
    i                     x    d Z ddlZddlmZ ddlZ ej
                  e      Zdeez  de	ee
f   fdZdedefdZy)	u.   PDF 텍스트 추출 및 언어 감지 모듈.    N)Pathpdf_pathreturnc                    t        j                  t        |             }g }|D ]'  }|j                         xs d}|j	                  |       ) dj                  |      }t        |      }|j                          |j                  dd      }||fS )z:Extract full text and page count from a PDF using PyMuPDF. 
 )	pymupdfopenstrget_textappendjoinlenclosereplace)r   docpagespagetextfulltext
page_counts          L/Users/manta/Documents/Projects/TheRoad-I1/backend/app/modules/pdf_parser.pyextract_text_from_pdfr      s~    
,,s8}
%CE}}$"T  yyHSJIIK+HZ    r   c                 n    | dd }t        d |D              }t        d |D              }|dkD  ry|dkD  ryy)	z4Simple language detection based on character ranges.Ni  c              3   >   K   | ]  }d |cxk  rdk  sn nd  yw)u   가u   힣   N .0cs     r   	<genexpr>z"detect_language.<locals>.<genexpr>   s     BfA(A(A1fs   
c              3   X   K   | ]"  }d |cxk  rdk  sn d|cxk  rdk  sn nd $ yw)u   ぀u   ゟu   ゠u   ヿr   Nr   r    s     r   r#   z"detect_language.<locals>.<genexpr>   s+     _fA(A(AXQRE^V^E^1fs   *
*2   kojaen)sum)r   sampleko_countja_counts       r   detect_languager-      sD    %4[FBfBBH_f__H"}"}r   )__doc__loggingpathlibr   r
   	getLogger__name__loggerr   tupleintr   r-   r   r   r   <module>r6      sU    4   			8	$ C$J  5c?  	# 	# 	r   