aoepub3_hack.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

# AozoraEpub3 + kindlegen で作成した mobi テキストで、部分的に辞書が
# 引けなくなる時の対策スクリプト。
#
# epub の 本文 xhtml (0001.xhtmlなど)について、句読点の後に、おそらく
# 無害であろうと思われる、空の <span /> タグを挿入している。
# 但し、句読点の後が、ascii文字(white-space, <ruby>タグなど)だった場合は、
# <span/>の挿入は行わない。

_version="0.9.1"
import shutil, tempfile, zipfile, os, sys, re
import StringIO

SENTENCE_END =(
  (u"。",),  # 通常
  (u"。", u"、")  # 酷い場合…
  )

# epub 内部の本文データのファイル名パターン
AOZORAEPUB3_BODYFILE = r'''OPS/xhtml/[\d]+\.xhtml$'''

#BUNRI_KINSOKU = u"—…‥〳〴〵"
BUNRI_KINSOKU = u"—…‥〳〴〵○"

#OWARI_KAKKO = u"」』）］｝〕〉》】〙〗｠" 
OWARI_KAKKO = u"」』）］｝〕〉》】〙〗｠〟’”"

GYOUTOU_KINSOKU = u"、。" + OWARI_KAKKO

#_isdebug = True
_isdebug = False

def expand_path(path):
  return os.path.expandvars( os.path.expanduser(path) )

def U(s):
  if isinstance(s, str): return s.decode("utf-8", "replace")
  return unicode(s)

def oprif(port, fmt, *args):
  if not args: port.write(U(fmt).encode("utf-8", "replace"))
  else:
    ustr = U(fmt) % tuple(
      U(a) if isinstance(a, str) else a for a in args)
    port.write(ustr.encode("utf-8", "replace"))
    
def prif(fmt, *args): oprif(sys.stdout, fmt, *args)
def eprif(fmt, *args): oprif(sys.stderr, fmt, *args)

def isascii(uc): return ord(uc) < 128

# for zipfile.ZipFile() in python2.6
class FileCtx(object):
  def __init__(self, fobj): self.fobj = fobj
  def __enter__(self): return self.fobj
  def __exit__(self, exc_type, exc_value, traceback):
    if not (self.fobj == sys.stdin):
      # print "  %s will close" % self.fobj # for debug
      self.fobj.close()
      # print u"  %s did close" % self.fobj # for debug
    return False # throw exception

class TempfileCtx(object):
  def __init__(self):
    (fd, self.tmpfile) = tempfile.mkstemp(".xhtml", "_aoepub3hack_")
    os.close(fd) # Fix me
    if _isdebug: prif("tmpfile = %s\n", self.tmpfile)
  def __enter__(self): return self.tmpfile
  def __exit__(self, exc_type, exc_value, traceback):
    os.unlink(self.tmpfile)
    return False # throw exception

class TempdirCtx(object):
  def __init__(self):
    self.tmpdir  = tempfile.mkdtemp(".tmp", "_aoepub3hack_")
    if _isdebug: sys.stdout.write("tmpdir = %s\n" % self.tmpdir)
  def __enter__(self): return self.tmpdir
  def __exit__(self, exc_type, exc_value, traceback):
    #print "delete", self.tmpdir
    # Memo:  in Mac, find /var/folders -type d -iname '_zipdir*'
    shutil.rmtree(self.tmpdir)
    return False # throw exception


def out(uline, port=sys.stdout):
  try:
    sline = uline.encode("utf-8", "strict")
  except UnicodeError:
    sline = uline.encode("utf-8", "replace")
    eprif("WARNING: fail to unicode encoding\n")
    eprif("    %s\n", uline.rstrip())
    eprif("  ->%s\n", sline.decode("utf-8", "replace").rstrip())
  port.write(sline)

def out_each_sentence(line, port=sys.stdout, sep=SENTENCE_END[0] ):
  class TextBuffer(object):
    def __init__(self):
      self.buf = None
    def putbuf(self, s):
      if self.buf is None: self.buf = StringIO.StringIO(u"")
      self.buf.write(s)
    def flushbuf(self):
      if self.buf is not None:
        out( self.buf.getvalue(), port )
        self.buf.close()
        self.buf = None
  try: uline = line.decode("utf-8", "strict")
  except UnicodeError:
    uline = line.decode("utf-8", "replace")
    eprif("WARNING: fial to unicode decoding\n")
    eprif("  %s\n", uline.rstrip())
  
