X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/8550d172b829f29b2fcd4723789fb2a5d9fff6eb..ac9ba215113d29655e142f23ab398c5c2765df1e:/src/librarian/epub.py?ds=sidebyside
diff --git a/src/librarian/epub.py b/src/librarian/epub.py
index 137796e..7dbfc49 100644
--- a/src/librarian/epub.py
+++ b/src/librarian/epub.py
@@ -1,174 +1,14 @@
-# -*- 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 Wolne Lektury. See NOTICE for more information.
#
-from __future__ import print_function, unicode_literals
-
-import os
-import os.path
import re
-import subprocess
-from six import BytesIO
from copy import deepcopy
-from mimetypes import guess_type
-
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.hyphenator import Hyphenator
-
-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
-
- >>> print(inner_xml(etree.fromstring('xyz')))
- xyz
- """
-
- nt = node.text if node.text is not None else ''
- 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('bxx')
- >>> set_inner_xml(e, 'xyz')
- >>> print(etree.tostring(e, encoding='unicode'))
- xyz
- """
-
- p = etree.fromstring('%s' % text)
- node.text = p.text
- node[:] = p[:]
-
-
-def node_name(node):
- """ Find out a node's name
-
- >>> print(node_name(etree.fromstring('XYZ')))
- XYZ
- """
-
- tempnode = deepcopy(node)
- for p in ('pe', 'pa', 'pt', 'pr', 'motyw'):
- for e in tempnode.findall('.//%s' % p):
- t = e.tail
- e.clear()
- e.tail = t
- etree.strip_tags(tempnode, '*')
- return tempnode.text
+from librarian import WLNS
-def xslt(xml, sheet, **kwargs):
- if isinstance(xml, etree._Element):
- xml = etree.ElementTree(xml)
- with open(sheet) as xsltf:
- 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
- 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.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_no):
- for child in source:
- if child.tag in ('pe', 'pa', 'pt', 'pr'):
- annotation = deepcopy(child)
- 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
- child.text = number
- if child.tag not in ('extra', 'uwaga'):
- find_annotations(annotations, child, part_no)
-
-
-class Stanza(object):
+class Stanza:
"""
Converts / verse endings into verse elements in a stanza.
@@ -247,543 +87,3 @@ def replace_by_verse(tree):
stanzas = tree.findall('.//' + WLNS('strofa'))
for stanza in stanzas:
Stanza(stanza).versify()
-
-
-def add_to_manifest(manifest, partno):
- """ Adds a node to the manifest section in content.opf file """
-
- partstr = 'part%d' % partno
- 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 """
-
- e = spine.makeelement(
- OPFNS('itemref'),
- attrib={'idref': 'part%d' % partno}
- )
- spine.append(e)
-
-
-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(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
-
- def append(self, toc):
- self.children.append(toc)
-
- def extend(self, toc):
- self.children.extend(toc.children)
-
- 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(
- "" %
- (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
-
- # 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
-
- 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:
- with open(get_resource('epub/emptyChunk.html')) as f:
- _empty_html_static.append(f.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=''
- )
- 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
- """
-
- 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=''
- )
- zip.writestr('OPS/title.html', squeeze_whitespace(html_string))
- # add a title page TOC entry
- toc.add(u"Strona tytuÅowa", "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=''
- )
- 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)
-
- functions.reg_mathml_epub(zip)
-
- if os.path.isdir(ilustr_path):
- ilustr_elements = set(ilustr.get('src')
- for ilustr in document.edoc.findall('//ilustr'))
- for i, filename in enumerate(os.listdir(ilustr_path)):
- if filename not in ilustr_elements:
- continue
- 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(
- '