X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/e57b146bf49e38b3bb57615110b27de5b4d1ae69..a4d6831b455ed69b196905427600d7163cdd3fe8:/librarian/epub.py diff --git a/librarian/epub.py b/librarian/epub.py index bfd7570..e52e524 100644 --- a/librarian/epub.py +++ b/librarian/epub.py @@ -16,14 +16,51 @@ import zipfile from tempfile import mkdtemp, NamedTemporaryFile from shutil import rmtree -from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, OutputFile -from librarian.cover import WLCover +from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, DCNS, OutputFile +from librarian.cover import DefaultEbookCover from librarian import functions, get_resource -functions.reg_person_name() - +from librarian.hyphenator import Hyphenator +functions.reg_person_name() +functions.reg_lang_code_3to2() + +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: + 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): + if hyph is not None: + texts = etree.XPath('/utwor/*[2]//text()')(source_tree) + for t in texts: + parent = t.getparent() + 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 + def inner_xml(node): """ returns node's text and children as a string @@ -247,7 +284,10 @@ class TOC(object): nav_label = nav_map.makeelement(NCXNS('navLabel')) text = nav_map.makeelement(NCXNS('text')) - text.text = child.name + 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) @@ -290,18 +330,41 @@ def chop(main_text): 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_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)] + 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: - main_xml_part.append(deepcopy(one_part)) - last_node_part = False + 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 @@ -310,7 +373,9 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s toc = TOC() for element in chunk_xml[0]: - if element.tag in ("naglowek_czesc", "naglowek_rozdzial", "naglowek_akt", "srodtytul"): + 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) @@ -325,7 +390,11 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s 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) + output_html = etree.tostring(html_tree, pretty_print = True, + xml_declaration = True, + encoding = "utf-8", + doctype='') return output_html, toc, chars @@ -343,7 +412,11 @@ def transform(wldoc, verbose=False, """ processes one input file and proceeds to its children """ replace_characters(wldoc.edoc.getroot()) - + + hyphenator = set_hyph_language(wldoc.edoc.getroot()) + 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) @@ -353,7 +426,11 @@ def transform(wldoc, verbose=False, 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)) + etree.tostring(html_tree, pretty_print = True, + xml_declaration = True, + encoding = "utf-8", + doctype='')) # add a title page TOC entry toc.add(u"Strona tytułowa", "title.html") elif wldoc.book_info.parts: @@ -364,7 +441,12 @@ def transform(wldoc, verbose=False, 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) + html_string = etree.tostring(html_tree, + pretty_print = True, + xml_declaration = True, + encoding = "utf-8", + doctype='') zip.writestr('OPS/part%d.html' % chunk_counter, html_string) add_to_manifest(manifest, chunk_counter) add_to_spine(spine, chunk_counter) @@ -448,7 +530,7 @@ def transform(wldoc, verbose=False, if cover: if cover is True: - cover = WLCover + cover = DefaultEbookCover cover_file = StringIO() bound_cover = cover(document.book_info) @@ -460,7 +542,9 @@ def transform(wldoc, verbose=False, 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, method="html", pretty_print=True)) + cover_tree, pretty_print = True, xml_declaration = True, encoding = "utf-8", + doctype='')) if bound_cover.uses_dc_cover: if document.book_info.cover_by: @@ -472,7 +556,7 @@ def transform(wldoc, verbose=False, '')) manifest.append(etree.fromstring( '' % (cover_name, bound_cover.mime_type()))) - spine.insert(0, etree.fromstring('')) + spine.insert(0, etree.fromstring('')) opf.getroot()[0].append(etree.fromstring('')) guide.append(etree.fromstring('')) @@ -509,7 +593,11 @@ def transform(wldoc, verbose=False, html_tree = xslt(annotations, get_resource('epub/xsltAnnotations.xsl')) chars = chars.union(used_chars(html_tree.getroot())) zip.writestr('OPS/annotations.html', etree.tostring( - html_tree, method="html", pretty_print=True)) + html_tree, pretty_print = True, + xml_declaration = True, + encoding = "utf-8", + doctype='')) toc.add("Wesprzyj Wolne Lektury", "support.html") manifest.append(etree.fromstring( @@ -528,7 +616,11 @@ def transform(wldoc, verbose=False, 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)) + html_tree, pretty_print = True, + xml_declaration = True, + encoding = "utf-8", + doctype='')) if not flags or not 'without-fonts' in flags: # strip fonts @@ -541,7 +633,7 @@ def transform(wldoc, verbose=False, 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)] + get_resource('fonts/' + fname), os.path.join(tmpdir, fname)] if verbose: print "Running font-optimizer" subprocess.check_call(optimizer_call) @@ -549,12 +641,13 @@ def transform(wldoc, verbose=False, 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))) + '' % (fname, fname))) rmtree(tmpdir) if cwd is not None: os.chdir(cwd) - - zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True)) + 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: @@ -562,7 +655,7 @@ def transform(wldoc, verbose=False, 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][0].set('content', str(document.book_info.url)) toc_file[0][1].set('content', str(toc.depth())) set_inner_xml(toc_file[1], ''.join(('', title, ''))) @@ -571,7 +664,9 @@ def transform(wldoc, verbose=False, 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.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)