python - googleTrans を使用した自動翻訳ツール

webベースの自動翻訳ツールにおいてコピペの手間を省く

• 論文翻訳時にpdfから 日本訳pdfを直接生成するツール を作成

• 現在版は、 Google翻訳 を使用．

  • エンジン部分でDeeplのpython APIを叩けば、Deeplの使用が同様に可能．

  • そのうち拡張予定．

制約

• ネットへの接続が必要．

• 無料版では、 500000字 までの制限があるらしい．

  • ただし、今のところ、公共無線LANから接続している(IPが都度変わる)からか、制約をくらったことはない．

使用方法 (1例)

• pathが通る場所 ( e.g. /usr/local/bin/ ) に translator__usingGoogleTrans.py を置いておく.

• importする他のコード( nkTextProcess.convert__text2pdf.py, nkTextProcess.extract__textFromPDF.py ) はpython path の通ったところにおいておく( nkTextProcess )、もしくは、同一ファイル内に入れておく．

• 実行権限を付与しておく

  • $ sudo chmod +x translator__usingGoogleTrans.py をしておく

  • shebang ( = pythonファイルの先頭のおまじない) はすでに設定．適宜変更．

実行コマンド

$ translator__usingGoogleTrans.py --input_pdf pdf/sample.pdf

$ translator__usingGoogleTrans.py --input_pdf pdf/sample.pdf --output_pdf pdf/output.pdf --english_text text_en.txt --japanese_text text_ja.txt --intermediate True --fontsize 8.0 --show True

コード

github

$ git clone https://github.com/wfw-pgr/nkTextProcess.git

translator__usingGoogleTrans.py ( スタンドアロン：ライブラリ同梱バージョン )

#!/usr/bin/env python3
import os, sys
import pypdf
import googletrans

from reportlab.pdfbase            import pdfmetrics, cidfonts
from reportlab.lib.styles         import getSampleStyleSheet, ParagraphStyle
from reportlab.platypus           import SimpleDocTemplate, Paragraph, PageBreak
from reportlab.platypus.flowables import Spacer
from reportlab.lib.pagesizes      import A4, mm, portrait

# ========================================================= #
# ===  extract__textFromPDF.py                          === #
# ========================================================= #

def extract__textFromPDF( inpFile=None, outFile=None, silent=True, \
                          remove_return=True, returnType="list" ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( inpFile is None ): sys.exit( "[extract__textFromPDF.py] inpFile == ???" )

    # ------------------------------------------------- #
    # --- [2] read pdf file                         --- #
    # ------------------------------------------------- #
    text_stack = []
    with open( inpFile, "rb" ) as f:
        # -- [2-1] open file                        --  #
        reader = pypdf.PdfReader( f )
        nPages = len( reader.pages )
        # -- [2-2] convert file into text           --  #
        for ik,apage in enumerate( reader.pages ):
            atext       = apage.extract_text()
            if ( remove_return ): atext = atext.replace( "\n", " " )
            text_stack += [ atext ]
        plaintext = "\n".join( text_stack )
    
    # ------------------------------------------------- #
    # --- [4] display / save in a file              --- #
    # ------------------------------------------------- #
    if ( not( silent ) ):
        for ik, atext in enumerate( text_stack ):
            print( "-"*70  )
            print( "---" + " page == {}".format( ik+1 ) )
            print( "-"*70 )
            print( atext )
            print()
    if ( outFile is not None ):
        with open( outFile, "w" ) as f:
            f.write( plaintext )

    # ------------------------------------------------- #
    # --- [5] return                                --- #
    # ------------------------------------------------- #
    if   ( returnType == "list" ):
        return( text_stack )
    elif ( returnTYype == "str" ):
        return( plaintext  )


# ========================================================= #
# ===  convert__text2pdf.py                             === #
# ========================================================= #

def convert__text2pdf( outFile=None, texts=None, fontsize=9.0, leading=None, \
                       output_by_page=True ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( outFile is None ): sys.exit( "[convert__text2pdf.py] outFile == ???" )
    if ( texts   is None ): sys.exit( "[convert__text2pdf.py] texts   == ???" )
    if ( leading is None ): leading = 2.0 * fontsize
    