  tbuf = TextBuffer()
  intag = False
  will_span = False
  for s in uline:
    if will_span:
      # 句読点の次の文字が、ascii文字('<' や white-space) でなければ
      # <span /> を書く
      if isascii(s) or (s in GYOUTOU_KINSOKU) or (s in BUNRI_KINSOKU): pass
      else:
        out(u"<span />", port)
      will_span = False
        
    tbuf.putbuf(s)
    if intag:
      if s == ">": intag = False
    else:
      if s == "<": intag = True
      else:
        if s in sep:
          # output sentence
          tbuf.flushbuf()
          will_span = True
          #out(u"<span />", port)
  tbuf.flushbuf()

def main(srcname, dstname, sep=SENTENCE_END[0]):
  with TempdirCtx() as tmpdir:
    with FileCtx(zipfile.ZipFile(srcname, "r")) as inzp:
      infolist = inzp.infolist()
      inzp.extractall(tmpdir)
  
      with FileCtx(zipfile.ZipFile(U(dstname).encode("utf-8", "replace"),
                                   "w", zipfile.ZIP_STORED)) as outzp:
        for zinfo in infolist:
          compression = zinfo.compress_type
          if compression not in (zipfile.ZIP_DEFLATED, zipfile.ZIP_STORED):
            compression = zipfile.ZIP_DEFLATED
          arcname = zinfo.filename
          srcpath = os.path.join(tmpdir, arcname)
          if not os.path.isfile(srcpath):
            eprif("ERROR: BUG?,  No such a file - %s\n", srcpath)
          else:
            if re.search(AOZORAEPUB3_BODYFILE, arcname, re.I):
              # 文の終わりで改行させる。
              with TempfileCtx() as convname:
                with open(convname, "w") as outport:
                  with open(srcpath, "rb") as inport:
                    inbody = False
                    rxbody = re.compile(r'''<body\s*>|<body\s[^<>]*>''', re.I)
                    for line in inport:
                      if not inbody:
                        if rxbody.search(line): inbody = True
                      if inbody:
                        out_each_sentence(line, outport, sep)
                      else:
                        outport.write(line)
                outzp.write(convname, arcname, compression)
            else:
              outzp.write(srcpath, arcname, compression)

if __name__ == '__main__':

  def usage():
    name = os.path.basename(sys.argv[0])
    sys.stderr.write( (u'''\
Usage:  %s {OPTIONS} input_epub3 output_epub3
  input_epub3 --- AozoraEpub3 が出力する epub3 ファイル名
  output_epub3 -- 変換後の epub3 ファイル名
  OPTIONS:
    -s 0|1  --- 0 ならば、句点(。) の後に <span /> を加える 
                1 ならば、句点(。) 及び、読点(、) の後に <span /> を加える 
                (デフォルトは、1)
    -V      --- バージョン番号の表示
    -h      --- Help表示
''' % name).encode("utf-8", "replace"))
    exit(1)

  (srcname, dstname, sep) = (None, None, SENTENCE_END[1])
  arglist = sys.argv[1:]
  while arglist:
    arg = arglist.pop(0)
    if arg.startswith("-"):
      if arg.startswith("-h"):
        usage()
      elif arg == "-s":
        try:
          sep = SENTENCE_END[ int( arglist.pop(0) ) ]
        except (ValueError, IndexError):
          usage()
      elif arg == "-V":
        prif("version %s\n", _version)
        exit(0)
      else: usage()
    else:
      if not srcname:
        srcname = expand_path( arg )
        if not os.path.isfile(srcname):
          eprif("No such a file - %s\n", srcname)
          exit(1)
      elif not dstname :
        dstname = expand_path( arg )
      else:
        usage()
  if srcname and dstname:
    #prif("src=%s dst=%s sep=%s\n", srcname, dstname, sep) # debug
    main(srcname, dstname, sep)
  else:
    usage()