def test_parentRule(self):
"CSSStyleDeclaration.parentRule"
s = css_parser.css.CSSStyleDeclaration()
sheet = css_parser.css.CSSStyleRule()
s.parentRule = sheet
self.assertEqual(sheet, s.parentRule)
sheet = css_parser.parseString('a{x:1}')
s = sheet.cssRules[0]
d = s.style
self.assertEqual(s, d.parentRule)
s = css_parser.parseString('''
@font-face {
font-weight: bold;
}
a {
font-weight: bolder;
}
@page {
font-weight: bolder;
}
''')
for r in s:
self.assertEqual(r.style.parentRule, r)
def test_handlers(self):
"css_parser.log"
s = self._setHandler()
css_parser.log.setLevel(logging.FATAL)
self.assertEqual(css_parser.log.getEffectiveLevel(), logging.FATAL)
css_parser.parseString('a { color: 1 }')
self.assertEqual(s.getvalue(), '')
css_parser.log.setLevel(logging.DEBUG)
css_parser.parseString('a { color: 1 }')
self.assertEqual(
s.getvalue(),
'ERROR Property: Invalid value for "CSS Level 2.1" property: 1 [1:5: color]\n'
)
s = self._setHandler()
css_parser.log.setLevel(logging.WARNING)
css_parser.parseUrl('http://example.com')
q = s.getvalue()[:38]
if q.startswith('WARNING URLError'):
pass
else:
self.assertEqual(q, 'ERROR Expected "text/css" mime type')
def test_comments(self):
"PropertyValue with comment"
# issue #45
for t in (
'green',
'green /* comment */',
'/* comment */green',
'/* comment */green/* comment */',
'/* comment */ green /* comment */',
'/* comment *//**/ green /* comment *//**/',
):
sheet = css_parser.parseString('body {color: %s; }' % t)
p = sheet.cssRules[0].style.getProperties()[0]
self.assertEqual(p.valid, True)
for t in (
'gree',
'gree /* comment */',
'/* comment */gree',
'/* comment */gree/* comment */',
'/* comment */ gree /* comment */',
'/* comment *//**/ gree /* comment *//**/',
):
sheet = css_parser.parseString('body {color: %s; }' % t)
p = sheet.cssRules[0].style.getProperties()[0]
self.assertEqual(p.valid, False)
def test_cssRules(self):
"CSSMediaRule.cssRules"
r = css_parser.css.CSSMediaRule()
self.assertEqual([], r.cssRules)
sr = css_parser.css.CSSStyleRule()
r.cssRules.append(sr)
self.assertEqual([sr], r.cssRules)
ir = css_parser.css.CSSImportRule()
self.assertRaises(xml.dom.HierarchyRequestErr, r.cssRules.append, ir)
s = css_parser.parseString('@media all { /*1*/a {x:1} }')
m = s.cssRules[0]
self.assertEqual(2, m.cssRules.length)
del m.cssRules[0]
self.assertEqual(1, m.cssRules.length)
m.cssRules.append('/*2*/')
self.assertEqual(2, m.cssRules.length)
m.cssRules.extend(css_parser.parseString('/*3*/x {y:2}').cssRules)
self.assertEqual(4, m.cssRules.length)
self.assertEqual(
'@media all {\n a {\n x: 1\n }\n /*2*/\n /*3*/\n x {\n y: 2\n }\n }',
m.cssText)
for rule in m.cssRules:
self.assertEqual(rule.parentStyleSheet, s)
self.assertEqual(rule.parentRule, m)
def test_parsevalidation(self):
style = 'color: 1'
t = 'a { %s }' % style
css_parser.log.setLevel(logging.DEBUG)
# sheet
s = self._setHandler()
css_parser.parseString(t)
self.assertNotEqual(len(s.getvalue()), 0)
s = self._setHandler()
css_parser.parseString(t, validate=False)
self.assertEqual(s.getvalue(), '')
# style
s = self._setHandler()
css_parser.parseStyle(style)
self.assertNotEqual(len(s.getvalue()), 0)
s = self._setHandler()
css_parser.parseStyle(style, validate=True)
self.assertNotEqual(len(s.getvalue()), 0)
s = self._setHandler()
css_parser.parseStyle(style, validate=False)
self.assertEqual(s.getvalue(), '')
def test_set(self):
"settings.set()"
css_parser.ser.prefs.useMinified()
text = 'a {filter: progid:DXImageTransform.Microsoft.BasicImage( rotation = 90 )}'
self.assertEqual(css_parser.parseString(text).cssText, ''.encode())
css_parser.settings.set('DXImageTransform.Microsoft', True)
self.assertEqual(css_parser.parseString(text).cssText,
'a{filter:progid:DXImageTransform.Microsoft.BasicImage(rotation=90)}'.encode())
css_parser.ser.prefs.useDefaults()
