path: root/src/epy_reader/tools/KindleUnpack/mobi_k8proc.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# vim:ts=4:sw=4:softtabstop=4:smarttab:expandtab

from __future__ import unicode_literals, division, absolute_import, print_function

from .compatibility_utils import PY2, bstr, utf8_str

if PY2:
    range = xrange

import os

import struct
# note:  struct pack, unpack, unpack_from all require bytestring format
# data all the way up to at least python 2.7.5, python 3 okay with bytestring

import re
# note: re requites the pattern to be the exact same type as the data to be searched in python3
# but u"" is not allowed for the pattern itself only b""

from .mobi_index import MobiIndex
from .mobi_utils import fromBase32
from .unipath import pathof

_guide_types = [b'cover',b'title-page',b'toc',b'index',b'glossary',b'acknowledgements',
                b'bibliography',b'colophon',b'copyright-page',b'dedication',
                b'epigraph',b'foreward',b'loi',b'lot',b'notes',b'preface',b'text']

# locate beginning and ending positions of tag with specific aid attribute
def locate_beg_end_of_tag(ml, aid):
    pattern = utf8_str(r'''<[^>]*\said\s*=\s*['"]%s['"][^>]*>''' % aid)
    aid_pattern = re.compile(pattern,re.IGNORECASE)
    for m in re.finditer(aid_pattern, ml):
        plt = m.start()
        pgt = ml.find(b'>',plt+1)
        return plt, pgt
    return 0, 0


# iterate over all tags in block in reverse order, i.e. last ta to first tag
def reverse_tag_iter(block):
    end = len(block)
    while True:
        pgt = block.rfind(b'>', 0, end)
        if pgt == -1:
            break
        plt = block.rfind(b'<', 0, pgt)
        if plt == -1:
            break
        yield block[plt:pgt+1]
        end = plt


class K8Processor:

    def __init__(self, mh, sect, files, debug=False):
        self.sect = sect
        self.files = files
        self.mi = MobiIndex(sect)
        self.mh = mh
        self.skelidx = mh.skelidx
        self.fragidx = mh.fragidx
        self.guideidx = mh.guideidx
        self.fdst = mh.fdst
        self.flowmap = {}
        self.flows = None
        self.flowinfo = []
        self.parts = None
        self.partinfo = []
        self.linked_aids = set()
        self.fdsttbl= [0,0xffffffff]
        self.DEBUG = debug

        # read in and parse the FDST info which is very similar in format to the Palm DB section
        # parsing except it provides offsets into rawML file and not the Palm DB file
        # this is needed to split up the final css, svg, etc flow section
        # that can exist at the end of the rawML file
        if self.fdst != 0xffffffff:
            header = self.sect.loadSection(self.fdst)
            if header[0:4] == b"FDST":
                num_sections, = struct.unpack_from(b'>L', header, 0x08)
                self.fdsttbl = struct.unpack_from(bstr('>%dL' % (num_sections*2)), header, 12)[::2] + (mh.rawSize, )
                sect.setsectiondescription(self.fdst,"KF8 FDST INDX")
                if self.DEBUG:
                    print("\nFDST Section Map:  %d sections" % num_sections)
                    for j in range(num_sections):
                        print("Section %d: 0x%08X - 0x%08X" % (j, self.fdsttbl[j],self.fdsttbl[j+1]))
            else:
                print("\nError: K8 Mobi with Missing FDST info")

        # read/process skeleton index info to create the skeleton table
        skeltbl = []
        if self.skelidx != 0xffffffff:
            # for i in range(2):
            #     fname = 'skel%04d.dat' % i
            #     data = self.sect.loadSection(self.skelidx + i)
            #     with open(pathof(fname), 'wb') as f:
            #         f.write(data)
            outtbl, ctoc_text = self.mi.getIndexData(self.skelidx, "KF8 Skeleton")
            fileptr = 0
            for [text, tagMap] in outtbl:
                # file number, skeleton name, fragtbl record count, start position, length
                skeltbl.append([fileptr, text, tagMap[1][0], tagMap[6][0], tagMap[6][1]])
                fileptr += 1
        self.skeltbl = skeltbl
        if self.DEBUG:
            print("\nSkel Table:  %d entries" % len(self.skeltbl))
            print("table: filenum, skeleton name, frag tbl record count, start position, length")
            for j in range(len(self.skeltbl)):
                print(self.skeltbl[j])

