Fix transform_abstracy + tests.

[librarian.git] / librarian / epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index 201799b..e9670d5 100644 (file)
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -1,44 +1,99 @@
 # -*- coding: utf-8 -*-
 #
 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
 # -*- coding: utf-8 -*-
 #
 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.

-# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.  
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.

 #
 #

-from __future__ import with_statement
+from __future__ import print_function, unicode_literals

 
 import os
 import os.path
 
 import os
 import os.path

-import shutil
-import sys
+import re
+import subprocess
+from six import BytesIO

 from copy import deepcopy
 from copy import deepcopy

+from mimetypes import guess_type
+

 from lxml import etree
 import zipfile
 from lxml import etree
 import zipfile

+from tempfile import mkdtemp, NamedTemporaryFile
+from shutil import rmtree
+
+from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, DCNS, OutputFile
+from librarian.cover import make_cover
+
+from librarian import functions, get_resource

 
 

-from librarian import XMLNamespace, RDFNS, DCNS, WLNS, XHTMLNS
+from librarian.hyphenator import Hyphenator

 
 

-NCXNS = XMLNamespace("http://www.daisy.org/z3986/2005/ncx/")
-OPFNS = XMLNamespace("http://www.idpf.org/2007/opf")
+functions.reg_person_name()
+functions.reg_lang_code_3to2()
+
+
+def squeeze_whitespace(s):
+    return re.sub(b'\\s+', b' ', s)
+
+
+def set_hyph_language(source_tree):
+    def get_short_lng_code(text):
+        result = ''
+        text = ''.join(text)
+        with open(get_resource('res/ISO-639-2_8859-1.txt'), 'rb') as f:
+            for line in f.read().decode('latin1').split('\n'):
+                list = line.strip().split('|')
+                if list[0] == text:
+                    result = list[2]
+        if result == '':
+            return text
+        else:
+            return result
+    bibl_lng = etree.XPath('//dc:language//text()',
+                           namespaces={'dc': str(DCNS)})(source_tree)
+    short_lng = get_short_lng_code(bibl_lng[0])
+    try:
+        return Hyphenator(get_resource('res/hyph-dictionaries/hyph_' +
+                                       short_lng + '.dic'))
+    except:
+        pass
+
+
+def hyphenate_and_fix_conjunctions(source_tree, hyph):
+    texts = etree.XPath('/utwor/*[2]//text()')(source_tree)
+    for t in texts:
+        parent = t.getparent()
+        if hyph is not None:
+            newt = ''
+            wlist = re.compile(r'\w+|[^\w]', re.UNICODE).findall(t)
+            for w in wlist:
+                newt += hyph.inserted(w, u'\u00AD')
+        else:
+            newt = t
+        newt = re.sub(r'(?<=\s\w)\s+', u'\u00A0', newt)
+        if t.is_text:
+            parent.text = newt
+        elif t.is_tail:
+            parent.tail = newt

 
 
 def inner_xml(node):
     """ returns node's text and children as a string
 
 
 
 def inner_xml(node):
     """ returns node's text and children as a string
 

-    >>> print inner_xml(etree.fromstring('<a>x<b>y</b>z</a>'))
+    >>> print(inner_xml(etree.fromstring('<a>x<b>y</b>z</a>')))

     x<b>y</b>z
     """
 
     nt = node.text if node.text is not None else ''
     x<b>y</b>z
     """
 
     nt = node.text if node.text is not None else ''

-    return ''.join([nt] + [etree.tostring(child) for child in node]) 
+    return ''.join([nt] + [etree.tostring(child, encoding='unicode') for child in node])
+

 
 def set_inner_xml(node, text):
     """ sets node's text and children from a string
 
     >>> e = etree.fromstring('<a>b<b>x</b>x</a>')
     >>> set_inner_xml(e, 'x<b>y</b>z')
 
 def set_inner_xml(node, text):
     """ sets node's text and children from a string
 
     >>> e = etree.fromstring('<a>b<b>x</b>x</a>')
     >>> set_inner_xml(e, 'x<b>y</b>z')

-    >>> print etree.tostring(e)
+    >>> print(etree.tostring(e, encoding='unicode'))

     <a>x<b>y</b>z</a>
     """
 
     <a>x<b>y</b>z</a>
     """
 

-    

     p = etree.fromstring('<x>%s</x>' % text)
     node.text = p.text
     node[:] = p[:]
     p = etree.fromstring('<x>%s</x>' % text)
     node.text = p.text
     node[:] = p[:]


@@ -47,7 +102,7 @@ def set_inner_xml(node, text):
 def node_name(node):
     """ Find out a node's name
 
 def node_name(node):
     """ Find out a node's name
 

-    >>> print node_name(etree.fromstring('<a>X<b>Y</b>Z</a>'))
+    >>> print(node_name(etree.fromstring('<a>X<b>Y</b>Z</a>')))

     XYZ
     """
 
     XYZ
     """
 


@@ -62,303 +117,599 @@ def node_name(node):
     return tempnode.text
 
 
     return tempnode.text
 
 