def test_attributes(self):
"css_parser.parseString(href, media)"
s = css_parser.parseString("a{}",
href="file:foo.css",
media="screen, projection, tv")
self.assertEqual(s.href, "file:foo.css")
self.assertEqual(s.media.mediaText, "screen, projection, tv")
s = css_parser.parseString("a{}",
href="file:foo.css",
media=["screen", "projection", "tv"])
self.assertEqual(s.media.mediaText, "screen, projection, tv")
def test_escapes(self):
"css_parser escapes"
css = r'\43\x { \43\x: \43\x !import\41nt }'
sheet = css_parser.parseString(css)
self.assertEqual(sheet.cssText, r'''C\x {
c\x: C\x !important
}'''.encode())
css = r'\ x{\ x :\ x ;y:1} '
sheet = css_parser.parseString(css)
self.assertEqual(sheet.cssText, r'''\ x {
\ x: \ x;
y: 1
}'''.encode())
def test_omitLastSemicolon(self):
"Preferences.omitLastSemicolon"
css = 'a { x: 1; y: 2 }'
s = css_parser.parseString(css)
self.assertEqual('a {\n x: 1;\n y: 2\n }'.encode(), s.cssText)
css_parser.ser.prefs.omitLastSemicolon = False
self.assertEqual('a {\n x: 1;\n y: 2;\n }'.encode(), s.cssText)
def test_defaultPropertyPriority(self):
"Preferences.defaultPropertyPriority"
css = 'a {\n color: green !IM\\portant\n }'
s = css_parser.parseString(css)
self.assertEqual(s.cssText, 'a {\n color: green !important\n }'.encode())
css_parser.ser.prefs.defaultPropertyPriority = False
self.assertEqual(s.cssText, css.encode())
def test_styleSheet(self):
"CSSImportRule.styleSheet"
def fetcher(url):
if url == "/root/level1/anything.css":
return None, '@import "level2/css.css" "title2";'
else:
return None, 'a { color: red }'
parser = css_parser.CSSParser(fetcher=fetcher)
sheet = parser.parseString('''@charset "ascii";
@import "level1/anything.css" tv "title";''',
href='/root/')
self.assertEqual(sheet.href, '/root/')
ir = sheet.cssRules[1]
self.assertEqual(ir.href, 'level1/anything.css')
self.assertEqual(ir.styleSheet.href, '/root/level1/anything.css')
# inherits ascii as no self charset is set
self.assertEqual(ir.styleSheet.encoding, 'ascii')
self.assertEqual(ir.styleSheet.ownerRule, ir)
self.assertEqual(ir.styleSheet.media.mediaText, 'tv')
self.assertEqual(ir.styleSheet.parentStyleSheet, None) # sheet
self.assertEqual(ir.styleSheet.title, 'title')
self.assertEqual(ir.styleSheet.cssText,
'@charset "ascii";\n@import "level2/css.css" "title2";'.encode())
ir2 = ir.styleSheet.cssRules[1]
self.assertEqual(ir2.href, 'level2/css.css')
self.assertEqual(ir2.styleSheet.href, '/root/level1/level2/css.css')
# inherits ascii as no self charset is set
self.assertEqual(ir2.styleSheet.encoding, 'ascii')
self.assertEqual(ir2.styleSheet.ownerRule, ir2)
self.assertEqual(ir2.styleSheet.media.mediaText, 'all')
self.assertEqual(ir2.styleSheet.parentStyleSheet, None) # ir.styleSheet
self.assertEqual(ir2.styleSheet.title, 'title2')
self.assertEqual(ir2.styleSheet.cssText,
'@charset "ascii";\na {\n color: red\n }'.encode())
sheet = css_parser.parseString('@import "CANNOT-FIND.css";')
ir = sheet.cssRules[0]
self.assertEqual(ir.href, "CANNOT-FIND.css")
self.assertEqual(type(ir.styleSheet), css_parser.css.CSSStyleSheet)
def fetcher(url):
if url.endswith('level1.css'):
return None, '@charset "ascii"; @import "level2.css";'.encode()
else:
return None, 'a { color: red }'.encode()
parser = css_parser.CSSParser(fetcher=fetcher)
sheet = parser.parseString('@charset "iso-8859-1";@import "level1.css";')
self.assertEqual(sheet.encoding, 'iso-8859-1')
sheet = sheet.cssRules[1].styleSheet
self.assertEqual(sheet.encoding, 'ascii')
sheet = sheet.cssRules[1].styleSheet
self.assertEqual(sheet.encoding, 'ascii')
def test_replaceUrls(self):
"css_parser.replaceUrls()"
css_parser.ser.prefs.keepAllProperties = True
css = r'''
@import "im1";
@import url(im2);
a {
background-image: url(c) !important;
background-\image: url(b);
background: url(a) no-repeat !important;
}'''
s = css_parser.parseString(css)
