X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/33db54c61d9818547c059e3e3cda2518a260bcda..edbda465ca1fceabe219d5caea95d513d61c6976:/librarian/epub.py?ds=inline diff --git a/librarian/epub.py b/librarian/epub.py index cb2166b..b063380 100644 --- a/librarian/epub.py +++ b/librarian/epub.py @@ -1,22 +1,25 @@ # -*- 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 import os import os.path -import shutil -import sys +import subprocess +from StringIO import StringIO from copy import deepcopy from lxml import etree import zipfile +from tempfile import mkdtemp, NamedTemporaryFile +from shutil import rmtree -from librarian import XMLNamespace, RDFNS, DCNS, WLNS, XHTMLNS +from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, OutputFile -NCXNS = XMLNamespace("http://www.daisy.org/z3986/2005/ncx/") -OPFNS = XMLNamespace("http://www.idpf.org/2007/opf") +from librarian import functions, get_resource + +functions.reg_person_name() def inner_xml(node): @@ -27,7 +30,7 @@ def inner_xml(node): """ 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) for child in node]) def set_inner_xml(node, text): """ sets node's text and children from a string @@ -38,7 +41,6 @@ def set_inner_xml(node, text): xyz """ - p = etree.fromstring('%s' % text) node.text = p.text node[:] = p[:] @@ -69,48 +71,46 @@ def xslt(xml, sheet): 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) - - def replace_characters(node): def replace_chars(text): if text is None: return None - return text.replace("&", "&")\ - .replace("---", "—")\ - .replace("--", "–")\ - .replace(",,", "„")\ - .replace('"', "”")\ - .replace("'", "’") - 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() - 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) - 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 - 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) def replace_by_verse(tree): """ Find stanzas and create new verses in place of a '/' character """ + stanzas = tree.findall('.//' + WLNS('strofa')) for node in stanzas: for child_node in node: @@ -137,6 +137,7 @@ def replace_by_verse(tree): 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, @@ -148,186 +149,338 @@ def add_to_manifest(manifest, partno): 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) +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 - content = nav_map.makeelement(NCXNS('content')) - content.set('src', 'part%d.html' % part_counter) - nav_point.append(content) - - nav_map.append(nav_point) - - -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)) - - 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#sub%d' % (part_counter, subcounter)) - nav_point.append(content) + 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_map[-1].append(nav_point) + def append(self, toc): + self.children.append(toc) + def extend(self, toc): + self.children.extend(toc.children) -def transform(input_file, output_file): - """ produces an epub + 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')) + 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( + "
%s
" % + (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')) as f: + t = unicode(f.read(), '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 - input_file and output_file should be filelike objects + last_node_part = False + for one_part in main_text: + name = one_part.tag + 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 in ("naglowek_czesc", "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, method="html", pretty_print=True) + return output_html, toc, chars + + +def transform(wldoc, verbose=False, + style=None, html_toc=False, + sample=None, cover=None, flags=None): + """ produces a EPUB file + + sample=n: generate sample e-book (with at least n paragraphs) + cover: a cover.Cover object + flags: less-advertising, without-fonts """ - 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()) + + # 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')) + chars = used_chars(html_tree.getroot()) + zip.writestr('OPS/title.html', + etree.tostring(html_tree, method="html", pretty_print=True)) + # 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, method="html", pretty_print=True) + zip.writestr('OPS/part%d.html' % chunk_counter, 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, 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') + + 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) + # write static elements 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', '' \ '' \ '') + 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: + cover_file = StringIO() + c = cover(document.book_info.author.readable(), document.book_info.title) + c.save(cover_file) + c_name = 'cover.%s' % c.ext() + zip.writestr(os.path.join('OPS', c_name), cover_file.getvalue()) + del cover_file + + cover_tree = etree.parse(get_resource('epub/cover.html')) + cover_tree.find('//' + XHTMLNS('img')).set('src', c_name) + zip.writestr('OPS/cover.html', etree.tostring( + cover_tree, method="html", pretty_print=True)) - metadata_el = input_xml.find('.//'+RDFNS('Description')) - metadatasource = etree.ElementTree(metadata_el) - - opf = xslt(metadatasource, res('xsltContent.xsl')) - - manifest = opf.find('.//' + OPFNS('manifest')) - spine = opf.find('.//' + OPFNS('spine')) + manifest.append(etree.fromstring( + '')) + manifest.append(etree.fromstring( + '' % (c_name, c.mime_type()))) + spine.insert(0, etree.fromstring('')) + opf.getroot()[0].append(etree.fromstring('')) + guide.append(etree.fromstring('')) - for fname in 'style.css', 'logo_wolnelektury.png': - zip.write(res(fname), os.path.join('OPS', fname)) annotations = etree.Element('annotations') - part_xml = etree.Element('utwor') - etree.SubElement(part_xml, 'master') toc_file = etree.fromstring('' \ '' \ - '' \ - 'Strona tytułowa' \ - '' \ - 'Początek utworu' \ - '') - - 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] + '') + nav_map = toc_file[-1] - replace_characters(main_text) - zip.writestr('OPS/title.html', - etree.tostring(xslt(input_xml, res('xsltTitle.xsl')), pretty_print=True)) - - # Search for table of contents elements and book division + if html_toc: + manifest.append(etree.fromstring( + '')) + spine.append(etree.fromstring( + '')) + guide.append(etree.fromstring('')) - 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) + toc, chunk_counter, chars, sample = transform_file(document, sample=sample) - 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: - nav_map.append(etree.fromstring( - 'Przypisy'\ - '' % {'i':stupid_i+2})) + toc.add("Przypisy", "annotations.html") manifest.append(etree.fromstring( '')) spine.append(etree.fromstring( '')) 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( - xslt(annotations, res("xsltAnnotations.xsl")), pretty_print=True)) + html_tree, method="html", pretty_print=True)) + + toc.add("Strona redakcyjna", "last.html") + manifest.append(etree.fromstring( + '')) + spine.append(etree.fromstring( + '')) + html_tree = xslt(document.edoc, get_resource('epub/xsltLast.xsl')) + chars.update(used_chars(html_tree.getroot())) + zip.writestr('OPS/last.html', etree.tostring( + html_tree, method="html", pretty_print=True)) + + if not flags or not 'without-fonts' in flags: + # strip fonts + tmpdir = mkdtemp('-librarian-epub') + cwd = os.getcwd() + + 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)] + if verbose: + print "Running font-optimizer" + subprocess.check_call(optimizer_call) + else: + subprocess.check_call(optimizer_call, stdout=subprocess.PIPE, stderr=subprocess.PIPE) + zip.write(os.path.join(tmpdir, fname), os.path.join('OPS', fname)) + manifest.append(etree.fromstring( + '' % (fname, fname))) + rmtree(tmpdir) + os.chdir(cwd) zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True)) - contents = [] - title = node_name(etree.ETXPath('.//'+DCNS('title'))(input_xml)[0]) + title = document.book_info.title attributes = "dtb:uid", "dtb:depth", "dtb:totalPageCount", "dtb:maxPageNumber" for st in attributes: meta = toc_file.makeelement(NCXNS('meta')) @@ -335,31 +488,15 @@ def transform(input_file, output_file): 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][1].set('content', str(toc.depth())) set_inner_xml(toc_file[1], ''.join(('', title, ''))) + + # 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)) zip.close() - -if __name__ == '__main__': - import html - - if len(sys.argv) < 2: - print >> sys.stderr, 'Usage: wl2epub [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) == '': - print 'empty content - skipping' - else: - transform(open(input, 'r'), open(output, 'w')) - - - + return OutputFile.from_filename(output_file.name)