X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/5f92f5a341c58a30c0a5c0d64ba3ed498bb9db85..2ada86c7bb585d419b95c95ce42d5b7a9fa45475:/librarian/epub.py diff --git a/librarian/epub.py b/librarian/epub.py index bf58a9f..be9488a 100644 --- a/librarian/epub.py +++ b/librarian/epub.py @@ -3,21 +3,23 @@ # This file is part of Librarian, licensed under GNU Affero GPLv3 or later. # 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 re import subprocess -from StringIO import StringIO +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 DefaultEbookCover +from librarian.cover import make_cover from librarian import functions, get_resource @@ -27,12 +29,16 @@ 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: + for line in f.read().decode('latin1').split('\n'): list = line.strip().split('|') if list[0] == text: result = list[2] @@ -51,30 +57,32 @@ def set_hyph_language(source_tree): def hyphenate_and_fix_conjunctions(source_tree, hyph): - if hyph is not None: - texts = etree.XPath('/utwor/*[2]//text()')(source_tree) - for t in texts: - parent = t.getparent() + 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') - newt = re.sub(r'(?<=\s\w)\s+', u'\u00A0', newt) - if t.is_text: - parent.text = newt - elif t.is_tail: - parent.tail = newt + 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')) + >>> print(inner_xml(etree.fromstring('xyz'))) xyz """ 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): @@ -82,7 +90,7 @@ def set_inner_xml(node, text): >>> e = etree.fromstring('bxx') >>> set_inner_xml(e, 'xyz') - >>> print etree.tostring(e) + >>> print(etree.tostring(e, encoding='unicode')) xyz """ @@ -94,7 +102,7 @@ def set_inner_xml(node, text): def node_name(node): """ Find out a node's name - >>> print node_name(etree.fromstring('XYZ')) + >>> print(node_name(etree.fromstring('XYZ'))) XYZ """ @@ -109,11 +117,13 @@ def node_name(node): 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: - return xml.xslt(etree.parse(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): @@ -162,8 +172,8 @@ class Stanza(object): >>> s = etree.fromstring("a c c/\\nbx/\\nyc/ \\nd") >>> Stanza(s).versify() - >>> print etree.tostring(s) - a c cbx/ + >>> print(etree.tostring(s, encoding='unicode')) + a ccbx/ ycd """ @@ -180,7 +190,7 @@ class Stanza(object): tail = self.stanza.tail self.stanza.clear() self.stanza.tail = tail - self.stanza.extend(self.verses) + 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") @@ -197,6 +207,8 @@ class Stanza(object): 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 @@ -313,8 +325,8 @@ class TOC(object): return "\n".join(texts) def html(self): - with open(get_resource('epub/toc.html')) as f: - t = unicode(f.read(), 'utf-8') + with open(get_resource('epub/toc.html'), 'rb') as f: + t = f.read().decode('utf-8') return t % self.html_part() @@ -404,9 +416,8 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s return output_html, toc, chars -def transform(wldoc, verbose=False, - style=None, html_toc=False, - sample=None, cover=None, flags=None): +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) @@ -419,7 +430,7 @@ def transform(wldoc, verbose=False, replace_characters(wldoc.edoc.getroot()) - hyphenator = set_hyph_language(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, @@ -428,17 +439,15 @@ def transform(wldoc, verbose=False, chars = set() if first: # write book title page - html_tree = xslt(wldoc.edoc, get_resource('epub/xsltTitle.xsl')) + html_tree = xslt(wldoc.edoc, get_resource('epub/xsltTitle.xsl'), outputtype=output_type) chars = used_chars(html_tree.getroot()) - zip.writestr( - 'OPS/title.html', - etree.tostring( - html_tree, pretty_print=True, xml_declaration=True, - encoding="utf-8", - doctype='' - ) + 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 tytułowa", "title.html") elif wldoc.book_info.parts: @@ -455,7 +464,7 @@ def transform(wldoc, verbose=False, doctype='' ) - zip.writestr('OPS/part%d.html' % chunk_counter, html_string) + 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 @@ -481,7 +490,7 @@ def transform(wldoc, verbose=False, toc.extend(chunk_toc) chars = chars.union(chunk_chars) - zip.writestr('OPS/part%d.html' % chunk_counter, chunk_html) + 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 @@ -501,6 +510,9 @@ def transform(wldoc, verbose=False, 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: @@ -522,20 +534,31 @@ def transform(wldoc, verbose=False, 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( + '' % (image_id, filename, guess_type(file_path)[0]))) + # write static elements mime = zipfile.ZipInfo() mime.filename = 'mimetype' mime.compress_type = zipfile.ZIP_STORED - mime.extra = '' - zip.writestr(mime, 'application/epub+zip') + mime.extra = b'' + zip.writestr(mime, b'application/epub+zip') zip.writestr( 'META-INF/container.xml', - '' - '' - '' - '' + b'' + b'' + b'' + b'' ) zip.write(get_resource('res/wl-logo-small.png'), os.path.join('OPS', 'logo_wolnelektury.png')) @@ -547,9 +570,9 @@ def transform(wldoc, verbose=False, if cover: if cover is True: - cover = DefaultEbookCover + cover = make_cover - cover_file = StringIO() + cover_file = BytesIO() bound_cover = cover(document.book_info) bound_cover.save(cover_file) cover_name = 'cover.%s' % bound_cover.ext() @@ -582,12 +605,12 @@ def transform(wldoc, verbose=False, annotations = etree.Element('annotations') toc_file = etree.fromstring( - '' - '' - '' + b'' + b'' + b'' ) nav_map = toc_file[-1] @@ -625,25 +648,25 @@ def transform(wldoc, verbose=False, '')) spine.append(etree.fromstring( '')) - html_string = open(get_resource('epub/support.html')).read() + html_string = open(get_resource('epub/support.html'), 'rb').read() chars.update(used_chars(etree.fromstring(html_string))) - zip.writestr('OPS/support.html', html_string) + zip.writestr('OPS/support.html', squeeze_whitespace(html_string)) toc.add("Strona redakcyjna", "last.html") manifest.append(etree.fromstring( '')) spine.append(etree.fromstring( '')) - html_tree = xslt(document.edoc, get_resource('epub/xsltLast.xsl')) + 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', etree.tostring( + zip.writestr('OPS/last.html', squeeze_whitespace(etree.tostring( html_tree, pretty_print=True, xml_declaration=True, encoding="utf-8", doctype='' - )) + ))) - if not flags or not 'without-fonts' in flags: + if not flags or 'without-fonts' not in flags: # strip fonts tmpdir = mkdtemp('-librarian-epub') try: @@ -657,11 +680,13 @@ def transform(wldoc, verbose=False, ''.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) + print("Running font-optimizer") + subprocess.check_call(optimizer_call, env=env) else: - subprocess.check_call(optimizer_call, stdout=subprocess.PIPE, stderr=subprocess.PIPE) + 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( '' % (fname, fname)))