css_parser.replaceUrls(s, lambda old: "NEW" + old)
self.assertEqual('@import "NEWim1";', s.cssRules[0].cssText)
self.assertEqual('NEWim2', s.cssRules[1].href)
self.assertEqual(
'''background-image: url(NEWc) !important;
background-\\image: url(NEWb);
background: url(NEWa) no-repeat !important''', s.cssRules[2].style.cssText)
css_parser.ser.prefs.keepAllProperties = False
# CSSStyleDeclaration
style = css_parser.parseStyle('''color: red;
background-image:
url(1.png),
url('2.png')''')
css_parser.replaceUrls(style, lambda url: 'prefix/' + url)
self.assertEqual(
style.cssText, '''color: red;
background-image: url(prefix/1.png), url(prefix/2.png)''')
def test_list(self):
"PropertyValue[index]"
# issue #41
css = """div.one {color: rgb(255, 0, 0);} """
sheet = css_parser.parseString(css)
pv = sheet.cssRules[0].style.getProperty('color').propertyValue
self.assertEqual(pv.value, 'rgb(255, 0, 0)')
self.assertEqual(pv[0].value, 'rgb(255, 0, 0)')
# issue #42
sheet = css_parser.parseString('body { font-family: "A", b, serif }')
pv = sheet.cssRules[0].style.getProperty('font-family').propertyValue
self.assertEqual(3, pv.length)
self.assertEqual(pv[0].value, 'A')
self.assertEqual(pv[1].value, 'b')
self.assertEqual(pv[2].value, 'serif')
def test_getUrls(self):
"css_parser.getUrls()"
css_parser.ser.prefs.keepAllProperties = True
css = r'''
@import "im1";
@import url(im2);
@import url( im3 );
@import url( "im4" );
@import url( 'im5' );
a {
background-image: url(a) !important;
background-\image: url(b);
background: url(c) no-repeat !important;
/* issue #46 */
src: local("xx"),
url("f.woff") format("woff"),
url("f.otf") format("opentype"),
url("f.svg#f") format("svg");
}'''
urls = set(css_parser.getUrls(css_parser.parseString(css)))
self.assertEqual(
urls,
set([
"im1", "im2", "im3", "im4", "im5", "a", "b", "c", 'f.woff',
'f.svg#f', 'f.otf'
]))
css_parser.ser.prefs.keepAllProperties = False
def test_keepComments(self):
"Preferences.keepComments"
s = css_parser.parseString('/*1*/ a { /*2*/ }')
css_parser.ser.prefs.keepComments = False
self.assertEqual(''.encode(), s.cssText)
css_parser.ser.prefs.keepEmptyRules = True
self.assertEqual('a {}'.encode(), s.cssText)
def html_css_stylesheet():
global _html_css_stylesheet
if _html_css_stylesheet is None:
with open(P('templates/html.css'), 'rb') as f:
html_css = f.read().decode('utf-8')
_html_css_stylesheet = parseString(html_css, validate=False)
return _html_css_stylesheet
def test_keepUsedNamespaceRulesOnly(self):
"Preferences.keepUsedNamespaceRulesOnly"
tests = {
# default == prefix => both are combined
'@namespace p "u"; @namespace "u"; p|a, a {top: 0}':
('@namespace "u";\na, a {\n top: 0\n }',
'@namespace "u";\na, a {\n top: 0\n }'),
'@namespace "u"; @namespace p "u"; p|a, a {top: 0}':
('@namespace p "u";\np|a, p|a {\n top: 0\n }',
'@namespace p "u";\np|a, p|a {\n top: 0\n }'),
# default and prefix
'@namespace p "u"; @namespace "d"; p|a, a {top: 0}':
('@namespace p "u";\n@namespace "d";\np|a, a {\n top: 0\n }',
'@namespace p "u";\n@namespace "d";\np|a, a {\n top: 0\n }'
),
# prefix only
'@namespace p "u"; @namespace "d"; p|a {top: 0}':
('@namespace p "u";\n@namespace "d";\np|a {\n top: 0\n }',
'@namespace p "u";\np|a {\n top: 0\n }'),
# default only
'@namespace p "u"; @namespace "d"; a {top: 0}':
('@namespace p "u";\n@namespace "d";\na {\n top: 0\n }',
'@namespace "d";\na {\n top: 0\n }'),
# prefix-ns only
'@namespace p "u"; @namespace d "d"; p|a {top: 0}':
('@namespace p "u";\n@namespace d "d";\np|a {\n top: 0\n }',
'@namespace p "u";\np|a {\n top: 0\n }'),
}
for test in tests:
s = css_parser.parseString(test)
expwith, expwithout = tests[test]
css_parser.ser.prefs.keepUsedNamespaceRulesOnly = False
self.assertEqual(s.cssText, expwith.encode())
css_parser.ser.prefs.keepUsedNamespaceRulesOnly = True