    # ------------------------------------------------- #
    # --- [2] preparation of style                  --- #
    # ------------------------------------------------- #
    margins  = [ 15.0*mm, 15.0*mm, 15.0*mm, 15.0*mm ]
    doc      = SimpleDocTemplate( outFile, pagesize=portrait(A4),\
                                  leftMargin   =margins[0],
                                  bottomMargin =margins[1],
                                  rightMargin  =margins[2],
                                  topMargin    =margins[3] )
    pdfmetrics.registerFont( cidfonts.UnicodeCIDFont( "HeiseiMin-W3" ) )
    style_dict ={
        "name":"Normal",
        "fontName":"HeiseiMin-W3",
        "fontSize":fontsize,
        'borderWidth':0,
        'borderColor':None,
        "leading":leading,              # -- space between lines :: gyokan -- #
        "firstLineIndent":fontsize*1.0, # -- indent  -- #
    }
    style    = ParagraphStyle( **style_dict )
    Story    = []
    Story   += [ Spacer( width=1.0, height=5.0*mm ) ]

    # ------------------------------------------------- #
    # --- [3] pack texts                            --- #
    # ------------------------------------------------- #
    for ik,atext in enumerate( texts ):
        p = Paragraph( atext, style )
        Story += [ p ]
        if ( output_by_page ):
            Story += [ PageBreak() ]
        else:
            Story += [ Spacer( width=2*mm, height=6.0*mm ) ]

    # ------------------------------------------------- #
    # --- [4] build and return                      --- #
    # ------------------------------------------------- #
    doc.build( Story )
    print( "[convert__text2pdf.py] output file :: {} ".format( outFile ) )
    return()


# ========================================================= #
# ===  translator__usingGoogleTrans.py                  === #
# ========================================================= #

def translator__usingGoogleTrans( input_pdfFile=None, output_pdfFile=None, \
                                  english_txtFile=None, japanese_txtFile=None, \
                                  fontsize=9.0, silent=True ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( input_pdfFile  is None ): sys.exit("[translator__usingGoogleTrans.py] input_pdfFile  == ???")
    if ( output_pdfFile is None ): output_pdfFile = input_pdfFile.replace( ".pdf", "_ja.pdf" )

    # ------------------------------------------------- #
    # --- [2] extract__textFromPDF                  --- #
    # ------------------------------------------------- #
    text_en = extract__textFromPDF( inpFile=input_pdfFile, outFile=english_txtFile, \
                                    remove_return=True )
    
    # ------------------------------------------------- #
    # --- [3] translator into japanese              --- #
    # ------------------------------------------------- #
    tr         = googletrans.Translator()
    text_stack = []
    for ik,apage in enumerate( text_en ):
        text_piece = ( tr.translate( apage, dest="ja", src="en" ) ).text
        text_stack+= [ text_piece ]
    text_ja = "\n\n".join( text_stack )

    # ------------------------------------------------- #
    # --- [4] save in a file                        --- #
    # ------------------------------------------------- #
    if ( not( silent ) ):
        print( "\n" + "-"*70 +"\n"  )
        print( text_ja )
        print( "\n" + "-"*70 +"\n"  )
        
    if ( japanese_txtFile is not None ):
        with open( japanese_txtFile, "w" ) as f:
            f.write( text_ja )

    # ------------------------------------------------- #
    # --- [5] convert into japanese pdf             --- #
    # ------------------------------------------------- #
    convert__text2pdf( outFile=output_pdfFile, texts=text_stack, fontsize=fontsize )
    
    # ------------------------------------------------- #
    # --- [6] return                                --- #
    # ------------------------------------------------- #
    return( text_ja )


# ========================================================= #
# ===   Execution of Pragram                            === #
# ========================================================= #

if ( __name__=="__main__" ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    import argparse
    parser = argparse.ArgumentParser()
    parser.add_argument( "--input_pdf"     , default=None, help="input pdf file"     )
    parser.add_argument( "--output_pdf"    , default=None, help="output pdf file"    )
    parser.add_argument( "--english_text"  , default=None, help="english_text_file"  )
    parser.add_argument( "--japanese_text" , default=None, help="japanese_text_file" )
    parser.add_argument( "--fontsize"      , type=float, default=9.0, help="font size"        )
    parser.add_argument( "--show"          , type=bool , default=False, help="display or not" )
    parser.add_argument( "--intermediate"  , type=bool , default=False, help="intermidiate file out" )
    
    args   = parser.parse_args()

    if ( not( args.input_pdf ) ):
        print( "[ How to use ] python translator__usingGoogleTrans.py --input_pdf xxx.pdf " )
        sys.exit()
    else:
        input_pdfFile = str( args.input_pdf )
    if ( args.intermediate ):
        if ( args.english_text  is None ): args.english_text  = "text_en.txt"
        if ( args.japanese_text is None ): args.japanese_text = "text_ja.txt"