-def xslt(xml, sheet):
+def xslt(xml, sheet, **kwargs):

     if isinstance(xml, etree._Element):
         xml = etree.ElementTree(xml)
     with open(sheet) as xsltf:
     if isinstance(xml, etree._Element):
         xml = etree.ElementTree(xml)
     with open(sheet) as xsltf:

-        return xml.xslt(etree.parse(xsltf))
-
-
-_resdir = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'epub')
-def res(fname):
-    return os.path.join(_resdir, fname)
+        transform = etree.XSLT(etree.parse(xsltf))
+        params = dict((key, transform.strparam(value)) for key, value in kwargs.items())
+        return transform(xml, **params)

 
 
 def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None
 
 
 def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None

-        return text.replace("&", "&amp;")\
-                   .replace("---", "&#8212;")\
-                   .replace("--", "&#8211;")\
-                   .replace(",,", "&#8222;")\
-                   .replace('"', "&#8221;")\
-                   .replace("'", "&#8217;")
-    if node.tag == 'extra':
+        return text.replace(u"\ufeff", u"")\
+                   .replace("---", u"\u2014")\
+                   .replace("--", u"\u2013")\
+                   .replace(",,", u"\u201E")\
+                   .replace('"', u"\u201D")\
+                   .replace("'", u"\u2019")
+    if node.tag in ('uwaga', 'extra'):
+        t = node.tail

         node.clear()
         node.clear()

-    else:
-        node.text = replace_chars(node.text)
-        node.tail = replace_chars(node.tail)
-        for child in node:
-            replace_characters(child)
+        node.tail = t
+    node.text = replace_chars(node.text)
+    node.tail = replace_chars(node.tail)
+    for child in node:
+        replace_characters(child)

 
 
 
 

-def find_annotations(annotations, source, part_number):
+def find_annotations(annotations, source, part_no):

     for child in source:
         if child.tag in ('pe', 'pa', 'pt', 'pr'):
             annotation = deepcopy(child)
     for child in source:
         if child.tag in ('pe', 'pa', 'pt', 'pr'):
             annotation = deepcopy(child)

-            annotation.set('number', str(len(annotations)+1))
-            annotation.set('part', str(part_number))
+            number = str(len(annotations) + 1)
+            annotation.set('number', number)
+            annotation.set('part', str(part_no))

             annotation.tail = ''
             annotations.append(annotation)
             tail = child.tail
             child.clear()
             child.tail = tail
             annotation.tail = ''
             annotations.append(annotation)
             tail = child.tail
             child.clear()
             child.tail = tail

-            child.text = str(len(annotations))
-        if child.tag not in ('extra', 'podtytul'):
-            find_annotations(annotations, child, part_number)
+            child.text = number
+        if child.tag not in ('extra', 'uwaga'):
+            find_annotations(annotations, child, part_no)
+
+
+class Stanza(object):
+    """
+    Converts / verse endings into verse elements in a stanza.
+
+    Slashes may only occur directly in the stanza. Any slashes in subelements
+    will be ignored, and the subelements will be put inside verse elements.
+
+    >>> s = etree.fromstring("<strofa>a <b>c</b> <b>c</b>/\\nb<x>x/\\ny</x>c/ \\nd</strofa>")
+    >>> Stanza(s).versify()
+    >>> print(etree.tostring(s, encoding='unicode'))
+    <strofa><wers_normalny>a <b>c</b><b>c</b></wers_normalny><wers_normalny>b<x>x/
+    y</x>c</wers_normalny><wers_normalny>d</wers_normalny></strofa>
+
+    """
+    def __init__(self, stanza_elem):
+        self.stanza = stanza_elem
+        self.verses = []
+        self.open_verse = None
+
+    def versify(self):
+        self.push_text(self.stanza.text)
+        for elem in self.stanza:
+            self.push_elem(elem)
+            self.push_text(elem.tail)
+        tail = self.stanza.tail
+        self.stanza.clear()
+        self.stanza.tail = tail
+        self.stanza.extend(verse for verse in self.verses if verse.text or len(verse) > 0)
+
+    def open_normal_verse(self):
+        self.open_verse = self.stanza.makeelement("wers_normalny")
+        self.verses.append(self.open_verse)
+
+    def get_open_verse(self):
+        if self.open_verse is None:
+            self.open_normal_verse()
+        return self.open_verse
+
+    def push_text(self, text):
+        if not text:
+            return
+        for i, verse_text in enumerate(re.split(r"/\s*\n", text)):
+            if i:
+                self.open_normal_verse()
+            if not verse_text.strip():
+                continue
+            verse = self.get_open_verse()
+            if len(verse):
+                verse[-1].tail = (verse[-1].tail or "") + verse_text
+            else:
+                verse.text = (verse.text or "") + verse_text
+
+    def push_elem(self, elem):
+        if elem.tag.startswith("wers"):
+            verse = deepcopy(elem)
+            verse.tail = None
+            self.verses.append(verse)
+            self.open_verse = verse
+        else:
+            appended = deepcopy(elem)
+            appended.tail = None
+            self.get_open_verse().append(appended)