self.assertEqual(s.cssText, expwithout.encode())
def test_minimizeColorHash(self):
"Preferences.minimizeColorHash"
css = 'a { color: #ffffff }'
s = css_parser.parseString(css)
self.assertEqual('a {\n color: #fff\n }'.encode(), s.cssText)
css_parser.ser.prefs.minimizeColorHash = False
self.assertEqual('a {\n color: #ffffff\n }'.encode(), s.cssText)
def test_propertyNameSpacer(self):
"Preferences.propertyNameSpacer"
css = 'a { x: 1; y: 2 }'
s = css_parser.parseString(css)
self.assertEqual('a {\n x: 1;\n y: 2\n }'.encode(), s.cssText)
css_parser.ser.prefs.propertyNameSpacer = ''
self.assertEqual('a {\n x:1;\n y:2\n }'.encode(), s.cssText)
def test_linesAfterRule(self):
"Preferences.linesAfterRule"
s = css_parser.parseString(
'div {color:red;} @media screen {.aclass {width: 200px}}')
expected_default = '''\
div {
color: red
}
@media screen {
.aclass {
width: 200px
}
}'''
self.assertEqual(expected_default.encode(), s.cssText)
css_parser.ser.prefs.linesAfterRules = 1 * '\n'
expected_changed = '''\
div {
color: red
}
@media screen {
.aclass {
width: 200px
}
}
'''
self.assertEqual(expected_changed.encode(), s.cssText)
def test_keepUnknownAtRules(self):
"Preferences.keepUnknownAtRules"
tests = {
'''@three-dee {
@background-lighting {
azimuth: 30deg;
elevation: 190deg;
}
h1 { color: red }
}
h1 { color: blue }''': ('''@three-dee {
@background-lighting {
azimuth: 30deg;
elevation: 190deg;
} h1 {
color: red
}
}
h1 {
color: blue
}''', '''h1 {
color: blue
}''')
}
for test in tests:
s = css_parser.parseString(test)
expwith, expwithout = tests[test]
css_parser.ser.prefs.keepUnknownAtRules = True
self.assertEqual(s.cssText, expwith.encode())
css_parser.ser.prefs.keepUnknownAtRules = False
self.assertEqual(s.cssText, expwithout.encode())
def test_incomplete(self):
"PropertyValue (incomplete)"
tests = {'url("a': 'url(a)', 'url(a': 'url(a)'}
for v, exp in tests.items():
s = css_parser.parseString('a { background: %s' % v)
v = s.cssRules[0].style.background
self.assertEqual(v, exp)
def do_filter_css(self, css):
from css_parser import parseString
from css_parser.css import CSSRule
sheet = parseString(css, validate=False)
rules = list(sheet.cssRules.rulesOfType(CSSRule.STYLE_RULE))
sel_map = {}
count = 0
for r in rules:
# Check if we have only class selectors for this rule
nc = [
x for x in r.selectorList if not x.selectorText.startswith('.')
]
if len(r.selectorList) > 1 and not nc:
# Replace all the class selectors with a single class selector
# This will be added to the class attribute of all elements
# that have one of these selectors.
replace_name = 'c_odt%d' % count
count += 1
for sel in r.selectorList:
s = sel.selectorText[1:]
if s not in sel_map:
sel_map[s] = []
sel_map[s].append(replace_name)
r.selectorText = '.' + replace_name
return sheet.cssText, sel_map
def _get_css_links(
content: bytes,
base_url: str,
headers: _Headers,
) -> Iterable[str]:
"""Get all links from a CSS file."""
parsed = css_parser.parseString(content)
return list(css_parser.getUrls(parsed))
def html_css_stylesheet():
global _html_css_stylesheet
if _html_css_stylesheet is None:
with open(pkg_resources.resource_filename('ebook_converter',
'data/html.css'), 'rb') as f:
html_css = f.read().decode('utf-8')
_html_css_stylesheet = parseString(html_css, validate=False)
return _html_css_stylesheet
def test_defaultAtKeyword(self):
"Preferences.defaultAtKeyword"
s = css_parser.parseString('@im\\port "x";')
self.assertEqual('@import "x";'.encode(), s.cssText)
css_parser.ser.prefs.defaultAtKeyword = True
self.assertEqual('@import "x";'.encode(), s.cssText)
css_parser.ser.prefs.defaultAtKeyword = False
self.assertEqual('@im\\port "x";'.encode(), s.cssText)
def get_css_links(
css_file: BinaryIO,
base_url: str,
headers: Optional[List[Tuple[str, str]]] = None,
) -> Iterable[str]:
"""Get all links from a CSS file."""
