def __call__(self, oeb, opts):
from ebook_converter.utils.imghdr import what
self.log = oeb.log
attr_path = XPath('//h:img[@src]')
for item in oeb.spine:
root = item.data
if not hasattr(root, 'xpath'):
continue
for img in attr_path(root):
raw = img.get('src', '')
if not raw.startswith('data:'):
continue
header, data = raw.partition(',')[0::2]
if not header.startswith('data:image/') or not data:
continue
if ';base64' in header:
data = re.sub(r'\s+', '', data)
try:
data = from_base64_bytes(data)
except Exception:
self.log.error('Found invalid base64 encoded data '
'URI, ignoring it')
continue
else:
data = urllib.parse.unquote(data)
data = as_bytes(data)
fmt = what(None, data)
if not fmt:
self.log.warn('Image encoded as data URL has unknown '
'format, ignoring')
continue
img.set(
'src',
item.relhref(self.convert_image_data_uri(data, fmt, oeb)))
def _create_html_root(self, hhcpath, log, encoding):
hhcdata = self._read_file(hhcpath)
hhcdata = hhcdata.decode(encoding)
hhcdata = xml_to_unicode(hhcdata, verbose=True,
strip_encoding_pats=True,
resolve_entities=True)[0]
hhcroot = html.fromstring(hhcdata)
toc = self._process_nodes(hhcroot)
log.debug('Found %d section nodes' % toc.count())
htmlpath = os.path.splitext(hhcpath)[0] + ".html"
base = os.path.dirname(os.path.abspath(htmlpath))
def unquote(x):
if isinstance(x, str):
x = x.encode('utf-8')
return _unquote(x).decode('utf-8')
def unquote_path(x):
y = unquote(x)
if (not os.path.exists(os.path.join(base, x)) and
os.path.exists(os.path.join(base, y))):
x = y
return x
def donode(item, parent, base, subpath):
for child in item:
title = child.title
if not title:
continue
raw = unquote_path(child.href or '')
rsrcname = os.path.basename(raw)
rsrcpath = os.path.join(subpath, rsrcname)
if (not os.path.exists(os.path.join(base, rsrcpath)) and
os.path.exists(os.path.join(base, raw))):
rsrcpath = raw
if '%' not in rsrcpath:
rsrcpath = urlquote(rsrcpath)
if not raw:
rsrcpath = ''
c = builder.DIV(builder.A(title, href=rsrcpath))
donode(child, c, base, subpath)
parent.append(c)
with open(htmlpath, 'wb') as f:
if toc.count() > 1:
path0 = toc[0].href
path0 = unquote_path(path0)
subpath = os.path.dirname(path0)
base = os.path.dirname(f.name)
root = builder.DIV()
donode(toc, root, base, subpath)
raw = html.tostring(builder.HTML(builder.BODY(root)),
encoding='utf-8',
pretty_print=True)
f.write(raw)
else:
f.write(as_bytes(hhcdata))
return htmlpath, toc
def write(self, name='styles.css'):
def join(style):
ans = ['%s : %s;' % (k, v) for k, v in style.items()]
if ans:
ans[-1] = ans[-1][:-1]
return '\n\t'.join(ans)
with open(name, 'wb') as f:
f.write(as_bytes(self.CSS))
for (w, sel) in [(self.text_styles, 'ts'),
(self.block_styles, 'bs')]:
for i, s in enumerate(w):
if not s:
continue
rsel = '.%s%d' % (sel, i)
s = join(s)
f.write(as_bytes(rsel + ' {\n\t' + s + '\n}\n\n'))
def get_table(raw, name):
''' Get the raw table bytes for the specified table in the font '''
name = as_bytes(name.lower())
for table_tag, table, table_index, table_offset, table_checksum in get_tables(
raw):
if table_tag.lower() == name:
return table, table_index, table_offset, table_checksum
return None, None, None, None
def __call__(self, stream, odir, log):
from ebook_converter.utils.zipfile import ZipFile
from ebook_converter.ebooks.metadata.odt import get_metadata
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
if not os.path.exists(odir):
os.makedirs(odir)
with CurrentDir(odir):
log('Extracting ODT file...')
stream.seek(0)
mi = get_metadata(stream, 'odt')
if not mi.title:
mi.title = 'Unknown'
if not mi.authors:
mi.authors = ['Unknown']
self.filter_load(stream, mi, log)