        # read/process the fragment index to create the fragment table
        fragtbl = []
        if self.fragidx != 0xffffffff:
            # for i in range(3):
            #     fname = 'frag%04d.dat' % i
            #     data = self.sect.loadSection(self.fragidx + i)
            #     with open(pathof(fname), 'wb') as f:
            #         f.write(data)
            outtbl, ctoc_text = self.mi.getIndexData(self.fragidx, "KF8 Fragment")
            for [text, tagMap] in outtbl:
                # insert position, ctoc offset (aidtext), file number, sequence number, start position, length
                ctocoffset = tagMap[2][0]
                ctocdata = ctoc_text[ctocoffset]
                fragtbl.append([int(text), ctocdata, tagMap[3][0], tagMap[4][0], tagMap[6][0], tagMap[6][1]])
        self.fragtbl = fragtbl
        if self.DEBUG:
            print("\nFragment Table: %d entries" % len(self.fragtbl))
            print("table: file position, link id text, file num, sequence number, start position, length")
            for j in range(len(self.fragtbl)):
                print(self.fragtbl[j])

        # read / process guide index for guide elements of opf
        guidetbl = []
        if self.guideidx != 0xffffffff:
            # for i in range(3):
            #     fname = 'guide%04d.dat' % i
            #     data = self.sect.loadSection(self.guideidx + i)
            #     with open(pathof(fname), 'wb') as f:
            #         f.write(data)
            outtbl, ctoc_text = self.mi.getIndexData(self.guideidx, "KF8 Guide elements)")
            for [text, tagMap] in outtbl:
                # ref_type, ref_title, frag number
                ctocoffset = tagMap[1][0]
                ref_title = ctoc_text[ctocoffset]
                ref_type = text
                fileno = None
                if 3 in tagMap:
                    fileno  = tagMap[3][0]
                if 6 in tagMap:
                    fileno = tagMap[6][0]
                guidetbl.append([ref_type, ref_title, fileno])
        self.guidetbl = guidetbl
        if self.DEBUG:
            print("\nGuide Table: %d entries" % len(self.guidetbl))
            print("table: ref_type, ref_title, fragtbl entry number")
            for j in range(len(self.guidetbl)):
                print(self.guidetbl[j])

    def buildParts(self, rawML):
        # now split the rawML into its flow pieces
        self.flows = []
        for j in range(0, len(self.fdsttbl)-1):
            start = self.fdsttbl[j]
            end = self.fdsttbl[j+1]
            self.flows.append(rawML[start:end])

        # the first piece represents the xhtml text
        text = self.flows[0]
        self.flows[0] = b''

        # walk the <skeleton> and fragment tables to build original source xhtml files
        # *without* destroying any file position information needed for later href processing
        # and create final list of file separation start: stop points and etc in partinfo
        if self.DEBUG:
            print("\nRebuilding flow piece 0: the main body of the ebook")
        self.parts = []
        self.partinfo = []
        fragptr = 0
        baseptr = 0
        cnt = 0
        filename = 'part%04d.xhtml' % cnt
        for [skelnum, skelname, fragcnt, skelpos, skellen] in self.skeltbl:
            baseptr = skelpos + skellen
            skeleton = text[skelpos: baseptr]
            aidtext = "0"
            for i in range(fragcnt):
                [insertpos, idtext, filenum, seqnum, startpos, length] = self.fragtbl[fragptr]
                aidtext = idtext[12:-2]
                if i == 0:
                    filename = 'part%04d.xhtml' % filenum
                slice = text[baseptr: baseptr + length]
                insertpos = insertpos - skelpos
                head = skeleton[:insertpos]
                tail = skeleton[insertpos:]
                actual_inspos = insertpos
                if (tail.find(b'>') < tail.find(b'<') or head.rfind(b'>') < head.rfind(b'<')):
                    # There is an incomplete tag in either the head or tail.
                    # This can happen for some badly formed KF8 files
                    print('The fragment table for %s has incorrect insert position. Calculating manually.' % skelname)
                    bp, ep = locate_beg_end_of_tag(skeleton, aidtext)
                    if bp != ep:
                        actual_inspos = ep + 1 + startpos
                if insertpos != actual_inspos:
                    print("fixed corrupt fragment table insert position", insertpos+skelpos, actual_inspos+skelpos)
                    insertpos = actual_inspos
                    self.fragtbl[fragptr][0] = actual_inspos + skelpos
                skeleton = skeleton[0:insertpos] + slice + skeleton[insertpos:]
                baseptr = baseptr + length
                fragptr += 1
            cnt += 1
            self.parts.append(skeleton)
            self.partinfo.append([skelnum, 'Text', filename, skelpos, baseptr, aidtext])