 
 
 def replace_by_verse(tree):
     """ Find stanzas and create new verses in place of a '/' character """
 
 
 def replace_by_verse(tree):
     """ Find stanzas and create new verses in place of a '/' character """

+

     stanzas = tree.findall('.//' + WLNS('strofa'))
     stanzas = tree.findall('.//' + WLNS('strofa'))

-    for node in stanzas:
-        for child_node in node:
-            if child_node.tag in ('slowo_obce', 'wyroznienie'):
-                foreign_verses = inner_xml(child_node).split('/\n')
-                if len(foreign_verses) > 1:
-                    new_foreign = ''
-                    for foreign_verse in foreign_verses:
-                        if foreign_verse.startswith('<wers'):
-                            new_foreign += foreign_verse
-                        else:
-                            new_foreign += ''.join(('<wers_normalny>', foreign_verse, '</wers_normalny>'))
-                    set_inner_xml(child_node, new_foreign)
-        verses = inner_xml(node).split('/\n')
-        if len(verses) > 1:
-            modified_inner_xml = ''
-            for verse in verses:
-                if verse.startswith('<wers') or verse.startswith('<extra'):
-                    modified_inner_xml += verse
-                else:
-                    modified_inner_xml += ''.join(('<wers_normalny>', verse, '</wers_normalny>'))
-            set_inner_xml(node, modified_inner_xml)
+    for stanza in stanzas:
+        Stanza(stanza).versify()

 
 
 def add_to_manifest(manifest, partno):
     """ Adds a node to the manifest section in content.opf file """
 
 
 def add_to_manifest(manifest, partno):
     """ Adds a node to the manifest section in content.opf file """

+

     partstr = 'part%d' % partno
     partstr = 'part%d' % partno

-    e = manifest.makeelement(OPFNS('item'), attrib={
-                                 'id': partstr,
-                                 'href': partstr + '.html',
-                                 'media-type': 'application/xhtml+xml',
-                             })
+    e = manifest.makeelement(
+        OPFNS('item'), attrib={'id': partstr, 'href': partstr + '.html',
+                               'media-type': 'application/xhtml+xml'}
+    )

     manifest.append(e)
 
 
 def add_to_spine(spine, partno):
     """ Adds a node to the spine section in content.opf file """
     manifest.append(e)
 
 
 def add_to_spine(spine, partno):
     """ Adds a node to the spine section in content.opf file """

-    e = spine.makeelement(OPFNS('itemref'), attrib={'idref': 'part%d' % partno});
-    spine.append(e)
-
-
-def add_nav_point(nav_map, counter, title, part_counter):
-    nav_point = nav_map.makeelement(NCXNS('navPoint'))
-    nav_point.set('id', 'NavPoint-%d' % counter)
-    nav_point.set('playOrder', str(counter))
-
-    nav_label = nav_map.makeelement(NCXNS('navLabel'))
-    text = nav_map.makeelement(NCXNS('text'))
-    text.text = title
-    nav_label.append(text)
-    nav_point.append(nav_label)

 
 

-    content = nav_map.makeelement(NCXNS('content'))
-    content.set('src', 'part%d.html' % part_counter)
-    nav_point.append(content)
+    e = spine.makeelement(OPFNS('itemref'), attrib={'idref': 'part%d' % partno})
+    spine.append(e)

 
 

-    nav_map.append(nav_point)

 
 

+class TOC(object):
+    def __init__(self, name=None, part_href=None):
+        self.children = []
+        self.name = name
+        self.part_href = part_href
+        self.sub_number = None

 
 

-def add_nav_point2(nav_map, counter, title, part_counter, subcounter):
-    nav_point = nav_map.makeelement(NCXNS('navPoint'))
-    nav_point.set('id', 'NavPoint-%d' % counter)
-    nav_point.set('playOrder', str(counter))
+    def add(self, name, part_href, level=0, is_part=True, index=None):
+        assert level == 0 or index is None
+        if level > 0 and self.children:
+            return self.children[-1].add(name, part_href, level - 1, is_part)
+        else:
+            t = TOC(name)
+            t.part_href = part_href
+            if index is not None:
+                self.children.insert(index, t)
+            else:
+                self.children.append(t)
+            if not is_part:
+                t.sub_number = len(self.children) + 1
+                return t.sub_number

 
 

-    nav_label = nav_map.makeelement(NCXNS('navLabel'))
-    text = nav_map.makeelement(NCXNS('text'))
-    text.text = title
-    nav_label.append(text)
-    nav_point.append(nav_label)
+    def append(self, toc):
+        self.children.append(toc)

 
 

-    content = nav_map.makeelement(NCXNS('content'))
-    content.set('src', 'part%d.html#sub%d' % (part_counter, subcounter))
-    nav_point.append(content)
+    def extend(self, toc):
+        self.children.extend(toc.children)

 
 

