def change_font_in_declaration(style, old_name, new_name=None):
changed = False
ff = style.getProperty('font-family')
if ff is not None:
fams = parse_font_family(css_text(ff.propertyValue))
nfams = list(
filter(None, [new_name if x == old_name else x for x in fams]))
if fams != nfams:
if nfams:
ff.propertyValue.cssText = serialize_font_family(nfams)
else:
style.removeProperty(ff.name)
changed = True
ff = style.getProperty('font')
if ff is not None:
props = parse_font(css_text(ff.propertyValue))
fams = props.get('font-family') or []
nfams = list(
filter(None, [new_name if x == old_name else x for x in fams]))
if fams != nfams:
props['font-family'] = nfams
if nfams:
ff.propertyValue.cssText = serialize_font(props)
else:
style.removeProperty(ff.name)
changed = True
return changed
def change_font_in_declaration(style, old_name, new_name=None):
changed = False
ff = style.getProperty('font-family')
if ff is not None:
fams = parse_font_family(css_text(ff.propertyValue))
nfams = list(filter(None, [new_name if x == old_name else x for x in fams]))
if fams != nfams:
if nfams:
ff.propertyValue.cssText = serialize_font_family(nfams)
else:
style.removeProperty(ff.name)
changed = True
ff = style.getProperty('font')
if ff is not None:
props = parse_font(css_text(ff.propertyValue))
fams = props.get('font-family') or []
nfams = list(filter(None, [new_name if x == old_name else x for x in fams]))
if fams != nfams:
props['font-family'] = nfams
if nfams:
ff.propertyValue.cssText = serialize_font(props)
else:
style.removeProperty(ff.name)
changed = True
return changed
def normalize_edge(name, cssvalue):
style = {}
if isinstance(cssvalue, PropertyValue):
primitives = [css_text(v) for v in cssvalue]
else:
primitives = [css_text(cssvalue)]
if len(primitives) == 1:
value, = primitives
values = (value, value, value, value)
elif len(primitives) == 2:
vert, horiz = primitives
values = (vert, horiz, vert, horiz)
elif len(primitives) == 3:
top, horiz, bottom = primitives
values = (top, horiz, bottom, horiz)
else:
values = primitives[:4]
if '-' in name:
l, _, r = name.partition('-')
for edge, value in zip(EDGES, values):
style['%s-%s-%s' % (l, edge, r)] = value
else:
for edge, value in zip(EDGES, values):
style['%s-%s' % (name, edge)] = value
return style
def normalize_edge(name, cssvalue):
style = {}
if isinstance(cssvalue, PropertyValue):
primitives = [css_text(v) for v in cssvalue]
else:
primitives = [css_text(cssvalue)]
if len(primitives) == 1:
value, = primitives
values = (value, value, value, value)
elif len(primitives) == 2:
vert, horiz = primitives
values = (vert, horiz, vert, horiz)
elif len(primitives) == 3:
top, horiz, bottom = primitives
values = (top, horiz, bottom, horiz)
else:
values = primitives[:4]
if '-' in name:
l, _, r = name.partition('-')
for edge, value in zip(EDGES, values):
style['%s-%s-%s' % (l, edge, r)] = value
else:
for edge, value in zip(EDGES, values):
style['%s-%s' % (name, edge)] = value
return style
def remove_links_to(container, predicate):
''' predicate must be a function that takes the arguments (name, href,
fragment=None) and returns True iff the link should be removed '''
from calibre.ebooks.oeb.base import iterlinks, OEB_DOCS, OEB_STYLES, XPath, XHTML
stylepath = XPath('//h:style')
styleattrpath = XPath('//*[@style]')
changed = set()
for name, mt in iteritems(container.mime_map):
removed = False
if mt in OEB_DOCS:
root = container.parsed(name)
for el, attr, href, pos in iterlinks(root,
find_links_in_css=False):
hname = container.href_to_name(href, name)
frag = href.partition('#')[-1]
if predicate(hname, href, frag):
if attr is None:
el.text = None
else:
if el.tag == XHTML('link') or el.tag == XHTML('img'):
extract(el)
else:
del el.attrib[attr]
removed = True
for tag in stylepath(root):
if tag.text and (tag.get('type')
or 'text/css').lower() == 'text/css':
sheet = container.parse_css(tag.text)
if remove_links_in_sheet(
partial(container.href_to_name, base=name), sheet,
predicate):
tag.text = css_text(sheet)
removed = True
for tag in styleattrpath(root):
style = tag.get('style')
if style:
style = container.parse_css(style, is_declaration=True)