        assembled_text = b''.join(self.parts)
        if self.DEBUG:
            outassembled = os.path.join(self.files.k8dir, 'assembled_text.dat')
            with open(pathof(outassembled),'wb') as f:
                f.write(assembled_text)

        # The primary css style sheet is typically stored next followed by any
        # snippets of code that were previously inlined in the
        # original xhtml but have been stripped out and placed here.
        # This can include local CDATA snippets and and svg sections.

        # The problem is that for most browsers and ereaders, you can not
        # use <img src="imageXXXX.svg" /> to import any svg image that itself
        # properly uses an <image/> tag to import some raster image - it
        # should work according to the spec but does not for almost all browsers
        # and ereaders and causes epub validation issues because those  raster
        # images are in manifest but not in xhtml text - since they only
        # referenced from an svg image

        # So we need to check the remaining flow pieces to see if they are css
        # or svg images.  if svg images, we must check if they have an <image />
        # and if so inline them into the xhtml text pieces.

        # there may be other sorts of pieces stored here but until we see one
        # in the wild to reverse engineer we won't be able to tell
        self.flowinfo.append([None, None, None, None])
        svg_tag_pattern = re.compile(br'''(<svg[^>]*>)''', re.IGNORECASE)
        image_tag_pattern = re.compile(br'''(<image[^>]*>)''', re.IGNORECASE)
        for j in range(1,len(self.flows)):
            flowpart = self.flows[j]
            nstr = '%04d' % j
            m = re.search(svg_tag_pattern, flowpart)
            if m is not None:
                # svg
                ptype = b'svg'
                start = m.start()
                m2 = re.search(image_tag_pattern, flowpart)
                if m2 is not None:
                    pformat = b'inline'
                    pdir = None
                    fname = None
                    # strip off anything before <svg if inlining
                    flowpart = flowpart[start:]
                else:
                    pformat = b'file'
                    pdir = "Images"
                    fname = 'svgimg' + nstr + '.svg'
            else:
                # search for CDATA and if exists inline it
                if flowpart.find(b'[CDATA[') >= 0:
                    ptype = b'css'
                    flowpart = b'<style type="text/css">\n' + flowpart + b'\n</style>\n'
                    pformat = b'inline'
                    pdir = None
                    fname = None
                else:
                    # css - assume as standalone css file
                    ptype = b'css'
                    pformat = b'file'
                    pdir = "Styles"
                    fname = 'style' + nstr + '.css'

            self.flows[j] = flowpart
            self.flowinfo.append([ptype, pformat, pdir, fname])

        if self.DEBUG:
            print("\nFlow Map:  %d entries" % len(self.flowinfo))
            for fi in self.flowinfo:
                print(fi)
            print("\n")

            print("\nXHTML File Part Position Information: %d entries" % len(self.partinfo))
            for pi in self.partinfo:
                print(pi)

        if False:  # self.Debug:
            # dump all of the locations of the aid tags used in TEXT
            # find id links only inside of tags
            #    inside any < > pair find all "aid=' and return whatever is inside the quotes
            #    [^>]* means match any amount of chars except for  '>' char
            #    [^'"] match any amount of chars except for the quote character
            #    \s* means match any amount of whitespace
            print("\npositions of all aid= pieces")
            id_pattern = re.compile(br'''<[^>]*\said\s*=\s*['"]([^'"]*)['"][^>]*>''',re.IGNORECASE)
            for m in re.finditer(id_pattern, rawML):
                [filename, partnum, start, end] = self.getFileInfo(m.start())
                [seqnum, idtext] = self.getFragTblInfo(m.start())
                value = fromBase32(m.group(1))
                print("  aid: %s value: %d at: %d -> part: %d, start: %d, end: %d" % (m.group(1), value, m.start(), partnum, start, end))
                print("       %s  fragtbl entry %d" % (idtext, seqnum))