text = css_file.read()
parsed = css_parser.parseString(text)
yield from css_parser.getUrls(parsed)
def test_CSSStyleSheet(self):
"CSSSerializer.do_CSSStyleSheet"
css = '/* κουρος */'
sheet = css_parser.parseString(css)
ans = sheet.cssText
if isinstance(ans, bytes):
ans = ans.decode('utf-8')
self.assertEqual(css, ans)
css = '@charset "utf-8";\n/* κουρος */'
sheet = css_parser.parseString(css)
ans = sheet.cssText
if isinstance(ans, bytes):
ans = ans.decode('utf-8')
self.assertEqual(css, ans)
sheet.cssRules[0].encoding = 'ascii'
self.assertEqual('@charset "ascii";\n/* \\3BA \\3BF \\3C5 \\3C1 \\3BF \\3C2 */'.encode(),
sheet.cssText)
def test_lineSeparator(self):
"Preferences.lineSeparator"
s = css_parser.parseString('a { x:1;y:2}')
self.assertEqual('a {\n x: 1;\n y: 2\n }'.encode(), s.cssText)
# cannot be indented as no split possible
css_parser.ser.prefs.lineSeparator = ''
self.assertEqual('a {x: 1;y: 2 }'.encode(), s.cssText)
# no valid css but should work
css_parser.ser.prefs.lineSeparator = 'XXX'
self.assertEqual('a {XXX x: 1;XXX y: 2XXX }'.encode(), s.cssText)
def test_parentRule(self):
"CSSVariablesDeclaration.parentRule"
s = css_parser.parseString('@variables { a:1}')
r = s.cssRules[0]
d = r.variables
self.assertEqual(r, d.parentRule)
d2 = css_parser.css.CSSVariablesDeclaration('b: 2')
r.variables = d2
self.assertEqual(r, d2.parentRule)
def __call__(self, oeb):
if not self.body_font_family:
return None
if not self.href:
iid, href = oeb.manifest.generate(u'page_styles', u'page_styles.css')
rules = [css_text(x) for x in self.rules]
rules = u'\n\n'.join(rules)
sheet = css_parser.parseString(rules, validate=False)
self.href = oeb.manifest.add(iid, href, guess_type(href)[0],
data=sheet).href
return self.href
def replace_css(self, css):
manifest = self.oeb.manifest
for item in manifest.values():
if item.media_type in OEB_STYLES:
manifest.remove(item)
id, href = manifest.generate('css', 'stylesheet.css')
sheet = css_parser.parseString(css, validate=False)
if self.transform_css_rules:
from calibre.ebooks.css_transform_rules import transform_sheet
transform_sheet(self.transform_css_rules, sheet)
item = manifest.add(id, href, CSS_MIME, data=sheet)
self.oeb.manifest.main_stylesheet = item
return href
def get_embed_font_info(self, family, failure_critical=True):
efi = []
body_font_family = None
if not family:
return body_font_family, efi
from calibre.utils.fonts.scanner import font_scanner, NoFonts
from calibre.utils.fonts.utils import panose_to_css_generic_family
try:
faces = font_scanner.fonts_for_family(family)
except NoFonts:
msg = (u'No embeddable fonts found for family: %r'%family)
if failure_critical:
raise ValueError(msg)
self.oeb.log.warn(msg)
return body_font_family, efi
if not faces:
msg = (u'No embeddable fonts found for family: %r'%family)
if failure_critical:
raise ValueError(msg)
self.oeb.log.warn(msg)
return body_font_family, efi
for i, font in enumerate(faces):
ext = 'otf' if font['is_otf'] else 'ttf'
fid, href = self.oeb.manifest.generate(id=u'font',
href=u'fonts/%s.%s'%(ascii_filename(font['full_name']).replace(u' ', u'-'), ext))
item = self.oeb.manifest.add(fid, href,
guess_type('dummy.'+ext)[0],
data=font_scanner.get_font_data(font))
item.unload_data_from_memory()
cfont = {
u'font-family':u'"%s"'%font['font-family'],
u'panose-1': u' '.join(map(unicode_type, font['panose'])),
u'src': u'url(%s)'%item.href,
}
if i == 0:
generic_family = panose_to_css_generic_family(font['panose'])
body_font_family = u"'%s',%s"%(font['font-family'], generic_family)
self.oeb.log(u'Embedding font: %s'%font['font-family'])
for k in (u'font-weight', u'font-style', u'font-stretch'):
if font[k] != u'normal':
cfont[k] = font[k]