if remove_links_in_declaration(
partial(container.href_to_name, base=name), style,
predicate):
removed = True
tag.set('style', css_text(style))
elif mt in OEB_STYLES:
removed = remove_links_in_sheet(
partial(container.href_to_name, base=name),
container.parsed(name), predicate)
if removed:
changed.add(name)
for i in changed:
container.dirty(i)
return changed
def remove_property_value(prop, predicate):
''' Remove the Values that match the predicate from this property. If all
values of the property would be removed, the property is removed from its
parent instead. Note that this means the property must have a parent (a
CSSStyleDeclaration). '''
removed_vals = list(filter(predicate, prop.propertyValue))
if len(removed_vals) == len(prop.propertyValue):
prop.parent.removeProperty(prop.name)
else:
x = css_text(prop.propertyValue)
for v in removed_vals:
x = x.replace(css_text(v), '').strip()
prop.propertyValue.cssText = x
return bool(removed_vals)
def remove_property_value(prop, predicate):
''' Remove the Values that match the predicate from this property. If all
values of the property would be removed, the property is removed from its
parent instead. Note that this means the property must have a parent (a
CSSStyleDeclaration). '''
removed_vals = list(filter(predicate, prop.propertyValue))
if len(removed_vals) == len(prop.propertyValue):
prop.parent.removeProperty(prop.name)
else:
x = css_text(prop.propertyValue)
for v in removed_vals:
x = x.replace(css_text(v), '').strip()
prop.propertyValue.cssText = x
return bool(removed_vals)
def test_border_condensation(self):
vals = 'red solid 5px'
css = '; '.join('border-%s-%s: %s' % (edge, p, v) for edge in EDGES
for p, v in zip(BORDER_PROPS, vals.split()))
style = parseStyle(css)
condense_rule(style)
for e, p in product(EDGES, BORDER_PROPS):
self.assertFalse(style.getProperty('border-%s-%s' % (e, p)))
self.assertFalse(style.getProperty('border-%s' % e))
self.assertFalse(style.getProperty('border-%s' % p))
self.assertEqual(style.getProperty('border').value, vals)
css = '; '.join('border-%s-%s: %s' % (edge, p, v)
for edge in ('top', )
for p, v in zip(BORDER_PROPS, vals.split()))
style = parseStyle(css)
condense_rule(style)
self.assertEqual(css_text(style), 'border-top: %s' % vals)
css += ';' + '; '.join(
'border-%s-%s: %s' % (edge, p, v)
for edge in ('right', 'left', 'bottom')
for p, v in zip(BORDER_PROPS,
vals.replace('red', 'green').split()))
style = parseStyle(css)
condense_rule(style)
self.assertEqual(len(style.getProperties()), 4)
self.assertEqual(style.getProperty('border-top').value, vals)
self.assertEqual(
style.getProperty('border-left').value,
vals.replace('red', 'green'))
def change_font(container, old_name, new_name=None):
'''
Change a font family from old_name to new_name. Changes all occurrences of
the font family in stylesheets, style tags and style attributes.
If the old_name refers to an embedded font, it is removed. You can set
new_name to None to remove the font family instead of changing it.
'''
changed = False
for name, mt in tuple(iteritems(container.mime_map)):
if mt in OEB_STYLES:
sheet = container.parsed(name)
if change_font_in_sheet(container, sheet, old_name, new_name, name):
container.dirty(name)
changed = True
elif mt in OEB_DOCS:
root = container.parsed(name)
for style in root.xpath('//*[local-name() = "style"]'):
if style.text and style.get('type', 'text/css').lower() == 'text/css':
sheet = container.parse_css(style.text)
if change_font_in_sheet(container, sheet, old_name, new_name, name):
container.dirty(name)
changed = True
for elem in root.xpath('//*[@style]'):
style = elem.get('style', '')
if style:
style = container.parse_css(style, is_declaration=True)
if change_font_in_declaration(style, old_name, new_name):
style = css_text(style).strip().rstrip(';').strip()
if style:
elem.set('style', style)
else:
del elem.attrib['style']
container.dirty(name)
changed = True
return changed
def change_font(container, old_name, new_name=None):
'''
Change a font family from old_name to new_name. Changes all occurrences of
the font family in stylesheets, style tags and style attributes.