-    nav_map[-1].append(nav_point)
+    def depth(self):
+        if self.children:
+            return max((c.depth() for c in self.children)) + 1
+        else:
+            return 0
+
+    def href(self):
+        src = self.part_href
+        if self.sub_number is not None:
+            src += '#sub%d' % self.sub_number
+        return src
+
+    def write_to_xml(self, nav_map, counter=1):
+        for child in self.children:
+            nav_point = nav_map.makeelement(NCXNS('navPoint'))
+            nav_point.set('id', 'NavPoint-%d' % counter)
+            nav_point.set('playOrder', str(counter))
+
+            nav_label = nav_map.makeelement(NCXNS('navLabel'))
+            text = nav_map.makeelement(NCXNS('text'))
+            if child.name is not None:
+                text.text = re.sub(r'\n', ' ', child.name)
+            else:
+                text.text = child.name
+            nav_label.append(text)
+            nav_point.append(nav_label)
+
+            content = nav_map.makeelement(NCXNS('content'))
+            content.set('src', child.href())
+            nav_point.append(content)
+            nav_map.append(nav_point)
+            counter = child.write_to_xml(nav_point, counter + 1)
+        return counter
+
+    def html_part(self, depth=0):
+        texts = []
+        for child in self.children:
+            texts.append(
+                "<div style='margin-left:%dem;'><a href='%s'>%s</a></div>" %
+                (depth, child.href(), child.name))
+            texts.append(child.html_part(depth + 1))
+        return "\n".join(texts)
+
+    def html(self):
+        with open(get_resource('epub/toc.html'), 'rb') as f:
+            t = f.read().decode('utf-8')
+        return t % self.html_part()
+
+
+def used_chars(element):
+    """ Lists characters used in an ETree Element """
+    chars = set((element.text or '') + (element.tail or ''))
+    for child in element:
+        chars = chars.union(used_chars(child))
+    return chars
+
+
+def chop(main_text):
+    """ divide main content of the XML file into chunks """
+
+    # prepare a container for each chunk
+    part_xml = etree.Element('utwor')
+    etree.SubElement(part_xml, 'master')
+    main_xml_part = part_xml[0]  # master

 
 

+    last_node_part = False

 
 

-def transform(input_file, output_file):
-    """ produces an epub
+    # the below loop are workaround for a problem with epubs in drama ebooks without acts
+    is_scene = False
+    is_act = False
+    for one_part in main_text:
+        name = one_part.tag
+        if name == 'naglowek_scena':
+            is_scene = True
+        elif name == 'naglowek_akt':
+            is_act = True

 
 

-    input_file and output_file should be filelike objects
+    for one_part in main_text:
+        name = one_part.tag
+        if is_act is False and is_scene is True:
+            if name == 'naglowek_czesc':
+                yield part_xml
+                last_node_part = True
+                main_xml_part[:] = [deepcopy(one_part)]
+            elif not last_node_part and name == "naglowek_scena":
+                yield part_xml
+                main_xml_part[:] = [deepcopy(one_part)]
+            else:
+                main_xml_part.append(deepcopy(one_part))
+                last_node_part = False
+        else:
+            if name == 'naglowek_czesc':
+                yield part_xml
+                last_node_part = True
+                main_xml_part[:] = [deepcopy(one_part)]
+            elif not last_node_part and name in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
+                yield part_xml
+                main_xml_part[:] = [deepcopy(one_part)]
+            else:
+                main_xml_part.append(deepcopy(one_part))
+                last_node_part = False
+    yield part_xml
+
+
+def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_static=[]):
+    """ transforms one chunk, returns a HTML string, a TOC object and a set of used characters """
+
+    toc = TOC()
+    for element in chunk_xml[0]:
+        if element.tag == "naglowek_czesc":
+            toc.add(node_name(element), "part%d.html#book-text" % chunk_no)
+        elif element.tag in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
+            toc.add(node_name(element), "part%d.html" % chunk_no)
+        elif element.tag in ('naglowek_podrozdzial', 'naglowek_scena'):
+            subnumber = toc.add(node_name(element), "part%d.html" % chunk_no, level=1, is_part=False)
+            element.set('sub', str(subnumber))
+    if empty:
+        if not _empty_html_static:
+            _empty_html_static.append(open(get_resource('epub/emptyChunk.html')).read())
+        chars = set()
+        output_html = _empty_html_static[0]
+    else:
+        find_annotations(annotations, chunk_xml, chunk_no)
+        replace_by_verse(chunk_xml)
+        html_tree = xslt(chunk_xml, get_resource('epub/xsltScheme.xsl'))
+        chars = used_chars(html_tree.getroot())
+        output_html = etree.tostring(
+            html_tree, pretty_print=True, xml_declaration=True,
+            encoding="utf-8",
+            doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" ' +
+                    '"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+        )
+    return output_html, toc, chars
+
+
+def transform(wldoc, verbose=False, style=None, html_toc=False,
+              sample=None, cover=None, flags=None, hyphenate=False, ilustr_path='', output_type='epub'):
+    """ produces a EPUB file
+
+    sample=n: generate sample e-book (with at least n paragraphs)
+    cover: a cover.Cover factory or True for default
+    flags: less-advertising, without-fonts, working-copy

     """
 
     """
 