rule = '@font-face { %s }'%('; '.join(u'%s:%s'%(k, v) for k, v in
iteritems(cfont)))
rule = css_parser.parseString(rule)
efi.append(rule)
return body_font_family, efi
def dup_data(self):
''' Duplicate data so that any changes we make to markup/CSS only
affect KF8 output and not MOBI 6 output '''
self._data_cache = {}
# Suppress css_parser logging output as it is duplicated anyway earlier
# in the pipeline
css_parser.log.setLevel(logging.CRITICAL)
for item in self.oeb.manifest:
if item.media_type in XML_DOCS:
self._data_cache[item.href] = copy.deepcopy(item.data)
elif item.media_type in OEB_STYLES:
# I can't figure out how to make an efficient copy of the
# in-memory CSSStylesheet, as deepcopy doesn't work (raises an
# exception)
self._data_cache[item.href] = css_parser.parseString(
item.data.cssText, validate=False)
def replace_resource_links(self):
''' Replace links to resources (raster images/fonts) with pointers to
the MOBI record containing the resource. The pointers are of the form:
kindle:embed:XXXX?mime=image/* The ?mime= is apparently optional and
not used for fonts. '''
def pointer(item, oref):
ref = urlnormalize(item.abshref(oref))
idx = self.resources.item_map.get(ref, None)
if idx is not None:
is_image = self.resources.records[idx-1][:4] not in {b'FONT'}
idx = to_ref(idx)
if is_image:
self.used_images.add(ref)
return 'kindle:embed:%s?mime=%s'%(idx,
self.resources.mime_map[ref])
else:
return 'kindle:embed:%s'%idx
return oref
for item in self.oeb.manifest:
if item.media_type in XML_DOCS:
root = self.data(item)
for tag in XPath('//h:img|//svg:image')(root):
for attr, ref in iteritems(tag.attrib):
if attr.split('}')[-1].lower() in {'src', 'href'}:
tag.attrib[attr] = pointer(item, ref)
for tag in XPath('//h:style')(root):
if tag.text:
sheet = css_parser.parseString(tag.text, validate=False)
replacer = partial(pointer, item)
css_parser.replaceUrls(sheet, replacer,
ignoreImportRules=True)
repl = sheet.cssText
if isbytestring(repl):
repl = repl.decode('utf-8')
tag.text = '\n'+ repl + '\n'
elif item.media_type in OEB_STYLES:
sheet = self.data(item)
replacer = partial(pointer, item)
css_parser.replaceUrls(sheet, replacer, ignoreImportRules=True)
def collect_global_css(self):
global_css = defaultdict(list)
for item in self.items:
stylizer = self.stylizers[item]
if float(self.context.margin_top) >= 0:
stylizer.page_rule['margin-top'] = '%gpt'%\
float(self.context.margin_top)
if float(self.context.margin_bottom) >= 0:
stylizer.page_rule['margin-bottom'] = '%gpt'%\
float(self.context.margin_bottom)
items = sorted(stylizer.page_rule.items())
css = ';\n'.join("%s: %s" % (key, val) for key, val in items)
css = ('@page {\n%s\n}\n'%css) if items else ''
rules = [css_text(r) for r in stylizer.font_face_rules + self.embed_font_rules]
raw = '\n\n'.join(rules)
css += '\n\n' + raw
global_css[css].append(item)
gc_map = {}
manifest = self.oeb.manifest
for css in global_css:
href = None
if css.strip():
id_, href = manifest.generate('page_css', 'page_styles.css')
sheet = css_parser.parseString(css, validate=False)
if self.transform_css_rules:
from calibre.ebooks.css_transform_rules import transform_sheet
transform_sheet(self.transform_css_rules, sheet)
manifest.add(id_, href, CSS_MIME, data=sheet)
gc_map[css] = href
ans = {}
for css, items in iteritems(global_css):
for item in items:
ans[item] = gc_map[css]
return ans
def do_filter_css(self, css):
from css_parser import parseString
from css_parser.css import CSSRule
sheet = parseString(css, validate=False)
rules = list(sheet.cssRules.rulesOfType(CSSRule.STYLE_RULE))
sel_map = {}
count = 0
for r in rules:
# Check if we have only class selectors for this rule
nc = [x for x in r.selectorList if not
x.selectorText.startswith('.')]
