def convert(self, stream, options, file_ext, log, accelerators): from ebook_converter.ebooks.metadata.toc import TOC from ebook_converter.ebooks.metadata.opf2 import OPFCreator from ebook_converter.utils.zipfile import ZipFile self.options = options self.log = log pages, images = [], [] toc = TOC() if file_ext == 'pmlz': log.debug('De-compressing content to temporary directory...') with TemporaryDirectory('_unpmlz') as tdir: zf = ZipFile(stream) zf.extractall(tdir) pmls = glob.glob(os.path.join(tdir, '*.pml')) for pml in pmls: html_name = os.path.splitext( os.path.basename(pml))[0] + '.html' html_path = os.path.join(os.getcwd(), html_name) pages.append(html_name) log.debug('Processing PML item %s...', pml) ttoc = self.process_pml(pml, html_path) toc += ttoc images = self.get_images(stream, tdir, True) else: toc = self.process_pml(stream, 'index.html') pages.append('index.html') if hasattr(stream, 'name'): images = self.get_images( stream, os.path.abspath(os.path.dirname(stream.name))) # We want pages to be orded alphabetically. pages.sort() manifest_items = [] for item in pages + images: manifest_items.append((item, None)) from ebook_converter.ebooks.metadata.meta import get_metadata log.debug('Reading metadata from input file...') mi = get_metadata(stream, 'pml') if 'images/cover.png' in images: mi.cover = 'images/cover.png' opf = OPFCreator(os.getcwd(), mi) log.debug('Generating manifest...') opf.create_manifest(manifest_items) opf.create_spine(pages) opf.set_toc(toc) with open('metadata.opf', 'wb') as opffile: with open('toc.ncx', 'wb') as tocfile: opf.render(opffile, tocfile, 'toc.ncx') return os.path.join(os.getcwd(), 'metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators): from ebook_converter.ebooks.metadata.opf2 import OPFCreator from ebook_converter.ebooks.pdf.pdftohtml import pdftohtml log.debug('Converting file to html...') # The main html file will be named index.html self.opts, self.log = options, log if options.new_pdf_engine: return self.convert_new(stream, accelerators) pdftohtml(os.getcwd(), stream.name, options.no_images) from ebook_converter.ebooks.metadata.meta import get_metadata log.debug('Retrieving document metadata...') mi = get_metadata(stream, 'pdf') opf = OPFCreator(os.getcwd(), mi) manifest = [('index.html', None)] images = os.listdir(os.getcwd()) images.remove('index.html') for i in images: manifest.append((i, None)) log.debug('Generating manifest...') opf.create_manifest(manifest) opf.create_spine(['index.html']) log.debug('Rendering manifest...') with open('metadata.opf', 'wb') as opffile: opf.render(opffile) if os.path.exists('toc.ncx'): ncxid = opf.manifest.id_for_path('toc.ncx') if ncxid: with open('metadata.opf', 'r+b') as f: raw = f.read().replace( b'<spine', b'<spine toc="%s"' % polyglot.as_bytes(ncxid)) f.seek(0) f.write(raw) return os.path.join(os.getcwd(), 'metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators): from ebook_converter.ebooks.metadata.meta import get_metadata from ebook_converter.ebooks.metadata.opf2 import OPFCreator from ebook_converter.ebooks.rtf2xml.ParseRtf import \ RtfInvalidCodeException from ebook_converter.ebooks.rtf.input import InlineClass self.opts = options self.log = log self.log('Converting RTF to XML...') try: xml = self.generate_xml(stream.name) except RtfInvalidCodeException as e: self.log.exception('Unable to parse RTF') raise ValueError('This RTF file has a feature calibre does not ' 'support. Convert it to HTML first and then try ' 'it.\n%s' % e) d = glob.glob(os.path.join('*_rtf_pict_dir', 'picts.rtf')) if d: imap = {} try: imap = self.extract_images(d[0]) except Exception: self.log.exception('Failed to extract images...') self.log('Parsing XML...') doc = etree.fromstring(xml) border_styles = self.convert_borders(doc) for pict in doc.xpath( '//rtf:pict[@num]', namespaces={'rtf': 'http://rtf2xml.sourceforge.net/'}): num = int(pict.get('num')) name = imap.get(num, None) if name is not None: pict.set('num', name) self.log('Converting XML to HTML...') inline_class = InlineClass(self.log) with open( pkg_resources.resource_filename('ebook_converter', 'data/rtf.xsl')) as fobj: styledoc = etree.fromstring(fobj.read()) extensions = {('calibre', 'inline-class'): inline_class} transform = etree.XSLT(styledoc, extensions=extensions) result = transform(doc) html = u'index.xhtml' with open(html, 'wb') as f: res = as_bytes(transform.tostring(result)) # res = res[:100].replace('xmlns:html', 'xmlns') + res[100:] # clean multiple \n res = re.sub(b'\n+', b'\n', res) # Replace newlines inserted by the 'empty_paragraphs' option in # rtf2xml with html blank lines # res = re.sub('\s*<body>', '<body>', res) # res = re.sub('(?