "আমি ঢাকায় বাস করি।"
    model = "msc/bangla_word2vec/bnwiki_word2vec.model"
    output = tokr.word2vec_based(text, model=model, sen_n=5, word_n=5)
    print(output)
    
  • Glove based augmentation

    from bnaug.sentence import TokenReplacement
    
    tokr = TokenReplacement()
    text = "আমি ঢাকায় বাস করি।"
    vector = "msc/bn_glove.300d.txt"
    output = tokr.glove_based(text, vector_path=vector, sen_n=5, word_n=5)
    print(output)
    

Back Translation

Back translation based augmentation first translate Bangla sentence to English and then again translate the English to Bangla.

from bnaug.sentence import BackTranslation

bt = BackTranslation()
text = "বাংলা ভাষা আন্দোলন তদানীন্তন পূর্ব পাকিস্তানে সংঘটিত একটি সাংস্কৃতিক ও রাজনৈতিক আন্দোলন। "
output = bt.get_augmented_sentences(text)
print(output)

Text Generation

  • Paraphrase generation
from bnaug.sentence import TextGeneration

tg = TextGeneration()
text = "বিমানটি যখন মাটিতে নামার জন্য এয়ারপোর্টের কাছাকাছি আসছে, তখন ল্যান্ডিং গিয়ারের খোপের ঢাকনাটি খুলে যায়।"
output = tg.parapharse_generation(text)
print(output)

Random Augmentation

  • Random remove part and generate new sentence

    At present it's removing word, stopwords, punctuations, numbers and generate new sentences

    from bnaug.sentence import RandomAugmentation
    
    raug = RandomAugmentation()
    sentence = "আমি ১০০ বাকি দিলাম"
    output = raug.random_remove(sentence)
    print(output)
    

    or apply individually

    from bnaug import randaug
    
    text = "১০০ বাকি দিলাম"
    output = randaug.remove_digits(text)
    print(output)
    
    text = "১০০! বাকি দিলাম?"
    output = randaug.remove_punctuations(text)
    print(output)
    
    text = "আমি ১০০ বাকি দিলাম"
    randaug.remove_stopwords(text)
    print(output)
    
    text = "আমি ১০০ বাকি দিলাম"
    randaug.remove_random_word(text)
    print(output)
    
    text = "আমি ১০০ বাকি দিলাম"
    randaug.remove_random_char(text)
    print(output)
    

Inspired from

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

bnaug-1.1.2.tar.gz (5.1 kB view details)

Uploaded Source

Built Distribution

bnaug-1.1.2-py3-none-any.whl (4.8 kB view details)

Uploaded Python 3

File details

Details for the file bnaug-1.1.2.tar.gz.

File metadata

  • Download URL: bnaug-1.1.2.tar.gz
  • Upload date:
  • Size: 5.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.18

File hashes

Hashes for bnaug-1.1.2.tar.gz
Algorithm Hash digest
SHA256 c524078fceb1b2edbef5b0a7e7a4cccc912333e6d5248412ef1a59a6ee18d2f9
MD5 a7ca049f0e36bde4f944b882849bd3d9
BLAKE2b-256 15efba3f00852c102db73029c8cec8ae96c10a2cada28a8b2db4610b114db91c

See more details on using hashes here.

File details

Details for the file bnaug-1.1.2-py3-none-any.whl.

File metadata

  • Download URL: bnaug-1.1.2-py3-none-any.whl
  • Upload date:
  • Size: 4.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.18

File hashes

Hashes for bnaug-1.1.2-py3-none-any.whl
Algorithm Hash digest
SHA256 eb069f70fa0f7af3fcf2f58d5a9d6f1cf8f0c18fe3b1030630bb75795e0e9fe5
MD5 6ce6e580f6e55c0a43501a73bcfd5766
BLAKE2b-256 198a7a0e8389bec7c8694270d13979d93e946261f2497465af5c9fc6f53782c9

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page