if len(r.selectorList) > 1 and not nc:
# Replace all the class selectors with a single class selector
# This will be added to the class attribute of all elements
# that have one of these selectors.
replace_name = 'c_odt%d'%count
count += 1
for sel in r.selectorList:
s = sel.selectorText[1:]
if s not in sel_map:
sel_map[s] = []
sel_map[s].append(replace_name)
r.selectorText = '.'+replace_name
return sheet.cssText, sel_map
def __init__(self, tree, path, oeb, opts, profile=None,
extra_css='', user_css='', base_css=''):
self.oeb, self.opts = oeb, opts
self.profile = profile
if self.profile is None:
# Use the default profile. This should really be using
# opts.output_profile, but I don't want to risk changing it, as
# doing so might well have hard to debug font size effects.
from calibre.customize.ui import output_profiles
for x in output_profiles():
if x.short_name == 'default':
self.profile = x
break
if self.profile is None:
# Just in case the default profile is removed in the future :)
self.profile = opts.output_profile
self.body_font_size = self.profile.fbase
self.logger = oeb.logger
item = oeb.manifest.hrefs[path]
basename = os.path.basename(path)
cssname = os.path.splitext(basename)[0] + '.css'
stylesheets = [html_css_stylesheet()]
if base_css:
stylesheets.append(parseString(base_css, validate=False))
style_tags = xpath(tree, '//*[local-name()="style" or local-name()="link"]')
# Add css_parser parsing profiles from output_profile
for profile in self.opts.output_profile.extra_css_modules:
cssprofiles.addProfile(profile['name'],
profile['props'],
profile['macros'])
parser = CSSParser(fetcher=self._fetch_css_file,
log=logging.getLogger('calibre.css'))
self.font_face_rules = []
for elem in style_tags:
if (elem.tag == XHTML('style') and elem.get('type', CSS_MIME) in OEB_STYLES and media_ok(elem.get('media'))):
text = elem.text if elem.text else u''
for x in elem:
t = getattr(x, 'text', None)
if t:
text += u'\n\n' + force_unicode(t, u'utf-8')
t = getattr(x, 'tail', None)
if t:
text += u'\n\n' + force_unicode(t, u'utf-8')
if text:
text = oeb.css_preprocessor(text)
# We handle @import rules separately
parser.setFetcher(lambda x: ('utf-8', b''))
stylesheet = parser.parseString(text, href=cssname,
validate=False)
parser.setFetcher(self._fetch_css_file)
for rule in stylesheet.cssRules:
if rule.type == rule.IMPORT_RULE:
ihref = item.abshref(rule.href)
if not media_ok(rule.media.mediaText):
continue
hrefs = self.oeb.manifest.hrefs
if ihref not in hrefs:
self.logger.warn('Ignoring missing stylesheet in @import rule:', rule.href)
continue
sitem = hrefs[ihref]
if sitem.media_type not in OEB_STYLES:
self.logger.warn('CSS @import of non-CSS file %r' % rule.href)
continue
stylesheets.append(sitem.data)
# Make links to resources absolute, since these rules will
# be folded into a stylesheet at the root
replaceUrls(stylesheet, item.abshref,
ignoreImportRules=True)
stylesheets.append(stylesheet)
elif (elem.tag == XHTML('link') and elem.get('href') and elem.get(
'rel', 'stylesheet').lower() == 'stylesheet' and elem.get(
'type', CSS_MIME).lower() in OEB_STYLES and media_ok(elem.get('media'))
):
href = urlnormalize(elem.attrib['href'])
path = item.abshref(href)
sitem = oeb.manifest.hrefs.get(path, None)
if sitem is None:
self.logger.warn(
'Stylesheet %r referenced by file %r not in manifest' %
(path, item.href))
continue
if not hasattr(sitem.data, 'cssRules'):
self.logger.warn(
'Stylesheet %r referenced by file %r is not CSS'%(path,
item.href))
continue
stylesheets.append(sitem.data)
csses = {'extra_css':extra_css, 'user_css':user_css}
for w, x in csses.items():
if x:
try:
text = x
stylesheet = parser.parseString(text, href=cssname,
validate=False)
stylesheets.append(stylesheet)
except:
self.logger.exception('Failed to parse %s, ignoring.'%w)
self.logger.debug('Bad css: ')
self.logger.debug(x)
rules = []
index = 0
self.stylesheets = set()
self.page_rule = {}
for sheet_index, stylesheet in enumerate(stylesheets):
href = stylesheet.href
self.stylesheets.add(href)
for rule in stylesheet.cssRules:
if rule.type == rule.MEDIA_RULE:
if media_ok(rule.media.mediaText):
for subrule in rule.cssRules:
rules.extend(self.flatten_rule(subrule, href, index, is_user_agent_sheet=sheet_index==0))