    # ------------------------------------------------- #
    # --- [2] call translator                       --- #
    # ------------------------------------------------- #
    print( "[translator__usingGoogleTrans.py] translation of {}".format( args.input_pdf ) )
    translator__usingGoogleTrans( input_pdfFile=args.input_pdf, output_pdfFile=args.output_pdf, \
                                  english_txtFile=args.english_text, \
                                  japanese_txtFile=args.japanese_text,\
                                  fontsize=args.fontsize, silent=not( args.show ) )

nkTextProcess 参照ver.

translator__usingGoogleTrans.py

#!/usr/bin/env python3
import os, sys
import googletrans
import nkTextProcess.extract__textFromPDF as etp
import nkTextProcess.convert__text2pdf    as t2p

# ========================================================= #
# ===  translator__usingGoogleTrans.py                  === #
# ========================================================= #

def translator__usingGoogleTrans( input_pdfFile=None, output_pdfFile=None, \
                                  english_txtFile=None, japanese_txtFile=None, \
                                  fontsize=9.0, silent=True ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( input_pdfFile  is None ): sys.exit("[translator__usingGoogleTrans.py] input_pdfFile  == ???")
    if ( output_pdfFile is None ): output_pdfFile = input_pdfFile.replace( ".pdf", "_ja.pdf" )

    # ------------------------------------------------- #
    # --- [2] extract__textFromPDF                  --- #
    # ------------------------------------------------- #
    text_en = etp.extract__textFromPDF( inpFile=input_pdfFile, outFile=english_txtFile, \
                                        remove_return=True )
    
    # ------------------------------------------------- #
    # --- [3] translator into japanese              --- #
    # ------------------------------------------------- #
    tr         = googletrans.Translator()
    text_stack = []
    for ik,apage in enumerate( text_en ):
        text_piece = ( tr.translate( apage, dest="ja", src="en" ) ).text
        text_stack+= [ text_piece ]
    text_ja = "\n\n".join( text_stack )

    # ------------------------------------------------- #
    # --- [4] save in a file                        --- #
    # ------------------------------------------------- #
    if ( not( silent ) ):
        print( "\n" + "-"*70 +"\n"  )
        print( text_ja )
        print( "\n" + "-"*70 +"\n"  )
        
    if ( japanese_txtFile is not None ):
        with open( japanese_txtFile, "w" ) as f:
            f.write( text_ja )

    # ------------------------------------------------- #
    # --- [5] convert into japanese pdf             --- #
    # ------------------------------------------------- #
    t2p.convert__text2pdf( outFile=output_pdfFile, texts=text_stack, fontsize=fontsize )
    
    # ------------------------------------------------- #
    # --- [6] return                                --- #
    # ------------------------------------------------- #
    return( text_ja )


# ========================================================= #
# ===   Execution of Pragram                            === #
# ========================================================= #

if ( __name__=="__main__" ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    import argparse
    parser = argparse.ArgumentParser()
    parser.add_argument( "--input_pdf"     , default=None, help="input pdf file"     )
    parser.add_argument( "--output_pdf"    , default=None, help="output pdf file"    )
    parser.add_argument( "--english_text"  , default=None, help="english_text_file"  )
    parser.add_argument( "--japanese_text" , default=None, help="japanese_text_file" )
    parser.add_argument( "--fontsize"      , type=float, default=9.0, help="font size"        )
    parser.add_argument( "--show"          , type=bool , default=False, help="display or not" )
    parser.add_argument( "--intermediate"  , type=bool , default=False, help="intermidiate file out" )
    
    args   = parser.parse_args()

    if ( not( args.input_pdf ) ):
        print( "[ How to use ] python translator__usingGoogleTrans.py --input_pdf xxx.pdf " )
        sys.exit()
    else:
        input_pdfFile = str( args.input_pdf )
    if ( args.intermediate ):
        if ( args.english_text  is None ): args.english_text  = "text_en.txt"
        if ( args.japanese_text is None ): args.japanese_text = "text_ja.txt"

    # ------------------------------------------------- #
    # --- [2] call translator                       --- #
    # ------------------------------------------------- #
    print( "[translator__usingGoogleTrans.py] translation of {}".format( args.input_pdf ) )
    translator__usingGoogleTrans( input_pdfFile=args.input_pdf, output_pdfFile=args.output_pdf, \
                                  english_txtFile=args.english_text, \
                                  japanese_txtFile=args.japanese_text,\
                                  fontsize=args.fontsize, silent=not( args.show ) )

convert__text2pdf.py

import os, sys
from reportlab.pdfbase            import pdfmetrics, cidfonts
from reportlab.lib.styles         import getSampleStyleSheet, ParagraphStyle
from reportlab.platypus           import SimpleDocTemplate, Paragraph, PageBreak
from reportlab.platypus.flowables import Spacer
from reportlab.lib.pagesizes      import A4, mm, portrait