-    input_xml = etree.parse(input_file)
+    def transform_file(wldoc, chunk_counter=1, first=True, sample=None):
+        """ processes one input file and proceeds to its children """
+
+        replace_characters(wldoc.edoc.getroot())
+
+        hyphenator = set_hyph_language(wldoc.edoc.getroot()) if hyphenate else None
+        hyphenate_and_fix_conjunctions(wldoc.edoc.getroot(), hyphenator)
+
+        # every input file will have a TOC entry,
+        # pointing to starting chunk
+        toc = TOC(wldoc.book_info.title, "part%d.html" % chunk_counter)
+        chars = set()
+        if first:
+            # write book title page
+            html_tree = xslt(wldoc.edoc, get_resource('epub/xsltTitle.xsl'), outputtype=output_type)
+            chars = used_chars(html_tree.getroot())
+            html_string = etree.tostring(
+                html_tree, pretty_print=True, xml_declaration=True,
+                encoding="utf-8",
+                doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"' +
+                        ' "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+            )
+            zip.writestr('OPS/title.html', squeeze_whitespace(html_string))
+            # add a title page TOC entry
+            toc.add(u"Strona tytułowa", "title.html")
+        elif wldoc.book_info.parts:
+            # write title page for every parent
+            if sample is not None and sample <= 0:
+                chars = set()
+                html_string = open(get_resource('epub/emptyChunk.html')).read()
+            else:
+                html_tree = xslt(wldoc.edoc, get_resource('epub/xsltChunkTitle.xsl'))
+                chars = used_chars(html_tree.getroot())
+                html_string = etree.tostring(
+                    html_tree, pretty_print=True, xml_declaration=True,
+                    encoding="utf-8",
+                    doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"' +
+                            ' "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+                )
+            zip.writestr('OPS/part%d.html' % chunk_counter, squeeze_whitespace(html_string))
+            add_to_manifest(manifest, chunk_counter)
+            add_to_spine(spine, chunk_counter)
+            chunk_counter += 1
+
+        if len(wldoc.edoc.getroot()) > 1:
+            # rdf before style master
+            main_text = wldoc.edoc.getroot()[1]
+        else:
+            # rdf in style master
+            main_text = wldoc.edoc.getroot()[0]
+            if main_text.tag == RDFNS('RDF'):
+                main_text = None
+
+        if main_text is not None:
+            for chunk_xml in chop(main_text):
+                empty = False
+                if sample is not None:
+                    if sample <= 0:
+                        empty = True
+                    else:
+                        sample -= len(chunk_xml.xpath('//strofa|//akap|//akap_cd|//akap_dialog'))
+                chunk_html, chunk_toc, chunk_chars = transform_chunk(chunk_xml, chunk_counter, annotations, empty)
+
+                toc.extend(chunk_toc)
+                chars = chars.union(chunk_chars)
+                zip.writestr('OPS/part%d.html' % chunk_counter, squeeze_whitespace(chunk_html))
+                add_to_manifest(manifest, chunk_counter)
+                add_to_spine(spine, chunk_counter)
+                chunk_counter += 1
+
+        for child in wldoc.parts():
+            child_toc, chunk_counter, chunk_chars, sample = transform_file(
+                child, chunk_counter, first=False, sample=sample)
+            toc.append(child_toc)
+            chars = chars.union(chunk_chars)
+
+        return toc, chunk_counter, chars, sample
+
+    document = deepcopy(wldoc)
+    del wldoc
+
+    if flags:
+        for flag in flags:
+            document.edoc.getroot().set(flag, 'yes')
+
+    document.clean_ed_note()
+    document.clean_ed_note('abstrakt')
+
+    # add editors info
+    editors = document.editors()
+    if editors:
+        document.edoc.getroot().set('editors', u', '.join(sorted(
+            editor.readable() for editor in editors)))
+    if document.book_info.funders:
+        document.edoc.getroot().set('funders', u', '.join(
+            document.book_info.funders))
+    if document.book_info.thanks:
+        document.edoc.getroot().set('thanks', document.book_info.thanks)
+
+    opf = xslt(document.book_info.to_etree(), get_resource('epub/xsltContent.xsl'))
+    manifest = opf.find('.//' + OPFNS('manifest'))
+    guide = opf.find('.//' + OPFNS('guide'))
+    spine = opf.find('.//' + OPFNS('spine'))

 
 

+    output_file = NamedTemporaryFile(prefix='librarian', suffix='.epub', delete=False)

     zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED)
 
     zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED)
 

+    functions.reg_mathml_epub(zip)
+
+    if os.path.isdir(ilustr_path):
+        for i, filename in enumerate(os.listdir(ilustr_path)):
+            file_path = os.path.join(ilustr_path, filename)
+            zip.write(file_path, os.path.join('OPS', filename))
+            image_id = 'image%s' % i
+            manifest.append(etree.fromstring(
+                '<item id="%s" href="%s" media-type="%s" />' % (image_id, filename, guess_type(file_path)[0])))
+
+    # write static elements

     mime = zipfile.ZipInfo()
     mime.filename = 'mimetype'
     mime.compress_type = zipfile.ZIP_STORED
     mime = zipfile.ZipInfo()
     mime.filename = 'mimetype'
     mime.compress_type = zipfile.ZIP_STORED