        return

    # get information fragment table entry by pos
    def getFragTblInfo(self, pos):
        for j in range(len(self.fragtbl)):
            [insertpos, idtext, filenum, seqnum, startpos, length] = self.fragtbl[j]
            if pos >= insertpos and pos < (insertpos + length):
                # why are these "in: and before: added here
                return seqnum, b'in: ' + idtext
            if pos < insertpos:
                return seqnum, b'before: ' + idtext
        return None, None

    # get information about the part (file) that exists at pos in original rawML
    def getFileInfo(self, pos):
        for [partnum, pdir, filename, start, end, aidtext] in self.partinfo:
            if pos >= start and pos < end:
                return filename, partnum, start, end
        return None, None, None, None

    # accessor functions to properly protect the internal structure
    def getNumberOfParts(self):
        return len(self.parts)

    def getPart(self,i):
        if i >= 0 and i < len(self.parts):
            return self.parts[i]
        return None

    def getPartInfo(self, i):
        if i >= 0 and i < len(self.partinfo):
            return self.partinfo[i]
        return None

    def getNumberOfFlows(self):
        return len(self.flows)

    def getFlow(self,i):
        # note flows[0] is empty - it was all of the original text
        if i > 0 and i < len(self.flows):
            return self.flows[i]
        return None

    def getFlowInfo(self,i):
        # note flowinfo[0] is empty - it was all of the original text
        if i > 0 and i < len(self.flowinfo):
            return self.flowinfo[i]
        return None

    def getIDTagByPosFid(self, posfid, offset):
        # first convert kindle:pos:fid and offset info to position in file
        # (fromBase32 can handle both string types on input)
        row = fromBase32(posfid)
        off = fromBase32(offset)
        [insertpos, idtext, filenum, seqnm, startpos, length] = self.fragtbl[row]
        pos = insertpos + off
        fname, pn, skelpos, skelend = self.getFileInfo(pos)
        if fname is None:
            # pos does not exist
            # default to skeleton pos instead
            print("Link To Position", pos, "does not exist, retargeting to top of target")
            pos = self.skeltbl[filenum][3]
            fname, pn, skelpos, skelend = self.getFileInfo(pos)
        # an existing "id=" or "name=" attribute must exist in original xhtml otherwise it would not have worked for linking.
        # Amazon seems to have added its own additional "aid=" inside tags whose contents seem to represent
        # some position information encoded into Base32 name.
        # so find the closest "id=" before position the file  by actually searching in that file
        idtext = self.getIDTag(pos)
        return fname, idtext

    def getIDTag(self, pos):
        # find the first tag with a named anchor (name or id attribute) before pos
        fname, pn, skelpos, skelend = self.getFileInfo(pos)
        if pn is None and skelpos is None:
            print("Error: getIDTag - no file contains ", pos)
        textblock = self.parts[pn]
        npos = pos - skelpos
        # if npos inside a tag then search all text before the its end of tag marker
        pgt = textblock.find(b'>',npos)
        plt = textblock.find(b'<',npos)
        if plt == npos or pgt < plt:
            npos = pgt + 1
        # find id and name attributes only inside of tags
        # use a reverse tag search since that is faster
        #    inside any < > pair find "id=" and "name=" attributes return it
        #    [^>]* means match any amount of chars except for  '>' char
        #    [^'"] match any amount of chars except for the quote character
        #    \s* means match any amount of whitespace
        textblock = textblock[0:npos]
        id_pattern = re.compile(br'''<[^>]*\sid\s*=\s*['"]([^'"]*)['"]''',re.IGNORECASE)
        name_pattern = re.compile(br'''<[^>]*\sname\s*=\s*['"]([^'"]*)['"]''',re.IGNORECASE)
        aid_pattern = re.compile(br'''<[^>]+\s(?:aid|AID)\s*=\s*['"]([^'"]+)['"]''')
        for tag in reverse_tag_iter(textblock):
            # any ids in the body should default to top of file
            if tag[0:6] == b'<body ':
                return b''
            if tag[0:6] != b'<meta ':
                m = id_pattern.match(tag) or name_pattern.match(tag)
                if m is not None:
                    return m.group(1)
                m = aid_pattern.match(tag)
                if m is not None:
                    self.linked_aids.add(m.group(1))
                    return b'aid-' + m.group(1)
        return b''

