\n' - sep = '' - else : - word = '\n' - sep = '' - - if num in self.dehyphen_rootid : - word = word[0:-1] - sep = '' - - parares += word + sep - - elif wtype == 'img' : - sep = '' - parares += '
' % num
- parares += sep
-
- elif wtype == 'imgsa' :
- sep = ' '
- parares += '' % num
- parares += sep
-
- elif wtype == 'svg' :
- sep = ''
- parares += '
' % num
- parares += sep
-
- if len(sep) > 0 : parares = parares[0:-1]
- if (type == 'full') or (type == 'end') :
- parares += ''
- return parares
-
-
- def buildTOCEntry(self, pdesc) :
- parares = ''
- sep =''
- tocentry = ''
- handle_links = len(self.link_id) > 0
-
- lstart = 0
-
- cnt = len(pdesc)
- for j in xrange( 0, cnt) :
-
- (wtype, num) = pdesc[j]
-
- if wtype == 'ocr' :
- word = self.ocrtext[num]
- sep = ' '
-
- if handle_links:
- link = self.link_id[num]
- if (link > 0):
- linktype = self.link_type[link-1]
- title = self.link_title[link-1]
- title = title.rstrip('. ')
- alt_title = parares[lstart:]
- alt_title = alt_title.strip()
- # now strip off the actual printed page number
- alt_title = alt_title.rstrip('01234567890ivxldIVXLD-.')
- alt_title = alt_title.rstrip('. ')
- # skip over any external links - can't have them in a books toc
- if linktype == 'external' :
- title = ''
- alt_title = ''
- linkpage = ''
- else :
- if len(self.link_page) >= link :
- ptarget = self.link_page[link-1] - 1
- linkpage = '%04d' % ptarget
- else :
- # just link to the current page
- linkpage = self.id[4:]
- if len(alt_title) >= len(title):
- title = alt_title
- if title != '' and linkpage != '':
- tocentry += title + '|' + linkpage + '\n'
- lstart = len(parares)
- if word == '_link_' : word = ''
- elif (link < 0) :
- if word == '_link_' : word = ''
-
- if word == '_lb_':
- word = ''
- sep = ''
-
- if num in self.dehyphen_rootid :
- word = word[0:-1]
- sep = ''
-
- parares += word + sep
-
- else :
- continue
-
- return tocentry
-
-
-
-
- # walk the document tree collecting the information needed
- # to build an html page using the ocrText
-
- def process(self):
-
- tocinfo = ''
- hlst = []
-
- # get the ocr text
- (pos, argres) = self.findinDoc('info.word.ocrText',0,-1)
- if argres : self.ocrtext = argres.split('|')
-
- # get information to dehyphenate the text
- self.dehyphen_rootid = self.getData('info.dehyphen.rootID',0,-1)
-
- # determine if first paragraph is continued from previous page
- (pos, self.parastems_stemid) = self.findinDoc('info.paraStems.stemID',0,-1)
- first_para_continued = (self.parastems_stemid != None)
-
- # determine if last paragraph is continued onto the next page
- (pos, self.paracont_stemid) = self.findinDoc('info.paraCont.stemID',0,-1)
- last_para_continued = (self.paracont_stemid != None)
-
- # collect link ids
- self.link_id = self.getData('info.word.link_id',0,-1)
-
- # collect link destination page numbers
- self.link_page = self.getData('info.links.page',0,-1)
-
- # collect link types (container versus external)
- (pos, argres) = self.findinDoc('info.links.type',0,-1)
- if argres : self.link_type = argres.split('|')
-
- # collect link destinations
- (pos, argres) = self.findinDoc('info.links.href',0,-1)
- if argres : self.link_href = argres.split('|')
-
- # collect link titles
- (pos, argres) = self.findinDoc('info.links.title',0,-1)
- if argres :
- self.link_title = argres.split('|')
+ raise ValueError('Invalid file format')
+ self.sections = []
+ for i in xrange(self.num_sections):
+ offset, a1,a2,a3,a4 = struct.unpack('>LBBBB', self.contents[78+i*8:78+i*8+8])
+ flags, val = a1, a2<<16|a3<<8|a4
+ self.sections.append( (offset, flags, val) )
+ def loadSection(self, section):
+ if section + 1 == self.num_sections:
+ end_off = len(self.contents)
else:
- self.link_title.append('')
-
- # get a descriptions of the starting points of the regions
- # and groups on the page
- (pagetype, pageDesc) = self.PageDescription()
- regcnt = len(pageDesc) - 1
-
- anchorSet = False
- breakSet = False
- inGroup = False
-
- # process each region on the page and convert what you can to html
-
- for j in xrange(regcnt):
-
