OCRfixr 1.1.0

A contextual spellchecker for OCR output

OCRfixr

OVERVIEW

This project aims to automate the boring work of manually correcting OCR output from Distributed Proofreaders' book digitization projects

1) CORRECTING MISREADS

OCRs can sometimes mistake similar-looking characters when scanning a book. For example, "l" and "1" are easily confused, potentially causing the OCR to misread the word "learn" as "1earn".

As written in book:

"The birds flevv south"

Corrected text:

"The birds flew south"

How OCRfixr Works:

OCRfixr fixes misreads by checking 1) possible spell corrections against the 2) local context of the word. For example, here's how OCRfixr would evaluate the following OCR mistake:

As written in book:

"Days there were when small trade came to the stoie. Then the young clerk read."

Method	Plausible Replacements
Spellcheck (TextBlob)	stone, store, stoke, stove, stowe, stole, soie
Context (BERT)	market, shop, town, city, store, table, village, door, light, markets, surface, place, window, docks, area

Since there is match for both a plausible spellcheck replacement and that word reasonably matches the context of the sentence, OCRfixr updates the word.

Corrected text:

"Days there were when small trade came to the store. Then the young clerk read."

Using OCRfixr

The package can be installed using pip.

pip install OCRfixr

By default, OCRfixr only returns the original string, with all changes incorporated:

>>> from ocrfixr import spellcheck

>>> text = "The birds flevv south"
>>> spellcheck(text).fix()
'The birds flew south'

Use return_fixes to also include all corrections made to the text:

>>> spellcheck(text, return_fixes = "T").fix()
['The birds flew south', {'flevv': 'flew'}]

Use full_results_by_paragraph for longer texts, to break out the text & associated changes by paragraph:

>>> text = "The birds flevv down\n south, but wefe quickly apprehended\n by border patrol agents"
>>> spellcheck(text, full_results_by_paragraph = "T").fix()
[['The birds flew down\n', {'flevv': 'flew'}],
 [' south, but were quickly apprehended\n', {'wefe': 'were'}],
 [' by border patrol agents', {}]]

Otherwise, the full text (+ any changes) will be returned in a single object:

>>> text = "The birds flevv down\n south, but wefe quickly apprehended\n by border patrol agents"
>>> spellcheck(text, return_fixes = "T").fix()
['The birds flew down\n south, but were quickly apprehended\n by border patrol agents',
 {'flevv': 'flew', 'wefe': 'were'}]

(Note: OCRfixr resets its BERT context window at the start of each new paragraph, so splitting by paragraph may be a useful debug feature)

Avoiding "Damn You, Autocorrect!"

By design, OCRfixr is change-averse:

• If spellcheck/context do not line up, no update is made.
• Likewise, if there is >1 word that lines up for spellcheck/context, no update is made.
• Only the top 15 context suggestions are considered, to limit low-probability matches.
• Proper nouns (anything starting with a capital letter) are not evaluated for spelling.

Word context is drawn from all sentences in the current paragraph, to maximize available information, while also not bogging down the BERT model.

2) UNSPLITTING WORDS

Sometimes, books split words across lines with a hyphen. These need to be correctly pieced back together into a single word before a new line is started.

To glue words back together that are split across lines, OCRfixr checks the hyphenated word against the same word list to see if the newly reunited halves created a valid word, along with a few other rules. OCRfixr then decides if the split word needs that hyphen or not.

By default, OCRfixr only returns the original string, with all changes incorporated:

>>> from ocrfixr import unsplit

>>> text = "He saw the red pirogue, adrift and afire in the mid-\ndle of the river!"
>>> print(text)
'He saw the red pirogue, adrift and afire in the mid-
dle of the river!'

>>> print(unsplit(text).fix())
'He saw the red pirogue, adrift and afire in the middle
of the river!

As before, use return_fixes to also include all corrections made to the text.

This method does not use full_results_by_paragraph, as unsplit is rules-based rather than context-specific. There is no paragraph-specific context window.

Credits

• TextBlob powers spellcheck suggestions
• transformers does the heavy lifting for BERT context modelling.
• SCOWL word list is Copyright 2000-2019 by Kevin Atkinson.
• All book data comes from Distributed Proofreaders. Support them here: https://www.pgdp.net/c/