If the old_name refers to an embedded font, it is removed. You can set
new_name to None to remove the font family instead of changing it.
'''
changed = False
for name, mt in tuple(iteritems(container.mime_map)):
if mt in OEB_STYLES:
sheet = container.parsed(name)
if change_font_in_sheet(container, sheet, old_name, new_name, name):
container.dirty(name)
changed = True
elif mt in OEB_DOCS:
root = container.parsed(name)
for style in root.xpath('//*[local-name() = "style"]'):
if style.text and style.get('type', 'text/css').lower() == 'text/css':
sheet = container.parse_css(style.text)
if change_font_in_sheet(container, sheet, old_name, new_name, name):
container.dirty(name)
changed = True
for elem in root.xpath('//*[@style]'):
style = elem.get('style', '')
if style:
style = container.parse_css(style, is_declaration=True)
if change_font_in_declaration(style, old_name, new_name):
style = css_text(style).strip().rstrip(';').strip()
if style:
elem.set('style', style)
else:
del elem.attrib['style']
container.dirty(name)
changed = True
return changed
def normalize_simple_composition(name, cssvalue, composition, check_inherit=True):
if check_inherit and css_text(cssvalue) == 'inherit':
style = {k:'inherit' for k in composition}
else:
style = {k:DEFAULTS[k] for k in composition}
try:
primitives = [css_text(v) for v in cssvalue]
except TypeError:
primitives = [css_text(cssvalue)]
while primitives:
value = primitives.pop()
for key in composition:
if cssprofiles.validate(key, value):
style[key] = value
break
return style
def normalize_style_declaration(decl, sheet_name):
ans = {}
for prop in iterdeclaration(decl):
if prop.name == 'font-family':
# Needed because of https://bitbucket.org/cthedot/cssutils/issues/66/incorrect-handling-of-spaces-in-font
prop.propertyValue.cssText = serialize_font_family(parse_font_family(css_text(prop.propertyValue)))
ans[prop.name] = Values(prop.propertyValue, sheet_name, prop.priority)
return ans
def normalize_style_declaration(decl, sheet_name):
ans = {}
for prop in iterdeclaration(decl):
if prop.name == 'font-family':
# Needed because of https://bitbucket.org/cthedot/cssutils/issues/66/incorrect-handling-of-spaces-in-font
prop.propertyValue.cssText = serialize_font_family(parse_font_family(css_text(prop.propertyValue)))
ans[prop.name] = Values(prop.propertyValue, sheet_name, prop.priority)
return ans
def font_family_data_from_sheet(sheet, families):
for rule in sheet.cssRules:
if rule.type == rule.STYLE_RULE:
font_family_data_from_declaration(rule.style, families)
elif rule.type == rule.FONT_FACE_RULE:
ff = rule.style.getProperty('font-family')
if ff is not None:
for f in parse_font_family(css_text(ff.propertyValue)):
families[f] = True
def font_family_data_from_sheet(sheet, families):
for rule in sheet.cssRules:
if rule.type == rule.STYLE_RULE:
font_family_data_from_declaration(rule.style, families)
elif rule.type == rule.FONT_FACE_RULE:
ff = rule.style.getProperty('font-family')
if ff is not None:
for f in parse_font_family(css_text(ff.propertyValue)):
families[f] = True
def remove_links_to(container, predicate):
''' predicate must be a function that takes the arguments (name, href,
fragment=None) and returns True iff the link should be removed '''
from calibre.ebooks.oeb.base import iterlinks, OEB_DOCS, OEB_STYLES, XPath, XHTML
stylepath = XPath('//h:style')
styleattrpath = XPath('//*[@style]')
changed = set()
for name, mt in iteritems(container.mime_map):
removed = False
if mt in OEB_DOCS:
root = container.parsed(name)
for el, attr, href, pos in iterlinks(root, find_links_in_css=False):
hname = container.href_to_name(href, name)
frag = href.partition('#')[-1]
if predicate(hname, href, frag):
if attr is None:
el.text = None
else:
if el.tag == XHTML('link') or el.tag == XHTML('img'):
extract(el)
else:
del el.attrib[attr]
removed = True
for tag in stylepath(root):
if tag.text and (tag.get('type') or 'text/css').lower() == 'text/css':
sheet = container.parse_css(tag.text)
if remove_links_in_sheet(partial(container.href_to_name, base=name), sheet, predicate):
tag.text = css_text(sheet)
removed = True
for tag in styleattrpath(root):
style = tag.get('style')
if style:
