spellCheck

Contextual word checker for better suggestions

Types of spelling mistakes

It is essential to understand that identifying whether a candidate is a spelling error is a big task.

Spelling errors are broadly classified as non- word errors (NWE) and real word errors (RWE). If the misspelt string is a valid word in the language, then it is called an RWE, else it is an NWE.

-- Monojit Choudhury et. al. (2007)

This package currently focuses on Out of Vocabulary (OOV) word or non-word error (NWE) correction using BERT model. The idea of using BERT was to use the context when correcting OOV. To improve this package, I would like to extend the functionality to identify RWE, optimising the package, and improving the documentation.

Install

The package can be installed using pip. You would require python 3.6+

pip install contextualSpellCheck

Usage

Note: For use in other languages check examples folder.

How to load the package in spacy pipeline

>>> import contextualSpellCheck
>>> import spacy
>>> nlp = spacy.load("en_core_web_sm") 
>>> 
>>> ## We require NER to identify if a token is a PERSON
>>> ## also require parser because we use `Token.sent` for context
>>> nlp.pipe_names
['tok2vec', 'tagger', 'parser', 'ner', 'attribute_ruler', 'lemmatizer']
>>> contextualSpellCheck.add_to_pipe(nlp)
>>> nlp.pipe_names
['tok2vec', 'tagger', 'parser', 'ner', 'attribute_ruler', 'lemmatizer', 'contextual spellchecker']
>>> 
>>> doc = nlp('Income was $9.4 milion compared to the prior year of $2.7 milion.')
>>> doc._.outcome_spellCheck
'Income was $9.4 million compared to the prior year of $2.7 million.'

Or you can add to spaCy pipeline manually!

>>> import spacy
>>> import contextualSpellCheck
>>> 
>>> nlp = spacy.load("en_core_web_sm")
>>> nlp.pipe_names
['tok2vec', 'tagger', 'parser', 'ner', 'attribute_ruler', 'lemmatizer']
>>> # You can pass the optional parameters to the contextualSpellCheck
>>> # eg. pass max edit distance use config={"max_edit_dist": 3}
>>> nlp.add_pipe("contextual spellchecker")
<contextualSpellCheck.contextualSpellCheck.ContextualSpellCheck object at 0x1049f82b0>
>>> nlp.pipe_names
['tok2vec', 'tagger', 'parser', 'ner', 'attribute_ruler', 'lemmatizer', 'contextual spellchecker']
>>> 
>>> doc = nlp("Income was $9.4 milion compared to the prior year of $2.7 milion.")
>>> print(doc._.performed_spellCheck)
True
>>> print(doc._.outcome_spellCheck)
Income was $9.4 million compared to the prior year of $2.7 million.

After adding contextual spellchecker in the pipeline, you use the pipeline normally. The spell check suggestions and other data can be accessed using extensions.

Using the pipeline

>>> doc = nlp(u'Income was $9.4 milion compared to the prior year of $2.7 milion.')
>>> 
>>> # Doc Extention
>>> print(doc._.contextual_spellCheck)
True
>>> print(doc._.performed_spellCheck)
True
>>> print(doc._.suggestions_spellCheck)
{milion: 'million', milion: 'million'}
>>> print(doc._.outcome_spellCheck)
Income was $9.4 million compared to the prior year of $2.7 million.
>>> print(doc._.score_spellCheck)
{milion: [('million', 0.59422), ('billion', 0.24349), (',', 0.08809), ('trillion', 0.01835), ('Million', 0.00826), ('%', 0.00672), ('##M', 0.00591), ('annually', 0.0038), ('##B', 0.00205), ('USD', 0.00113)], milion: [('billion', 0.65934), ('million', 0.26185), ('trillion', 0.05391), ('##M', 0.0051), ('Million', 0.00425), ('##B', 0.00268), ('USD', 0.00153), ('##b', 0.00077), ('millions', 0.00059), ('%', 0.00041)]}
>>> 
>>> # Token Extention
>>> print(doc[4]._.get_require_spellCheck)
True
>>> print(doc[4]._.get_suggestion_spellCheck)
'million'
>>> print(doc[4]._.score_spellCheck)
[('million', 0.59422), ('billion', 0.24349), (',', 0.08809), ('trillion', 0.01835), ('Million', 0.00826), ('%', 0.00672), ('##M', 0.00591), ('annually', 0.0038), ('##B', 0.00205), ('USD', 0.00113)]
>>> 
>>> # Span Extention
>>> print(doc[2:6]._.get_has_spellCheck)
True
>>> print(doc[2:6]._.score_spellCheck)
{$: [], 9.4: [], milion: [('million', 0.59422), ('billion', 0.24349), (',', 0.08809), ('trillion', 0.01835), ('Million', 0.00826), ('%', 0.00672), ('##M', 0.00591), ('annually', 0.0038), ('##B', 0.00205), ('USD', 0.00113)], compared: []}