# NOTE(gryf): Here is a workaround for ODF2XHTML.xhtml() method,
# which expects, that all lines are strings.
html = ''.join([str(l) for l in self.lines])
# A blanket img specification like this causes problems
# with EPUB output as the containing element often has
# an absolute height and width set that is larger than
# the available screen real estate
html = html.replace('img { width: 100%; height: 100%; }', '')
# odf2xhtml creates empty title tag
html = html.replace('<title></title>',
'<title>%s</title>' % (mi.title, ))
try:
html = self.fix_markup(html, log)
except:
log.exception('Failed to filter CSS, conversion may be slow')
with open('index.xhtml', 'wb') as f:
f.write(as_bytes(html))
zf = ZipFile(stream, 'r')
self.extract_pictures(zf)
opf = OPFCreator(os.path.abspath(os.getcwd()), mi)
opf.create_manifest([(os.path.abspath(f2), None)
for f2 in walk(os.getcwd())])
opf.create_spine([os.path.abspath('index.xhtml')])
with open('metadata.opf', 'wb') as f:
opf.render(f)
return os.path.abspath('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, oeb, output_path, input_plugin, opts, log):
self.log, self.opts, self.oeb = log, opts, oeb
if self.opts.epub_inline_toc:
from ebook_converter.ebooks.mobi.writer8.toc import TOCAdder
opts.mobi_toc_at_start = not opts.epub_toc_at_end
opts.mobi_passthrough = False
opts.no_inline_toc = False
TOCAdder(oeb, opts, replace_previous_inline_toc=True, ignore_existing_toc=True)
if self.opts.epub_flatten:
from ebook_converter.ebooks.oeb.transforms.filenames import FlatFilenames
FlatFilenames()(oeb, opts)
else:
from ebook_converter.ebooks.oeb.transforms.filenames import UniqueFilenames
UniqueFilenames()(oeb, opts)
self.workaround_ade_quirks()
self.workaround_webkit_quirks()
self.upshift_markup()
from ebook_converter.ebooks.oeb.transforms.rescale import RescaleImages
RescaleImages(check_colorspaces=True)(oeb, opts)
from ebook_converter.ebooks.oeb.transforms.split import Split
split = Split(not self.opts.dont_split_on_page_breaks,
max_flow_size=self.opts.flow_size*1024
)
split(self.oeb, self.opts)
from ebook_converter.ebooks.oeb.transforms.cover import CoverManager
cm = CoverManager(
no_default_cover=self.opts.no_default_epub_cover,
no_svg_cover=self.opts.no_svg_cover,
preserve_aspect_ratio=self.opts.preserve_cover_aspect_ratio)
cm(self.oeb, self.opts, self.log)
self.workaround_sony_quirks()
if self.oeb.toc.count() == 0:
self.log.warn('This EPUB file has no Table of Contents. '
'Creating a default TOC')
first = next(iter(self.oeb.spine))
self.oeb.toc.add('Start', first.href)
identifiers = oeb.metadata['identifier']
_uuid = None
for x in identifiers:
if (x.get(base.tag('opf', 'scheme'), None).lower() == 'uuid' or
str(x).startswith('urn:uuid:')):
_uuid = str(x).split(':')[-1]
break
encrypted_fonts = getattr(input_plugin, 'encrypted_fonts', [])
if _uuid is None:
self.log.warn('No UUID identifier found')
_uuid = str(uuid.uuid4())
oeb.metadata.add('identifier', _uuid, scheme='uuid', id=_uuid)
if encrypted_fonts and not _uuid.startswith('urn:uuid:'):
# Apparently ADE requires this value to start with urn:uuid:
# for some absurd reason, or it will throw a hissy fit and refuse
# to use the obfuscated fonts.
for x in identifiers:
if str(x) == _uuid:
x.content = 'urn:uuid:' + _uuid
with TemporaryDirectory('_epub_output') as tdir:
from ebook_converter.customize.ui import plugin_for_output_format
metadata_xml = None
extra_entries = []
if self.is_periodical:
if self.opts.output_profile.epub_periodical_format == 'sony':
from ebook_converter.ebooks.epub.periodical import sony_metadata
metadata_xml, atom_xml = sony_metadata(oeb)
extra_entries = [('atom.xml', 'application/atom+xml', atom_xml)]
oeb_output = plugin_for_output_format('oeb')
oeb_output.convert(oeb, tdir, input_plugin, opts, log)
opf = [x for x in os.listdir(tdir) if x.endswith('.opf')][0]
self.condense_ncx([os.path.join(tdir, x) for x in os.listdir(tdir)
if x.endswith('.ncx')][0])
if self.opts.epub_version == '3':
self.upgrade_to_epub3(tdir, opf)
encryption = None
if encrypted_fonts:
encryption = self.encrypt_fonts(encrypted_fonts, tdir, _uuid)
from ebook_converter.ebooks.epub import initialize_container
with initialize_container(output_path, os.path.basename(opf),
extra_entries=extra_entries) as epub:
epub.add_dir(tdir)
if encryption is not None:
epub.writestr('META-INF/encryption.xml', as_bytes(encryption))
if metadata_xml is not None:
epub.writestr('META-INF/metadata.xml',
metadata_xml.encode('utf-8'))
if opts.extract_to is not None:
from ebook_converter.utils.zipfile import ZipFile
if os.path.exists(opts.extract_to):
if os.path.isdir(opts.extract_to):
shutil.rmtree(opts.extract_to)
else:
os.remove(opts.extract_to)
os.mkdir(opts.extract_to)
with ZipFile(output_path) as zf:
zf.extractall(path=opts.extract_to)
self.log.info('EPUB extracted to', opts.extract_to)
def get_css(self, oeb_book):
css = b''
for item in oeb_book.manifest:
if item.media_type == 'text/css':
css += as_bytes(item.data.cssText) + b'\n\n'
return css
def encode_string(raw):
ans = bytearray(as_bytes(raw))
ans.insert(0, len(ans))
return bytes(ans)