style = container.parse_css(style, is_declaration=True)
if remove_links_in_declaration(partial(container.href_to_name, base=name), style, predicate):
removed = True
tag.set('style', css_text(style))
elif mt in OEB_STYLES:
removed = remove_links_in_sheet(partial(container.href_to_name, base=name), container.parsed(name), predicate)
if removed:
changed.add(name)
tuple(map(container.dirty, changed))
return changed
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 collect_font_face_rules(self, container, processed, spine_name, sheet,
sheet_name):
if sheet_name in processed:
sheet_rules = processed[sheet_name]
else:
sheet_rules = []
if sheet_name != spine_name:
processed[sheet_name] = sheet_rules
for rule, base_name, rule_index in iterrules(
container, sheet_name, rules=sheet,
rule_type='FONT_FACE_RULE'):
cssdict = {}
for prop in iterdeclaration(rule.style):
if prop.name == 'font-family':
cssdict['font-family'] = [
icu_lower(x) for x in parse_font_family(
css_text(prop.propertyValue))
]
elif prop.name.startswith('font-'):
cssdict[prop.name] = prop.propertyValue[0].value
elif prop.name == 'src':
for val in prop.propertyValue:
x = val.value
fname = container.href_to_name(x, sheet_name)
if container.has_name(fname):
cssdict['src'] = fname
break
else:
container.log.warn(
'The @font-face rule refers to a font file that does not exist in the book: %s'
% css_text(prop.propertyValue))
if 'src' not in cssdict:
continue
ff = cssdict.get('font-family')
if not ff or ff[0] in bad_fonts:
continue
normalize_font_properties(cssdict)
prepare_font_rule(cssdict)
sheet_rules.append(cssdict)
self.font_rule_map[spine_name].extend(sheet_rules)
def font_family_data_from_declaration(style, families):
font_families = []
f = style.getProperty('font')
if f is not None:
f = normalize_font(f.propertyValue, font_family_as_list=True).get('font-family', None)
if f is not None:
font_families = [unquote(x) for x in f]
f = style.getProperty('font-family')
if f is not None:
font_families = parse_font_family(css_text(f.propertyValue))
for f in font_families:
families[f] = families.get(f, False)
def font_family_data_from_declaration(style, families):
font_families = []
f = style.getProperty('font')
if f is not None:
f = normalize_font(f.propertyValue, font_family_as_list=True).get('font-family', None)
if f is not None:
font_families = [unquote(x) for x in f]
f = style.getProperty('font-family')
if f is not None:
font_families = parse_font_family(css_text(f.propertyValue))
for f in font_families:
families[f] = families.get(f, False)
def process_fonts(self):
''' Make sure all fonts are embeddable. Also remove some fonts that cause problems. '''
from calibre.ebooks.oeb.base import urlnormalize, css_text
from calibre.utils.fonts.utils import remove_embed_restriction
processed = set()
for item in list(self.oeb.manifest):
if not hasattr(item.data, 'cssRules'):
continue
for i, rule in enumerate(item.data.cssRules):
if rule.type == rule.FONT_FACE_RULE:
try:
s = rule.style
src = s.getProperty('src').propertyValue[0].uri
except:
continue
path = item.abshref(src)
ff = self.oeb.manifest.hrefs.get(urlnormalize(path), None)
if ff is None:
continue
raw = nraw = ff.data
if path not in processed:
processed.add(path)
try:
nraw = remove_embed_restriction(raw)
except:
continue
if nraw != raw:
ff.data = nraw
self.oeb.container.write(path, nraw)
elif iswindows and rule.type == rule.STYLE_RULE:
from tinycss.fonts3 import parse_font_family, serialize_font_family
s = rule.style
f = s.getProperty(u'font-family')
if f is not None:
font_families = parse_font_family(
css_text(f.propertyValue))
ff = [
x for x in font_families if x.lower() != u'courier'
]
if len(ff) != len(font_families):
if 'courier' not in self.filtered_font_warnings:
# See https://bugs.launchpad.net/bugs/1665835
self.filtered_font_warnings.add(u'courier')
self.log.warn(
u'Removing courier font family as it does not render on windows'
)
f.propertyValue.cssText = serialize_font_family(
ff or [u'monospace'])
def sort_css(self):
from calibre.gui2.dialogs.confirm_delete import confirm
if confirm(_('Sorting CSS rules can in rare cases change the effective styles applied to the book.'