- (etype, start) = pageDesc[j]
- (ntype, end) = pageDesc[j+1]
+ end_off = self.sections[section + 1][0]
+ off = self.sections[section][0]
+ return self.contents[off:end_off]
+
+# cleanup unicode filenames
+# borrowed from calibre from calibre/src/calibre/__init__.py
+# added in removal of control (<32) chars
+# and removal of . at start and end
+# and with some (heavily edited) code from Paul Durrant's kindlenamer.py
+def sanitizeFileName(name):
+ # substitute filename unfriendly characters
+ name = name.replace(u"<",u"[").replace(u">",u"]").replace(u" : ",u" – ").replace(u": ",u" – ").replace(u":",u"—").replace(u"/",u"_").replace(u"\\",u"_").replace(u"|",u"_").replace(u"\"",u"\'")
+ # delete control characters
+ name = u"".join(char for char in name if ord(char)>=32)
+ # white space to single space, delete leading and trailing while space
+ name = re.sub(ur"\s", u" ", name).strip()
+ # remove leading dots
+ while len(name)>0 and name[0] == u".":
+ name = name[1:]
+ # remove trailing dots (Windows doesn't like them)
+ if name.endswith(u'.'):
+ name = name[:-1]
+ return name
+
+def fixKey(key):
+ def fixByte(b):
+ return b ^ ((b ^ (b<<1) ^ (b<<2) ^ (b<<3) ^ (b<<4) ^ (b<<5) ^ (b<<6) ^ (b<<7) ^ 0x80) & 0x80)
+ return "".join([chr(fixByte(ord(a))) for a in key])
+
+def deXOR(text, sp, table):
+ r=''
+ j = sp
+ for i in xrange(len(text)):
+ r += chr(ord(table[j]) ^ ord(text[i]))
+ j = j + 1
+ if j == len(table):
+ j = 0
+ return r
+
+class EreaderProcessor(object):
+ def __init__(self, sect, user_key):
+ self.section_reader = sect.loadSection
+ data = self.section_reader(0)
+ version, = struct.unpack('>H', data[0:2])
+ self.version = version
+ logging.info('eReader file format version %s', version)
+ if version != 272 and version != 260 and version != 259:
+ raise ValueError('incorrect eReader version %d (error 1)' % version)
+ data = self.section_reader(1)
+ self.data = data
+ des = Des(fixKey(data[0:8]))
+ cookie_shuf, cookie_size = struct.unpack('>LL', des.decrypt(data[-8:]))
+ if cookie_shuf < 3 or cookie_shuf > 0x14 or cookie_size < 0xf0 or cookie_size > 0x200:
+ raise ValueError('incorrect eReader version (error 2)')
+ input = des.decrypt(data[-cookie_size:])
+ def unshuff(data, shuf):
+ r = [''] * len(data)
+ j = 0
+ for i in xrange(len(data)):
+ j = (j + shuf) % len(data)
+ r[j] = data[i]
+ assert len("".join(r)) == len(data)
+ return "".join(r)
+ r = unshuff(input[0:-8], cookie_shuf)
+
+ drm_sub_version = struct.unpack('>H', r[0:2])[0]
+ self.num_text_pages = struct.unpack('>H', r[2:4])[0] - 1
+ self.num_image_pages = struct.unpack('>H', r[26:26+2])[0]
+ self.first_image_page = struct.unpack('>H', r[24:24+2])[0]
+ # Default values
+ self.num_footnote_pages = 0
+ self.num_sidebar_pages = 0
+ self.first_footnote_page = -1
+ self.first_sidebar_page = -1
+ if self.version == 272:
+ self.num_footnote_pages = struct.unpack('>H', r[46:46+2])[0]
+ self.first_footnote_page = struct.unpack('>H', r[44:44+2])[0]
+ if (sect.bkType == "Book"):
+ self.num_sidebar_pages = struct.unpack('>H', r[38:38+2])[0]
+ self.first_sidebar_page = struct.unpack('>H', r[36:36+2])[0]
+ # self.num_bookinfo_pages = struct.unpack('>H', r[34:34+2])[0]
+ # self.first_bookinfo_page = struct.unpack('>H', r[32:32+2])[0]
+ # self.num_chapter_pages = struct.unpack('>H', r[22:22+2])[0]
+ # self.first_chapter_page = struct.unpack('>H', r[20:20+2])[0]
+ # self.num_link_pages = struct.unpack('>H', r[30:30+2])[0]
+ # self.first_link_page = struct.unpack('>H', r[28:28+2])[0]
+ # self.num_xtextsize_pages = struct.unpack('>H', r[54:54+2])[0]
+ # self.first_xtextsize_page = struct.unpack('>H', r[52:52+2])[0]
+
+ # **before** data record 1 was decrypted and unshuffled, it contained data
+ # to create an XOR table and which is used to fix footnote record 0, link records, chapter records, etc
+ self.xortable_offset = struct.unpack('>H', r[40:40+2])[0]
+ self.xortable_size = struct.unpack('>H', r[42:42+2])[0]
+ self.xortable = self.data[self.xortable_offset:self.xortable_offset + self.xortable_size]
+ else:
+ # Nothing needs to be done
