示例#1
0
 def __init__(self, rsrcmgr, outfp, codec='utf-8', pageno=1, laparams=None,
              showpageno=False, imagewriter=None):
     TextConverter.__init__(self, rsrcmgr, outfp, codec=codec, pageno=pageno, laparams=laparams)
     self.showpageno = showpageno
     self.imagewriter = imagewriter
     self.text = ''
     return
 def __init__(self, rsrcmgr, codec='cp1251', pageno=1, laparams=None):
     dummy = io.StringIO()
     TextConverter.__init__(self, rsrcmgr, outfp=dummy, codec=codec, pageno=pageno, laparams=laparams)
     self.annots = set()
     # self.sentences = list()
     self.current_sentence = ''
     self.current_sentence_should_be_added_to_these_annotations_on_end = set()
示例#3
0
 def __init__(self, rsrcmgr, codec='utf-8', pageno=1, laparams=None):
     dummy = io.StringIO()
     TextConverter.__init__(self,
                            rsrcmgr,
                            outfp=dummy,
                            codec=codec,
                            pageno=pageno,
                            laparams=laparams)
     self.annots = set()
示例#4
0
 def __init__(self, rsrcmgr, pageno=1, laparams=None, showpageno=False):
     TextConverter.__init__(self,
                            rsrcmgr,
                            None,
                            codec=None,
                            pageno=pageno,
                            laparams=laparams,
                            showpageno=showpageno)
     self.__text = []
示例#5
0
 def __init__(self,
              rsrcmgr,
              outfp,
              codec='utf-8',
              pageno=1,
              laparams=None,
              showpageno=False,
              imagewriter=None):
     TextConverter.__init__(self,
                            rsrcmgr,
                            outfp,
                            codec=codec,
                            pageno=pageno,
                            laparams=laparams,
                            showpageno=showpageno,
                            imagewriter=imagewriter)
     return
示例#6
0
 def __init__(self,
              rsrcmgr,
              outfp,
              codec='utf-8',
              pageno=1,
              laparams=None,
              showpageno=False,
              imagewriter=None):
     TextConverter.__init__(self,
                            rsrcmgr,
                            outfp,
                            codec=codec,
                            pageno=pageno,
                            laparams=laparams)
     self.showpageno = showpageno
     self.imagewriter = imagewriter
     self.textlines = {}  # per pagenumber
     self.pagenumber = 0
     return
示例#7
0
 def __init__(self, *args, **kwargs):
     TextConverter.__init__(self, *args, **kwargs)
 def __init__(self, *args, **kwargs):
     TextConverter.__init__(self, *args, **kwargs)
     pass
示例#9
0
文件: pdf2text.py 项目: zaim/bukutip
 def __init__(self, *args, **kwargs):
     self.results = []
     TextConverter.__init__(self, *args, **kwargs)
示例#10
0
 def __init__(self, rsrcmgr, pageno=1, laparams=None, showpageno=False):
     TextConverter.__init__(self, rsrcmgr, None, codec=None, pageno=pageno, laparams=laparams,
                            showpageno=showpageno)
     self.__text = []
示例#11
0
 def __init__(self, separator=',', threshold=1.5, *args, **kwargs):
     TextConverter.__init__(self, *args, **kwargs)
     self.separator = separator
     self.threshold = threshold
示例#12
0
 def __init__(self, rsrcmgr, outfp, laparams=laparams):
     TextConverter.__init__(self, rsrcmgr, outfp, laparams=laparams)
     return
示例#13
0
文件: f.py 项目: kamekame/alpha
 def __init__(self, *args, **kwargs):
     TextConverter.__init__(self, *args, **kwargs)
     self.separator = separator
     self.threshold = threshold
示例#14
0
 def __init__(self, *args, **kwargs):
     TextConverter.__init__(self, *args, **kwargs)
     self.pages = {}