' Are you sure you want to proceed?'), 'edit-book-confirm-sort-css', parent=self, config_set=tprefs):
c = self.textCursor()
c.beginEditBlock()
c.movePosition(c.Start), c.movePosition(c.End, c.KeepAnchor)
text = unicode_type(c.selectedText()).replace(PARAGRAPH_SEPARATOR, '\n').rstrip('\0')
from calibre.ebooks.oeb.polish.css import sort_sheet
text = css_text(sort_sheet(current_container(), text))
c.insertText(text)
c.movePosition(c.Start)
c.endEditBlock()
self.setTextCursor(c)
def sort_css(self):
from calibre.gui2.dialogs.confirm_delete import confirm
if confirm(_('Sorting CSS rules can in rare cases change the effective styles applied to the book.'
' Are you sure you want to proceed?'), 'edit-book-confirm-sort-css', parent=self, config_set=tprefs):
c = self.textCursor()
c.beginEditBlock()
c.movePosition(c.Start), c.movePosition(c.End, c.KeepAnchor)
text = unicode_type(c.selectedText()).replace(PARAGRAPH_SEPARATOR, '\n').rstrip('\0')
from calibre.ebooks.oeb.polish.css import sort_sheet
text = css_text(sort_sheet(current_container(), text))
c.insertText(text)
c.movePosition(c.Start)
c.endEditBlock()
self.setTextCursor(c)
def change_font_in_sheet(container, sheet, old_name, new_name, sheet_name):
changed = False
removals = []
for rule in sheet.cssRules:
if rule.type == rule.STYLE_RULE:
changed |= change_font_in_declaration(rule.style, old_name, new_name)
elif rule.type == rule.FONT_FACE_RULE:
ff = rule.style.getProperty('font-family')
if ff is not None:
families = {x for x in parse_font_family(css_text(ff.propertyValue))}
if old_name in families:
changed = True
removals.append(rule)
for rule in reversed(removals):
remove_embedded_font(container, sheet, rule, sheet_name)
return changed
def change_font_in_sheet(container, sheet, old_name, new_name, sheet_name):
changed = False
removals = []
for rule in sheet.cssRules:
if rule.type == rule.STYLE_RULE:
changed |= change_font_in_declaration(rule.style, old_name, new_name)
elif rule.type == rule.FONT_FACE_RULE:
ff = rule.style.getProperty('font-family')
if ff is not None:
families = {x for x in parse_font_family(css_text(ff.propertyValue))}
if old_name in families:
changed = True
removals.append(rule)
for rule in reversed(removals):
remove_embedded_font(container, sheet, rule, sheet_name)
return changed
def get_font_properties(rule, default=None):
'''
Given a CSS rule, extract normalized font properties from
it. Note that shorthand font property should already have been expanded
by the CSS flattening code.