-    mime.extra = ''
-    zip.writestr(mime, 'application/epub+zip')
-
-    zip.writestr('META-INF/container.xml', '<?xml version="1.0" ?><container version="1.0" ' \
-                       'xmlns="urn:oasis:names:tc:opendocument:xmlns:container">' \
-                       '<rootfiles><rootfile full-path="OPS/content.opf" ' \
-                       'media-type="application/oebps-package+xml" />' \
-                       '</rootfiles></container>')
-
-    metadata_el = input_xml.find('.//'+RDFNS('Description'))
-    metadatasource = etree.ElementTree(metadata_el)
+    mime.extra = b''
+    zip.writestr(mime, b'application/epub+zip')
+    zip.writestr(
+        'META-INF/container.xml',
+        b'<?xml version="1.0" ?>'
+        b'<container version="1.0" '
+        b'xmlns="urn:oasis:names:tc:opendocument:xmlns:container">'
+        b'<rootfiles><rootfile full-path="OPS/content.opf" '
+        b'media-type="application/oebps-package+xml" />'
+        b'</rootfiles></container>'
+    )
+    zip.write(get_resource('res/wl-logo-small.png'),
+              os.path.join('OPS', 'logo_wolnelektury.png'))
+    zip.write(get_resource('res/jedenprocent.png'),
+              os.path.join('OPS', 'jedenprocent.png'))
+    if not style:
+        style = get_resource('epub/style.css')
+    zip.write(style, os.path.join('OPS', 'style.css'))
+
+    if cover:
+        if cover is True:
+            cover = make_cover
+
+        cover_file = BytesIO()
+        bound_cover = cover(document.book_info)
+        bound_cover.save(cover_file)
+        cover_name = 'cover.%s' % bound_cover.ext()
+        zip.writestr(os.path.join('OPS', cover_name), cover_file.getvalue())
+        del cover_file
+
+        cover_tree = etree.parse(get_resource('epub/cover.html'))
+        cover_tree.find('//' + XHTMLNS('img')).set('src', cover_name)
+        zip.writestr('OPS/cover.html', etree.tostring(
+            cover_tree, pretty_print=True, xml_declaration=True,
+            encoding="utf-8",
+            doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" ' +
+                    '"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+        ))
+
+        if bound_cover.uses_dc_cover:
+            if document.book_info.cover_by:
+                document.edoc.getroot().set('data-cover-by', document.book_info.cover_by)
+            if document.book_info.cover_source:
+                document.edoc.getroot().set('data-cover-source', document.book_info.cover_source)

 
 

-    opf = xslt(metadatasource, res('xsltContent.xsl'))
-
-    manifest = opf.find('.//' + OPFNS('manifest'))
-    spine = opf.find('.//' + OPFNS('spine'))
-
-    for fname in 'style.css', 'logo_wolnelektury.png':
-        zip.write(res(fname), os.path.join('OPS', fname))
+        manifest.append(etree.fromstring(
+            '<item id="cover" href="cover.html" media-type="application/xhtml+xml" />'))
+        manifest.append(etree.fromstring(
+            '<item id="cover-image" href="%s" media-type="%s" />' % (cover_name, bound_cover.mime_type())))
+        spine.insert(0, etree.fromstring('<itemref idref="cover"/>'))
+        opf.getroot()[0].append(etree.fromstring('<meta name="cover" content="cover-image"/>'))
+        guide.append(etree.fromstring('<reference href="cover.html" type="cover" title="Okładka"/>'))

 
     annotations = etree.Element('annotations')
 
     annotations = etree.Element('annotations')

-    part_xml = etree.Element('utwor')
-    etree.SubElement(part_xml, 'master')

 
 

-    toc_file = etree.fromstring('<?xml version="1.0" encoding="utf-8"?><!DOCTYPE ncx PUBLIC ' \
-                               '"-//NISO//DTD ncx 2005-1//EN" "http://www.daisy.org/z3986/2005/ncx-2005-1.dtd">' \
-                               '<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" xml:lang="pl" ' \
-                               'version="2005-1"><head></head><docTitle></docTitle><navMap>' \
-                               '<navPoint id="NavPoint-1" playOrder="1"><navLabel>' \
-                               '<text>Strona tytułowa</text></navLabel><content src="title.html" />' \
-                               '</navPoint><navPoint id="NavPoint-2" playOrder="2"><navLabel>' \
-                               '<text>Początek utworu</text></navLabel><content src="part1.html" />' \
-                               '</navPoint></navMap></ncx>')
-
-    main_xml_part = part_xml[0] # było [0][0], master
-    nav_map = toc_file[-1] # było [-1][-1]
-    depth = 1 # navmap
-
-    if len(input_xml.getroot()) > 1:
-        # rdf before style master
-        main_text = input_xml.getroot()[1]
-    else:
-        # rdf in style master
-        main_text = input_xml.getroot()[0]
-
-    replace_characters(main_text)
-    zip.writestr('OPS/title.html', 
-                 etree.tostring(xslt(input_xml, res('xsltTitle.xsl')), pretty_print=True))
+    toc_file = etree.fromstring(
+        b'<?xml version="1.0" encoding="utf-8"?><!DOCTYPE ncx PUBLIC '
+        b'"-//NISO//DTD ncx 2005-1//EN" '
+        b'"http://www.daisy.org/z3986/2005/ncx-2005-1.dtd">'
+        b'<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" xml:lang="pl" '
+        b'version="2005-1"><head></head><docTitle></docTitle><navMap>'
+        b'</navMap></ncx>'
+    )
+    nav_map = toc_file[-1]
+
+    if html_toc:
+        manifest.append(etree.fromstring(
+            '<item id="html_toc" href="toc.html" media-type="application/xhtml+xml" />'))
+        spine.append(etree.fromstring(
+            '<itemref idref="html_toc" />'))
+        guide.append(etree.fromstring('<reference href="toc.html" type="toc" title="Spis treści"/>'))