+ pass
+ # self.num_bookinfo_pages = 0
+ # self.num_chapter_pages = 0
+ # self.num_link_pages = 0
+ # self.num_xtextsize_pages = 0
+ # self.first_bookinfo_page = -1
+ # self.first_chapter_page = -1
+ # self.first_link_page = -1
+ # self.first_xtextsize_page = -1
+
+ logging.debug('self.num_text_pages %d', self.num_text_pages)
+ logging.debug('self.num_footnote_pages %d, self.first_footnote_page %d', self.num_footnote_pages , self.first_footnote_page)
+ logging.debug('self.num_sidebar_pages %d, self.first_sidebar_page %d', self.num_sidebar_pages , self.first_sidebar_page)
+ self.flags = struct.unpack('>L', r[4:8])[0]
+ reqd_flags = (1<<9) | (1<<7) | (1<<10)
+ if (self.flags & reqd_flags) != reqd_flags:
+ print "Flags: 0x%X" % self.flags
+ raise ValueError('incompatible eReader file')
+ des = Des(fixKey(user_key))
+ if version == 259:
+ if drm_sub_version != 7:
+ raise ValueError('incorrect eReader version %d (error 3)' % drm_sub_version)
+ encrypted_key_sha = r[44:44+20]
+ encrypted_key = r[64:64+8]
+ elif version == 260:
+ if drm_sub_version != 13 and drm_sub_version != 11:
+ raise ValueError('incorrect eReader version %d (error 3)' % drm_sub_version)
+ if drm_sub_version == 13:
+ encrypted_key = r[44:44+8]
+ encrypted_key_sha = r[52:52+20]
+ else:
+ encrypted_key = r[64:64+8]
+ encrypted_key_sha = r[44:44+20]
+ elif version == 272:
+ encrypted_key = r[172:172+8]
+ encrypted_key_sha = r[56:56+20]
+ self.content_key = des.decrypt(encrypted_key)
+ if sha1(self.content_key).digest() != encrypted_key_sha:
+ raise ValueError('Incorrect Name and/or Credit Card')
+
+ def getNumImages(self):
+ return self.num_image_pages
+
+ def getImage(self, i):
+ sect = self.section_reader(self.first_image_page + i)
+ name = sect[4:4+32].strip('\0')
+ data = sect[62:]
+ return sanitizeFileName(unicode(name,'windows-1252')), data
+
+
+ # def getChapterNamePMLOffsetData(self):
+ # cv = ''
+ # if self.num_chapter_pages > 0:
+ # for i in xrange(self.num_chapter_pages):
+ # chaps = self.section_reader(self.first_chapter_page + i)
+ # j = i % self.xortable_size
+ # offname = deXOR(chaps, j, self.xortable)
+ # offset = struct.unpack('>L', offname[0:4])[0]
+ # name = offname[4:].strip('\0')
+ # cv += '%d|%s\n' % (offset, name)
+ # return cv
+
+ # def getLinkNamePMLOffsetData(self):
+ # lv = ''
+ # if self.num_link_pages > 0:
+ # for i in xrange(self.num_link_pages):
+ # links = self.section_reader(self.first_link_page + i)
+ # j = i % self.xortable_size
+ # offname = deXOR(links, j, self.xortable)
+ # offset = struct.unpack('>L', offname[0:4])[0]
+ # name = offname[4:].strip('\0')
+ # lv += '%d|%s\n' % (offset, name)
+ # return lv
+
+ # def getExpandedTextSizesData(self):
+ # ts = ''
+ # if self.num_xtextsize_pages > 0:
+ # tsize = deXOR(self.section_reader(self.first_xtextsize_page), 0, self.xortable)
+ # for i in xrange(self.num_text_pages):
+ # xsize = struct.unpack('>H', tsize[0:2])[0]
+ # ts += "%d\n" % xsize
+ # tsize = tsize[2:]
+ # return ts
+
+ # def getBookInfo(self):
+ # bkinfo = ''
+ # if self.num_bookinfo_pages > 0:
+ # info = self.section_reader(self.first_bookinfo_page)
+ # bkinfo = deXOR(info, 0, self.xortable)
+ # bkinfo = bkinfo.replace('\0','|')
+ # bkinfo += '\n'
+ # return bkinfo
+
+ def getText(self):
+ des = Des(fixKey(self.content_key))
+ r = ''
+ for i in xrange(self.num_text_pages):
+ logging.debug('get page %d', i)
+ r += zlib.decompress(des.decrypt(self.section_reader(1 + i)))
+
+ # now handle footnotes pages
+ if self.num_footnote_pages > 0:
+ r += '\n'
+ # the record 0 of the footnote section must pass through the Xor Table to make it useful
+ sect = self.section_reader(self.first_footnote_page)
+ fnote_ids = deXOR(sect, 0, self.xortable)
+ # the remaining records of the footnote sections need to be decoded with the content_key and zlib inflated
+ des = Des(fixKey(self.content_key))
+ for i in xrange(1,self.num_footnote_pages):
+ logging.debug('get footnotepage %d', i)
+ id_len = ord(fnote_ids[2])
+ id = fnote_ids[3:3+id_len]
+ fmarker = '' % int(simgsrc))
- else:
- hlst.append('