'''
props = {}
s = rule.style
for q in ('font-family', 'src', 'font-weight', 'font-stretch',
'font-style'):
g = 'uri' if q == 'src' else 'value'
try:
val = s.getProperty(q).propertyValue[0]
val = getattr(val, g)
if q == 'font-family':
val = parse_font_family(
css_text(s.getProperty(q).propertyValue))
if val and val[0] == 'inherit':
val = None
except (IndexError, KeyError, AttributeError, TypeError, ValueError):
val = None if q in {'src', 'font-family'} else default
if q in {'font-weight', 'font-stretch', 'font-style'}:
val = unicode_type(val).lower() if (val or val == 0) else val
if val == 'inherit':
val = default
if q == 'font-weight':
val = {'normal': '400', 'bold': '700'}.get(val, val)
if val not in {
'100', '200', '300', '400', '500', '600', '700', '800',
'900', 'bolder', 'lighter'
}:
val = default
if val == 'normal':
val = '400'
elif q == 'font-style':
if val not in {'normal', 'italic', 'oblique'}:
val = default
elif q == 'font-stretch':
if val not in {
'normal', 'ultra-condensed', 'extra-condensed',
'condensed', 'semi-condensed', 'semi-expanded', 'expanded',
'extra-expanded', 'ultra-expanded'
}:
val = default
props[q] = val
return props
def normalize_font(cssvalue, font_family_as_list=False):
# See https://developer.mozilla.org/en-US/docs/Web/CSS/font
composition = font_composition
val = css_text(cssvalue)
if val == 'inherit':
ans = {k:'inherit' for k in composition}
elif val in {'caption', 'icon', 'menu', 'message-box', 'small-caption', 'status-bar'}:
ans = {k:DEFAULTS[k] for k in composition}
else:
ans = {k:DEFAULTS[k] for k in composition}
ans.update(parse_font(val))
if font_family_as_list:
if isinstance(ans['font-family'], string_or_bytes):
ans['font-family'] = [x.strip() for x in ans['font-family'].split(',')]
else:
if not isinstance(ans['font-family'], string_or_bytes):
ans['font-family'] = serialize_font_family(ans['font-family'])
return ans
def process_fonts(self):
''' Make sure all fonts are embeddable. Also remove some fonts that cause problems. '''
from calibre.ebooks.oeb.base import urlnormalize, css_text
from calibre.utils.fonts.utils import remove_embed_restriction
processed = set()
for item in list(self.oeb.manifest):
if not hasattr(item.data, 'cssRules'):
continue
for i, rule in enumerate(item.data.cssRules):
if rule.type == rule.FONT_FACE_RULE:
try:
s = rule.style
src = s.getProperty('src').propertyValue[0].uri
except:
continue
path = item.abshref(src)
ff = self.oeb.manifest.hrefs.get(urlnormalize(path), None)
if ff is None:
continue
raw = nraw = ff.data
if path not in processed:
processed.add(path)
try:
nraw = remove_embed_restriction(raw)
except:
continue
if nraw != raw:
ff.data = nraw
self.oeb.container.write(path, nraw)
elif iswindows and rule.type == rule.STYLE_RULE:
from tinycss.fonts3 import parse_font_family, serialize_font_family
s = rule.style
f = s.getProperty(u'font-family')
if f is not None:
font_families = parse_font_family(css_text(f.propertyValue))
ff = [x for x in font_families if x.lower() != u'courier']
if len(ff) != len(font_families):
if 'courier' not in self.filtered_font_warnings:
# See https://bugs.launchpad.net/bugs/1665835
self.filtered_font_warnings.add(u'courier')
self.log.warn(u'Removing courier font family as it does not render on windows')
f.propertyValue.cssText = serialize_font_family(ff or [u'monospace'])
def get_font_properties(rule, default=None):
'''
Given a CSS rule, extract normalized font properties from
it. Note that shorthand font property should already have been expanded
by the CSS flattening code.