 
 

-    # Search for table of contents elements and book division
+    toc, chunk_counter, chars, sample = transform_file(document, sample=sample)

 
 

-    stupid_i = stupid_j = stupid_k = 1
-    last_node_part = False
-    for one_part in main_text:
-        name = one_part.tag
-        if name in ("naglowek_czesc", "naglowek_rozdzial", "naglowek_akt", "srodtytul"):
-            if name == "naglowek_czesc":
-                stupid_k = 1
-                last_node_part = True
-                find_annotations(annotations, part_xml, stupid_j)
-                replace_by_verse(part_xml)
-                zip.writestr('OPS/part%d.html' % stupid_j,
-                            etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
-                main_xml_part[:] = [deepcopy(one_part)]
-                # add to manifest and spine
-                add_to_manifest(manifest, stupid_j)
-                add_to_spine(spine, stupid_j)
-                name_toc = node_name(one_part)
-                # build table of contents
-                # i+2 because of title page
-                add_nav_point(nav_map, stupid_i+2, name_toc, stupid_j + 1)
-                stupid_i += 1
-                stupid_j += 1
-            else:
-                if last_node_part:
-                    main_xml_part.append(one_part)
-                    last_node_part = False
-                    name_toc = node_name(one_part)
-                    add_nav_point(nav_map, stupid_i + 1, name_toc, stupid_j)
-                else:
-                    stupid_k = 1
-                    find_annotations(annotations, part_xml, stupid_j)
-                    replace_by_verse(part_xml)
-                    zip.writestr('OPS/part%d.html' % stupid_j,
-                                 etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
-                    # start building a new part
-                    main_xml_part[:] = [deepcopy(one_part)]
-                    add_to_manifest(manifest, stupid_j)
-                    add_to_spine(spine, stupid_j)
-                    name_toc = node_name(one_part)
-                    add_nav_point(nav_map, stupid_i + 2, name_toc, stupid_j + 1) # title page
-                    stupid_j += 1
-                    stupid_i += 1
-        else:
-            if name in ('naglowek_podrozdzial', 'naglowek_scena'):
-                depth = 2
-                name_toc =  node_name(one_part)
-                add_nav_point2(nav_map, stupid_i + 2, name_toc, stupid_j, stupid_k)
-                one_part.set('sub', str(stupid_k))
-                stupid_k += 1
-                stupid_i += 1
-            main_xml_part.append(deepcopy(one_part))
-            last_node_part = False
-    find_annotations(annotations, part_xml, stupid_j)
-    replace_by_verse(part_xml)
-    add_to_manifest(manifest, stupid_j)
-    add_to_spine(spine, stupid_j)
-
-    zip.writestr('OPS/part%d.html' % stupid_j,
-                 etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
+    if len(toc.children) < 2:
+        toc.add(u"Początek utworu", "part1.html")

 
     # Last modifications in container files and EPUB creation
     if len(annotations) > 0:
 
     # Last modifications in container files and EPUB creation
     if len(annotations) > 0:

-        nav_map.append(etree.fromstring(
-            '<navPoint id="NavPoint-%(i)d" playOrder="%(i)d" ><navLabel><text>Przypisy</text>'\
-            '</navLabel><content src="annotations.html" /></navPoint>' % {'i':stupid_i+2}))
+        toc.add("Przypisy", "annotations.html")

         manifest.append(etree.fromstring(
             '<item id="annotations" href="annotations.html" media-type="application/xhtml+xml" />'))
         spine.append(etree.fromstring(
             '<itemref idref="annotations" />'))
         replace_by_verse(annotations)
         manifest.append(etree.fromstring(
             '<item id="annotations" href="annotations.html" media-type="application/xhtml+xml" />'))
         spine.append(etree.fromstring(
             '<itemref idref="annotations" />'))
         replace_by_verse(annotations)

+        html_tree = xslt(annotations, get_resource('epub/xsltAnnotations.xsl'))
+        chars = chars.union(used_chars(html_tree.getroot()))

         zip.writestr('OPS/annotations.html', etree.tostring(
         zip.writestr('OPS/annotations.html', etree.tostring(

