U

��Z^V�@s&dd
lZddl
mZ
Gdd�d�Zd
S)�N)
�Iterablec@s8eZ
dZd
Zgd�
eed�
dd�Zeed�dd�ZdS)	�OcrQualityDictionaryz3Manages a dictionary for simple OCR quality checks.)
�wordlistc

Cs"t�|_
|j
�d
d�|
D�
�

dS)z�Construct a dictionary from a list of words.

        Words for which capitalization is important should be capitalized in the
        dictionary. Words that contain spaces or other punctuation will never match.
        c
ss|]
}
|
V
qdS)
N���.0�
wrr�2/usr/lib/python3/dist-packages/ocrmypdf/quality.py�	<genexpr>"sz0OcrQualityDictionary.__init__.<locals>.<genexpr>N)�set�
dictionary�update)�selfrrrr	�__init__s
zOcrQualityDictionary.__init__)�ocr_text�returncCs�t�
d
d|
�}t�
dd|�}t�d|�}dd�|D�
}d}|D]0}||jksf||��kr>|��|jkr>|d7}q>|dkr�|t|�
}nd	}|S)
a
Check how many unique words in the OCR text match a dictionary.

        Words with mixed capitalized are only considered a match if the test word
        matches that capitalization.

        Returns:
            number of words that match / number
        z[0-9_]+�
 z\W+z\s+c
Ssh|]}
t|
�
dkr|
�qS)
�)
�lenrrrr	�	<setcomp>0sz=OcrQualityDictionary.measure_words_matched.<locals>.<setcomp>r
�
g)�re�sub�splitr�lowerr)rr�textZtext_words_listZ
text_wordsZmatchesrZ	hit_ratiorrr	�measure_words_matched$s 	







�
�



z*OcrQualityDictionary.measure_words_matchedN)	�__name__�
__module__�__qualname__�__doc__r�strr�floatrrrrr	rs
	r)r�typingrrrrrr	�<module>s