<=\n)\n{2}', # u'<p>\u00a0</p>\n'.encode('utf-8'), res) f.write(res) self.write_inline_css(inline_class, border_styles) stream.seek(0) mi = get_metadata(stream, 'rtf') if not mi.title: mi.title = 'Unknown' if not mi.authors: mi.authors = ['Unknown'] opf = OPFCreator(os.getcwd(), mi) opf.create_manifest([(u'index.xhtml', None)]) opf.create_spine([u'index.xhtml']) opf.render(open(u'metadata.opf', 'wb')) return os.path.abspath(u'metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators): from ebook_converter.ebooks.metadata.fb2 import ensure_namespace from ebook_converter.ebooks.metadata.fb2 import get_fb2_data from ebook_converter.ebooks.metadata.opf2 import OPFCreator from ebook_converter.ebooks.metadata.meta import get_metadata from ebook_converter.ebooks.chardet import xml_to_unicode self.log = log log.debug('Parsing XML...') raw = get_fb2_data(stream)[0] raw = raw.replace(b'\0', b'') raw = xml_to_unicode(raw, strip_encoding_pats=True, assume_utf8=True, resolve_entities=True)[0] try: doc = etree.fromstring(raw) except etree.XMLSyntaxError: doc = etree.fromstring(raw.replace('& ', '&')) if doc is None: raise ValueError('The FB2 file is not valid XML') doc = ensure_namespace(doc) try: fb_ns = doc.nsmap[doc.prefix] except Exception: fb_ns = FB2NS NAMESPACES = {'f': fb_ns, 'l': const.XLINK_NS} stylesheets = doc.xpath('//*[local-name() = "stylesheet" and ' '@type="text/css"]') css = '' for s in stylesheets: css += etree.tostring( s, encoding='unicode', method='text', with_tail=False) + '\n\n' if css: import css_parser import logging parser = css_parser.CSSParser(fetcher=None, log=logging.getLogger('calibre.css')) XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % const.XHTML_NS text = XHTML_CSS_NAMESPACE + css log.debug('Parsing stylesheet...') stylesheet = parser.parseString(text) stylesheet.namespaces['h'] = const.XHTML_NS css = stylesheet.cssText if isinstance(css, bytes): css = css.decode('utf-8', 'replace') css = css.replace('h|style', 'h|span') css = re.sub(r'name\s*=\s*', 'class=', css) self.extract_embedded_content(doc) log.debug('Converting XML to HTML...') with open( pkg_resources.resource_filename('ebook_converter', 'data/fb2.xsl')) as f: ss = f.read() ss = ss.replace("__FB_NS__", fb_ns) if options.no_inline_fb2_toc: log.info('Disabling generation of inline FB2 TOC') ss = re.compile(r'<!-- BUILD TOC -->.*<!-- END BUILD TOC -->', re.DOTALL).sub('', ss) styledoc = etree.fromstring(ss) transform = etree.XSLT(styledoc) result = transform(doc) # Handle links of type note and cite notes = { a.get('href')[1:]: a for a in result.xpath('//a[@link_note and @href]') if a.get('href').startswith('#') } cites = { a.get('link_cite'): a for a in result.xpath('//a[@link_cite]') if not a.get('href', '') } all_ids = {x for x in result.xpath('//*/@id')} for cite, a in cites.items(): note = notes.get(cite, None) if note: c = 1 while 'cite%d' % c in all_ids: c += 1 if not note.get('id', None): note.set('id', 'cite%d' % c) all_ids.add(note.get('id')) a.set('href', '#%s' % note.get('id')) for x in result.xpath('//*[@link_note or @link_cite]'): x.attrib.pop('link_note', None) x.attrib.pop('link_cite', None) for img in result.xpath('//img[@src]'): src = img.get('src') img.set('src', self.binary_map.get(src, src)) index = transform.tostring(result) with open('index.xhtml', 'wb') as f: f.write(index.encode('utf-8')) with open('inline-styles.css', 'wb') as f: f.write(css.encode('utf-8')) stream.seek(0) mi = get_metadata(stream, 'fb2') if not mi.title: mi.title = 'Unknown' if not mi.authors: mi.authors = ['Unknown'] cpath = None if mi.cover_data and mi.cover_data[1]: with open('fb2_cover_calibre_mi.jpg', 'wb') as f: f.write(mi.cover_data[1]) cpath = os.path.abspath('fb2_cover_calibre_mi.jpg') else: for img in doc.xpath('//f:coverpage/f:image', namespaces=NAMESPACES): href = img.get('{%s}href' % const.XLINK_NS, img.get('href', None)) if href is not None: if href.startswith('#'): href = href[1:] cpath = os.path.abspath(href) break opf = OPFCreator(os.getcwd(), mi) entries = [(f2, mimetypes.guess_type(f2)[0]) for f2 in os.listdir(u'.')] opf.create_manifest(entries) opf.create_spine(['index.xhtml']) if cpath: opf.guide.set_cover(cpath) with open('metadata.opf', 'wb') as f: opf.render(f) return os.path.join(os.getcwd(), 'metadata.opf')