-                            xslt(annotations, res("xsltAnnotations.xsl")), pretty_print=True))
-
-    zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True))
-    contents = []
-    title = node_name(etree.ETXPath('.//'+DCNS('title'))(input_xml)[0])
+            html_tree, pretty_print=True, xml_declaration=True,
+            encoding="utf-8",
+            doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" ' +
+                    '"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+        ))
+
+    toc.add("Wesprzyj Wolne Lektury", "support.html")
+    manifest.append(etree.fromstring(
+        '<item id="support" href="support.html" media-type="application/xhtml+xml" />'))
+    spine.append(etree.fromstring(
+        '<itemref idref="support" />'))
+    html_string = open(get_resource('epub/support.html'), 'rb').read()
+    chars.update(used_chars(etree.fromstring(html_string)))
+    zip.writestr('OPS/support.html', squeeze_whitespace(html_string))
+
+    toc.add("Strona redakcyjna", "last.html")
+    manifest.append(etree.fromstring(
+        '<item id="last" href="last.html" media-type="application/xhtml+xml" />'))
+    spine.append(etree.fromstring(
+        '<itemref idref="last" />'))
+    html_tree = xslt(document.edoc, get_resource('epub/xsltLast.xsl'), outputtype=output_type)
+    chars.update(used_chars(html_tree.getroot()))
+    zip.writestr('OPS/last.html', squeeze_whitespace(etree.tostring(
+        html_tree, pretty_print=True, xml_declaration=True,
+        encoding="utf-8",
+        doctype='<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" ' +
+                '"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">'
+    )))
+
+    if not flags or 'without-fonts' not in flags:
+        # strip fonts
+        tmpdir = mkdtemp('-librarian-epub')
+        try:
+            cwd = os.getcwd()
+        except OSError:
+            cwd = None
+
+        os.chdir(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'font-optimizer'))
+        for fname in 'DejaVuSerif.ttf', 'DejaVuSerif-Bold.ttf', 'DejaVuSerif-Italic.ttf', 'DejaVuSerif-BoldItalic.ttf':
+            optimizer_call = ['perl', 'subset.pl', '--chars',
+                              ''.join(chars).encode('utf-8'),
+                              get_resource('fonts/' + fname),
+                              os.path.join(tmpdir, fname)]
+            env = {"PERL_USE_UNSAFE_INC": "1"}
+            if verbose:
+                print("Running font-optimizer")
+                subprocess.check_call(optimizer_call, env=env)
+            else:
+                dev_null = open(os.devnull, 'w')
+                subprocess.check_call(optimizer_call, stdout=dev_null, stderr=dev_null, env=env)
+            zip.write(os.path.join(tmpdir, fname), os.path.join('OPS', fname))
+            manifest.append(etree.fromstring(
+                '<item id="%s" href="%s" media-type="application/x-font-truetype" />' % (fname, fname)))
+        rmtree(tmpdir)
+        if cwd is not None:
+            os.chdir(cwd)
+    zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True,
+                 xml_declaration=True, encoding="utf-8"))
+    title = document.book_info.title

     attributes = "dtb:uid", "dtb:depth", "dtb:totalPageCount", "dtb:maxPageNumber"
     for st in attributes:
         meta = toc_file.makeelement(NCXNS('meta'))
         meta.set('name', st)
         meta.set('content', '0')
         toc_file[0].append(meta)
     attributes = "dtb:uid", "dtb:depth", "dtb:totalPageCount", "dtb:maxPageNumber"
     for st in attributes:
         meta = toc_file.makeelement(NCXNS('meta'))
         meta.set('name', st)
         meta.set('content', '0')
         toc_file[0].append(meta)

-    toc_file[0][0].set('content', ''.join((title, 'WolneLektury.pl')))
-    toc_file[0][1].set('content', str(depth))
+    toc_file[0][0].set('content', str(document.book_info.url))
+    toc_file[0][1].set('content', str(toc.depth()))

     set_inner_xml(toc_file[1], ''.join(('<text>', title, '</text>')))
     set_inner_xml(toc_file[1], ''.join(('<text>', title, '</text>')))

-    zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True))
-    zip.close()
-
-
-if __name__ == '__main__':
-    import html
-
-    if len(sys.argv) < 2:
-        print >> sys.stderr, 'Usage: wl2epub <input file> [output file]'
-        sys.exit(1)
-
-    input = sys.argv[1]
-    if len(sys.argv) > 2:
-        output = sys.argv[2]
-    else:
-        basename, ext = os.path.splitext(input)
-        output = basename + '.epub' 
-
-    print input
-    if html.transform(input, is_file=True) == '<empty />':
-        print 'empty content - skipping'
-    transform(open(input, 'r'), open(output, 'w'))
-

 
 

+    # write TOC
+    if html_toc:
+        toc.add(u"Spis treści", "toc.html", index=1)
+        zip.writestr('OPS/toc.html', toc.html().encode('utf-8'))
+    toc.write_to_xml(nav_map)
+    zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True,
+                 xml_declaration=True, encoding="utf-8"))
+    zip.close()

 
 

+    return OutputFile.from_filename(output_file.name)