objavi/epub_utils.py

   1 """Module for dealing with booki -> epub conversions."""
   2
   3 import os, sys
   4 from pprint import pprint
   5 #import zipfile
   6 from cStringIO import StringIO
   7
   8 import lxml, lxml.html, lxml.cssselect
   9 from lxml import etree
  10
  11 from objavi.cgi_utils import log
  12 from objavi.config import DC, XHTML, XHTMLNS, NAVPOINT_ID_TEMPLATE
  13
  14 from booki.bookizip import get_metadata, get_metadata_schemes
  15
  16 ##Construct NCX
  17 BARE_NCX = ('<!DOCTYPE ncx PUBLIC "-//NISO//DTD ncx 2005-1//EN" '
  18             '"http://www.daisy.org/z3986/2005/ncx-2005-1.dtd"> '
  19             '<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" version="2005-1" />')
  20
  21 def add_ncxtext(parent, tag, text):
  22     """put text in a <text> subelement (as required by navLabel, navInfo)."""
  23     el = etree.SubElement(parent, tag)
  24     el2 = etree.SubElement(el, 'text')
  25     el2.text = text
  26
  27
  28 def make_ncx(toc, metadata, filemap):
  29     log(filemap)
  30     tree = etree.parse(StringIO(BARE_NCX))
  31     root = tree.getroot()
  32     head = etree.SubElement(root, 'head')
  33     add_ncxtext(root, 'docTitle', get_metadata(metadata, 'title')[0])
  34     navmap = etree.SubElement(root, 'navMap')
  35     counter, maxdepth = 0, 0
  36     for subtoc in toc:
  37         counter, maxdepth = write_navtree(navmap, subtoc, counter, 1, maxdepth, filemap)
  38     ids = get_metadata(metadata, 'identifier')
  39     if not ids: #workaround for one-time broken booki
  40         ids = [time.strftime('http://book.cc/UNKNOWN/%Y.%m.%d-%H.%M.%S')]
  41
  42     for name, content in (('dtb:uid', ids[0]),
  43                           ('dtb:depth', str(maxdepth)),
  44                           ('dtb:totalPageCount', '0'),
  45                           ('dtb:maxPageNumber', '0')
  46                           ):
  47         etree.SubElement(head, 'meta', name=name, content=content)
  48     return etree.tostring(tree, pretty_print=True, encoding='utf-8')
  49
  50
  51 def write_navtree(parent, subtoc, counter, depth, maxdepth, filemap):
  52     #subtoc has this structure:
  53     #{
  54     #  "title":    division title (optional),
  55     #  "url":      filename and possible fragment ID,
  56     #  "type":     string indicating division type (optional),
  57     #  "role":     epub guide type (optional),
  58     #  "children": list of TOC structures (optional)
  59     #}
  60     counter += 1
  61     if depth > maxdepth:
  62         maxdepth = depth
  63
  64     title = subtoc.get('title', '')
  65     url = subtoc['url']
  66     children = subtoc.get('children', [])
  67
  68     if url is None and children:
  69         # if the section has no url, it begins with its first child
  70         url = children[0]['url']
  71
  72     if filemap:
  73         url = filemap.get(url, url)
  74
  75     navpoint = make_navpoint(parent, counter, title, url)
  76     for point in children:
  77         counter, maxdepth = write_navtree(navpoint, point, counter, depth + 1, maxdepth, filemap)
  78
  79     return counter, maxdepth
  80
  81 def make_navpoint(parent, n, title, url):
  82     log((parent, n, title, url))
  83     """Make the actual navpoint node"""
  84     navpoint = etree.SubElement(parent, 'navPoint',
  85                                 id=(NAVPOINT_ID_TEMPLATE % (n - 1)),
  86                                 playOrder=str(n))
  87     add_ncxtext(navpoint, 'navLabel', title)
  88     etree.SubElement(navpoint, 'content', src=url)
  89     return navpoint
  90