Ë jË5j´ãó<—dZddlmZddlZddlmZGd„d«Zy)z!Utilities to measure OCR quality.é)Ú annotationsN)ÚIterablecó —eZdZdZdd„Zdd„Zy)ÚOcrQualityDictionaryz3Manages a dictionary for simple OCR quality checks.có$—t|«|_y)zÜConstruct a dictionary from a list of words. Words for which capitalization is important should be capitalized in the dictionary. Words that contain spaces or other punctuation will never match. N)ÚsetÚ dictionary)ÚselfÚwordlists úF/var/www/html/qr/venv/lib/python3.12/site-packages/ocrmypdf/quality.pyÚ__init__zOcrQualityDictionary.__init__s€ô ˜h›-ˆócóœ—tjdd|«}tjdd|«}tjd|«}|Dchc]}t|«dk\sŒ|’Œ}}d}|D]F}||jvs1||j «k7sŒ%|j «|jvsŒB|dz }ŒH|dkDr|t|«z }|Sd}|Scc}w) aCheck how many unique words in the OCR text match a dictionary. Words with mixed capitalized are only considered a match if the test word matches that capitalization. Returns: number of words that match / number z[0-9_]+ú z\W+z\s+érég)ÚreÚsubÚsplitÚlenr Úlower)r Úocr_textÚtextÚtext_words_listÚwÚ text_wordsÚmatchesÚ hit_ratios r Úmeasure_words_matchedz*OcrQualityDictionary.measure_words_matchedsÍ€ôv‰vj # xÓ0ˆÜv‰vf˜c 4Ó(ˆÜŸ(™( 6¨4Ó0ˆØ!0Ö@˜A´C¸³F¸a³K’aÐ@ˆ Ð@àˆØò ˆAØD—O‘OÑ#ØQ—W‘W“Y“ 1§7¡7£9°·±Ò#?à˜1‘ ‘ð  ð 29¸1²Gœc *›oÑ-ˆ ØÐðCFˆ ØÐùòAs Á C ÁC N)r z Iterable[str])rÚstrÚreturnÚfloat)Ú__name__Ú __module__Ú __qualname__Ú__doc__r r©rr rr s„Ù=ó(ôrr)r&Ú __future__rrÚcollections.abcrrr'rr úr*sðñ(å"ã Ý$÷ ò r