Extensions

To make the usage easy, contextual spellchecker provides custom spacy extensions which your code can consume. This makes it easier for the user to get the desired data. contextualSpellCheck provides extensions on the doc, span and token level. Below tables summaries the extensions.

spaCy.Doc level extensions

spaCy.Span level extensions

spaCy.Token level extensions

API

At present, there is a simple GET API to get you started. You can run the app in your local and play with it.

Query: You can use the endpoint http://127.0.0.1:5000/?query=YOUR-QUERY Note: Your browser can handle the text encoding

GET Request: http://localhost:5000/?query=Income%20was%20$9.4%20milion%20compared%20to%20the%20prior%20year%20of%20$2.7%20milion.

Response:

{
    "success": true,
    "input": "Income was $9.4 milion compared to the prior year of $2.7 milion.",
    "corrected": "Income was $9.4 milion compared to the prior year of $2.7 milion.",
    "suggestion_score": {
        "milion": [
            [
                "million",
                0.59422
            ],
            [
                "billion",
                0.24349
            ],
            ...
        ],
        "milion:1": [
            [
                "billion",
                0.65934
            ],
            [
                "million",
                0.26185
            ],
            ...
        ]
    }
}

Task List

• use cython for part of the code to improve performance (#39)
• Improve metric for candidate selection (#40)
• Add examples for other langauges (#41)
• Update the logic of misspell identification (OOV) (#44)
• better candidate generation (solved by #44?)
• add metric by testing on datasets
• Improve documentation
• Improve logging in code
• Add support for Real Word Error (RWE) (Big Task)
• add multi mask out capability

Support and contribution

If you like the project, please ⭑ the project and show your support! Also, if you feel, the current behaviour is not as expected, please feel free to raise an issue. If you can help with any of the above tasks, please open a PR with necessary changes to documentation and tests.

Cite

If you are using contextualSpellCheck in your academic work, please consider citing the library using the below BibTex entry:

@misc{Goel_Contextual_Spell_Check_2021,
author = {Goel, Rajat},
doi = {10.5281/zenodo.4642379},
month = {3},
title = {{Contextual Spell Check}},
url = {https://github.com/R1j1t/contextualSpellCheck},
year = {2021}
}

Reference

Below are some of the projects/work I referred to while developing this package

1. Explosion AI.Architecture. May 2020. url:https://spacy.io/api.
2. Monojit Choudhury et al. “How difficult is it to develop a perfect spell-checker? A cross-linguistic analysis through complex network approach”. In:arXiv preprint physics/0703198(2007).
3. Jacob Devlin et al. BERT: Pre-training of Deep Bidirectional Transform-ers for Language Understanding. 2019. arXiv:1810.04805 [cs.CL].
4. Hugging Face.Fast Coreference Resolution in spaCy with Neural Net-works. May 2020. url:https://github.com/huggingface/neuralcoref.
5. Ines.Chapter 3: Processing Pipelines. May 20202. url:https://course.spacy.io/en/chapter3.
6. Eric Mays, Fred J Damerau, and Robert L Mercer. “Context based spellingcorrection”. In:Information Processing & Management27.5 (1991), pp. 517–522.
7. Peter Norvig. How to Write a Spelling Corrector. May 2020. url:http://norvig.com/spell-correct.html.
8. Yifu Sun and Haoming Jiang.Contextual Text Denoising with MaskedLanguage Models. 2019. arXiv:1910.14080 [cs.CL].
9. Thomas Wolf et al. “Transformers: State-of-the-Art Natural LanguageProcessing”. In:Proceedings of the 2020 Conference on Empirical Methodsin Natural Language Processing: System Demonstrations. Online: Associ-ation for Computational Linguistics, Oct. 2020, pp. 38–45. url:https://www.aclweb.org/anthology/2020.emnlp-demos.6.