# ========================================================= #
# ===  convert__text2pdf.py                             === #
# ========================================================= #

def convert__text2pdf( outFile=None, texts=None, fontsize=9.0, leading=None, \
                       output_by_page=True ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( outFile is None ): sys.exit( "[convert__text2pdf.py] outFile == ???" )
    if ( texts   is None ): sys.exit( "[convert__text2pdf.py] texts   == ???" )
    if ( leading is None ): leading = 2.0 * fontsize
    
    # ------------------------------------------------- #
    # --- [2] preparation of style                  --- #
    # ------------------------------------------------- #
    margins  = [ 15.0*mm, 15.0*mm, 15.0*mm, 15.0*mm ]
    doc      = SimpleDocTemplate( outFile, pagesize=portrait(A4),\
                                  leftMargin   =margins[0],
                                  bottomMargin =margins[1],
                                  rightMargin  =margins[2],
                                  topMargin    =margins[3] )
    pdfmetrics.registerFont( cidfonts.UnicodeCIDFont( "HeiseiMin-W3" ) )
    style_dict ={
        "name":"Normal",
        "fontName":"HeiseiMin-W3",
        "fontSize":fontsize,
        'borderWidth':0,
        'borderColor':None,
        "leading":leading,              # -- space between lines :: gyokan -- #
        "firstLineIndent":fontsize*1.0, # -- indent  -- #
    }
    style    = ParagraphStyle( **style_dict )
    Story    = []
    Story   += [ Spacer( width=1.0, height=5.0*mm ) ]

    # ------------------------------------------------- #
    # --- [3] pack texts                            --- #
    # ------------------------------------------------- #
    for ik,atext in enumerate( texts ):
        p = Paragraph( atext, style )
        Story += [ p ]
        if ( output_by_page ):
            Story += [ PageBreak() ]
        else:
            Story += [ Spacer( width=2*mm, height=6.0*mm ) ]

    # ------------------------------------------------- #
    # --- [4] build and return                      --- #
    # ------------------------------------------------- #
    doc.build( Story )
    print( "[convert__text2pdf.py] output file :: {} ".format( outFile ) )
    return()


# ========================================================= #
# ===   Execution of Pragram                            === #
# ========================================================= #

if ( __name__=="__main__" ):

    outFile = "pdf/sample.pdf"
    texts   = [ 60*"これはサンプルです.({})".format( ik ) for ik in range(6) ]
    
    convert__text2pdf( outFile=outFile, texts=texts )
    

extract__textFromPDF.py

import os, sys
import pypdf

# ========================================================= #
# ===  extract__textFromPDF.py                          === #
# ========================================================= #

def extract__textFromPDF( inpFile=None, outFile=None, silent=True, \
                          remove_return=True, returnType="list" ):

    # ------------------------------------------------- #
    # --- [1] arguments                             --- #
    # ------------------------------------------------- #
    if ( inpFile is None ): sys.exit( "[ex1.py] inpFile == ???" )

    # ------------------------------------------------- #
    # --- [2] read pdf file                         --- #
    # ------------------------------------------------- #
    text_stack = []
    with open( inpFile, "rb" ) as f:
        # -- [2-1] open file                        --  #
        reader = pypdf.PdfReader( f )
        nPages = len( reader.pages )
        # -- [2-2] convert file into text           --  #
        for ik,apage in enumerate( reader.pages ):
            atext       = apage.extract_text()
            if ( remove_return ): atext = atext.replace( "\n", " " )
            text_stack += [ atext ]
        plaintext = "\n".join( text_stack )
    
    # ------------------------------------------------- #
    # --- [4] display / save in a file              --- #
    # ------------------------------------------------- #
    if ( not( silent ) ):
        for ik, atext in enumerate( text_stack ):
            print( "-"*70  )
            print( "---" + " page == {}".format( ik+1 ) )
            print( "-"*70 )
            print( atext )
            print()
    if ( outFile is not None ):
        with open( outFile, "w" ) as f:
            f.write( plaintext )

    # ------------------------------------------------- #
    # --- [5] return                                --- #
    # ------------------------------------------------- #
    if   ( returnType == "list" ):
        return( text_stack )
    elif ( returnTYype == "str" ):
        return( plaintext  )


# ========================================================= #
# ===   Execution of Pragram                            === #
# ========================================================= #

if ( __name__=="__main__" ):
    inpFile = "pdf/sample.pdf"
    outFile = "dat/sample.txt"
    ret = extract__textFromPDF( inpFile=inpFile, outFile=outFile )
    print( ret )