index += 1
else:
rules.extend(self.flatten_rule(rule, href, index, is_user_agent_sheet=sheet_index==0))
index = index + 1
rules.sort(key=itemgetter(0)) # sort by specificity
self.rules = rules
self._styles = {}
pseudo_pat = re.compile(u':{1,2}(%s)' % ('|'.join(INAPPROPRIATE_PSEUDO_CLASSES)), re.I)
select = Select(tree, ignore_inappropriate_pseudo_classes=True)
for _, _, cssdict, text, _ in rules:
fl = pseudo_pat.search(text)
try:
matches = tuple(select(text))
except SelectorError as err:
self.logger.error('Ignoring CSS rule with invalid selector: %r (%s)' % (text, as_unicode(err)))
continue
if fl is not None:
fl = fl.group(1)
if fl == 'first-letter' and getattr(self.oeb,
'plumber_output_format', '').lower() in {u'mobi', u'docx'}:
# Fake first-letter
for elem in matches:
for x in elem.iter('*'):
if x.text:
punctuation_chars = []
text = unicode_type(x.text)
while text:
category = unicodedata.category(text[0])
if category[0] not in {'P', 'Z'}:
break
punctuation_chars.append(text[0])
text = text[1:]
special_text = u''.join(punctuation_chars) + \
(text[0] if text else u'')
span = x.makeelement('{%s}span' % XHTML_NS)
span.text = special_text
span.set('data-fake-first-letter', '1')
span.tail = text[1:]
x.text = None
x.insert(0, span)
self.style(span)._update_cssdict(cssdict)
break
else: # Element pseudo-class
for elem in matches:
self.style(elem)._update_pseudo_class(fl, cssdict)
else:
for elem in matches:
self.style(elem)._update_cssdict(cssdict)
for elem in xpath(tree, '//h:*[@style]'):
self.style(elem)._apply_style_attr(url_replacer=item.abshref)
num_pat = re.compile(r'[0-9.]+$')
for elem in xpath(tree, '//h:img[@width or @height]'):
style = self.style(elem)
# Check if either height or width is not default
is_styled = style._style.get('width', 'auto') != 'auto' or \
style._style.get('height', 'auto') != 'auto'
if not is_styled:
# Update img style dimension using width and height
upd = {}
for prop in ('width', 'height'):
val = elem.get(prop, '').strip()
try:
del elem.attrib[prop]
except:
pass
if val:
if num_pat.match(val) is not None:
val += 'px'
upd[prop] = val
if upd:
style._update_cssdict(upd)
def html_css_stylesheet():
global _html_css_stylesheet
if _html_css_stylesheet is None:
html_css = open(P('templates/html.css'), 'rb').read()
_html_css_stylesheet = parseString(html_css, validate=False)
return _html_css_stylesheet
def extract_css_into_flows(self):
inlines = defaultdict(list) # Ensure identical <style>s not repeated
sheets = {}
passthrough = getattr(self.opts, 'mobi_passthrough', False)
for item in self.oeb.manifest:
if item.media_type in OEB_STYLES:
sheet = self.data(item)
if not passthrough and not self.opts.expand_css and hasattr(item.data, 'cssText'):
condense_sheet(sheet)
sheets[item.href] = len(self.flows)
self.flows.append(sheet)
def fix_import_rules(sheet):
changed = False
for rule in sheet.cssRules.rulesOfType(CSSRule.IMPORT_RULE):
if rule.href:
href = item.abshref(rule.href)
idx = sheets.get(href, None)
if idx is not None:
idx = to_ref(idx)
rule.href = 'kindle:flow:%s?mime=text/css'%idx
changed = True
return changed
for item in self.oeb.spine:
root = self.data(item)
for link in XPath('//h:link[@href]')(root):
href = item.abshref(link.get('href'))
idx = sheets.get(href, None)
if idx is not None:
idx = to_ref(idx)
link.set('href', 'kindle:flow:%s?mime=text/css'%idx)
for tag in XPath('//h:style')(root):
p = tag.getparent()
idx = p.index(tag)
raw = tag.text
if not raw or not raw.strip():
extract(tag)
continue
sheet = css_parser.parseString(raw, validate=False)
if fix_import_rules(sheet):
raw = force_unicode(sheet.cssText, 'utf-8')
repl = etree.Element(XHTML('link'), type='text/css',
rel='stylesheet')
repl.tail='\n'
p.insert(idx, repl)
extract(tag)
inlines[raw].append(repl)
for raw, elems in iteritems(inlines):
idx = to_ref(len(self.flows))
self.flows.append(raw)
for link in elems:
link.set('href', 'kindle:flow:%s?mime=text/css'%idx)
for item in self.oeb.manifest:
if item.media_type in OEB_STYLES:
sheet = self.data(item)
if hasattr(sheet, 'cssRules'):
fix_import_rules(sheet)
for i, sheet in enumerate(tuple(self.flows)):
if hasattr(sheet, 'cssText'):
self.flows[i] = force_unicode(sheet.cssText, 'utf-8')