    # do we need to do deep copying
    def setParts(self, parts):
        assert(len(parts) == len(self.parts))
        for i in range(len(parts)):
            self.parts[i] = parts[i]

    # do we need to do deep copying
    def setFlows(self, flows):
        assert(len(flows) == len(self.flows))
        for i in range(len(flows)):
            self.flows[i] = flows[i]

    # get information about the part (file) that exists at pos in original rawML
    def getSkelInfo(self, pos):
        for [partnum, pdir, filename, start, end, aidtext] in self.partinfo:
            if pos >= start and pos < end:
                return [partnum, pdir, filename, start, end, aidtext]
        return [None, None, None, None, None, None]

    # fileno is actually a reference into fragtbl (a fragment)
    def getGuideText(self):
        guidetext = b''
        for [ref_type, ref_title, fileno] in self.guidetbl:
            if ref_type == b'thumbimagestandard':
                continue
            if ref_type not in _guide_types and not ref_type.startswith(b'other.'):
                if ref_type == b'start':
                    ref_type = b'text'
                else:
                    ref_type = b'other.' + ref_type
            [pos, idtext, filenum, seqnm, startpos, length] = self.fragtbl[fileno]
            [pn, pdir, filename, skelpos, skelend, aidtext] = self.getSkelInfo(pos)
            idtext = self.getIDTag(pos)
            linktgt = filename.encode('utf-8')
            if idtext != b'':
                linktgt += b'#' + idtext
            guidetext += b'<reference type="'+ref_type+b'" title="'+ref_title+b'" href="'+utf8_str(pdir)+b'/'+linktgt+b'" />\n'
        # opf is encoded utf-8 so must convert any titles properly
        guidetext = (guidetext.decode(self.mh.codec)).encode("utf-8")
        return guidetext

    def getPageIDTag(self, pos):
        # find the first tag with a named anchor (name or id attribute) before pos
        # but page map offsets need to little more leeway so if the offset points
        # into a tag look for the next ending tag "/>" or "</" and start your search from there.
        fname, pn, skelpos, skelend = self.getFileInfo(pos)
        if pn is None and skelpos is None:
            print("Error: getIDTag - no file contains ", pos)
        textblock = self.parts[pn]
        npos = pos - skelpos
        # if npos inside a tag then search all text before next ending tag
        pgt = textblock.find(b'>',npos)
        plt = textblock.find(b'<',npos)
        if plt == npos or pgt < plt:
            # we are in a tag
            # so find first ending tag
            pend1 = textblock.find(b'/>', npos)
            pend2 = textblock.find(b'</', npos)
            if pend1 != -1 and pend2 != -1:
                pend = min(pend1, pend2)
            else:
                pend = max(pend1, pend2)
            if pend != -1:
                npos = pend
            else:
                npos = pgt + 1
        # find id and name attributes only inside of tags
        # use a reverse tag search since that is faster
        #    inside any < > pair find "id=" and "name=" attributes return it
        #    [^>]* means match any amount of chars except for  '>' char
        #    [^'"] match any amount of chars except for the quote character
        #    \s* means match any amount of whitespace
        textblock = textblock[0:npos]
        id_pattern = re.compile(br'''<[^>]*\sid\s*=\s*['"]([^'"]*)['"]''',re.IGNORECASE)
        name_pattern = re.compile(br'''<[^>]*\sname\s*=\s*['"]([^'"]*)['"]''',re.IGNORECASE)
        for tag in reverse_tag_iter(textblock):
            # any ids in the body should default to top of file
            if tag[0:6] == b'<body ':
                return b''
            if tag[0:6] != b'<meta ':
                m = id_pattern.match(tag) or name_pattern.match(tag)
                if m is not None:
                    return m.group(1)
        return b''