'''
props = {}
s = rule.style
for q in ('font-family', 'src', 'font-weight', 'font-stretch',
'font-style'):
g = 'uri' if q == 'src' else 'value'
try:
val = s.getProperty(q).propertyValue[0]
val = getattr(val, g)
if q == 'font-family':
val = parse_font_family(css_text(s.getProperty(q).propertyValue))
if val and val[0] == 'inherit':
val = None
except (IndexError, KeyError, AttributeError, TypeError, ValueError):
val = None if q in {'src', 'font-family'} else default
if q in {'font-weight', 'font-stretch', 'font-style'}:
val = unicode_type(val).lower() if (val or val == 0) else val
if val == 'inherit':
val = default
if q == 'font-weight':
val = {'normal':'400', 'bold':'700'}.get(val, val)
if val not in {'100', '200', '300', '400', '500', '600', '700',
'800', '900', 'bolder', 'lighter'}:
val = default
if val == 'normal':
val = '400'
elif q == 'font-style':
if val not in {'normal', 'italic', 'oblique'}:
val = default
elif q == 'font-stretch':
if val not in {'normal', 'ultra-condensed', 'extra-condensed',
'condensed', 'semi-condensed', 'semi-expanded',
'expanded', 'extra-expanded', 'ultra-expanded'}:
val = default
props[q] = val
return props
def test_border_condensation(self):
vals = 'red solid 5px'
css = '; '.join('border-%s-%s: %s' % (edge, p, v) for edge in EDGES for p, v in zip(BORDER_PROPS, vals.split()))
style = parseStyle(css)
condense_rule(style)
for e, p in product(EDGES, BORDER_PROPS):
self.assertFalse(style.getProperty('border-%s-%s' % (e, p)))
self.assertFalse(style.getProperty('border-%s' % e))
self.assertFalse(style.getProperty('border-%s' % p))
self.assertEqual(style.getProperty('border').value, vals)
css = '; '.join('border-%s-%s: %s' % (edge, p, v) for edge in ('top',) for p, v in zip(BORDER_PROPS, vals.split()))
style = parseStyle(css)
condense_rule(style)
self.assertEqual(css_text(style), 'border-top: %s' % vals)
css += ';' + '; '.join('border-%s-%s: %s' % (edge, p, v) for edge in ('right', 'left', 'bottom') for p, v in
zip(BORDER_PROPS, vals.replace('red', 'green').split()))
style = parseStyle(css)
condense_rule(style)
self.assertEqual(len(style.getProperties()), 4)
self.assertEqual(style.getProperty('border-top').value, vals)
self.assertEqual(style.getProperty('border-left').value, vals.replace('red', 'green'))
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 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 __init__(self, oeb, opts, replace_previous_inline_toc=True, ignore_existing_toc=False):
self.oeb, self.opts, self.log = oeb, opts, oeb.log
self.title = opts.toc_title or DEFAULT_TITLE
self.at_start = opts.mobi_toc_at_start
self.generated_item = None
self.added_toc_guide_entry = False
self.has_toc = oeb.toc and oeb.toc.count() > 1
self.tocitem = tocitem = None
if replace_previous_inline_toc:
tocitem = self.tocitem = find_previous_calibre_inline_toc(oeb)
if ignore_existing_toc and 'toc' in oeb.guide:
oeb.guide.remove('toc')
if 'toc' in oeb.guide:
# Remove spurious toc entry from guide if it is not in spine or it
# does not have any hyperlinks
href = urlnormalize(oeb.guide['toc'].href.partition('#')[0])
if href in oeb.manifest.hrefs:
item = oeb.manifest.hrefs[href]
if (hasattr(item.data, 'xpath') and
XPath('//h:a[@href]')(item.data)):
if oeb.spine.index(item) < 0:
oeb.spine.add(item, linear=False)
return
elif self.has_toc:
oeb.guide.remove('toc')
else:
oeb.guide.remove('toc')
if (not self.has_toc or 'toc' in oeb.guide or opts.no_inline_toc or
getattr(opts, 'mobi_passthrough', False)):
return
self.log('\tGenerating in-line ToC')
embed_css = ''
s = getattr(oeb, 'store_embed_font_rules', None)
if getattr(s, 'body_font_family', None):
css = [css_text(x) for x in s.rules] + [
'body { font-family: %s }'%s.body_font_family]
embed_css = '\n\n'.join(css)
root = safe_xml_fromstring(TEMPLATE.format(xhtmlns=XHTML_NS,
title=self.title, embed_css=embed_css,
extra_css=(opts.extra_css or '')))
parent = XPath('//h:ul')(root)[0]
parent.text = '\n\t'
for child in self.oeb.toc:
self.process_toc_node(child, parent)
if tocitem is not None:
href = tocitem.href
if oeb.spine.index(tocitem) > -1:
oeb.spine.remove(tocitem)
tocitem.data = root
else:
id, href = oeb.manifest.generate('contents', 'contents.xhtml')
tocitem = self.generated_item = oeb.manifest.add(id, href, XHTML_MIME,
data=root)
if self.at_start:
oeb.spine.insert(0, tocitem, linear=True)
else:
oeb.spine.add(tocitem, linear=False)
oeb.guide.add('toc', 'Table of Contents', href)
def __init__(self, oeb, opts, replace_previous_inline_toc=True, ignore_existing_toc=False):
self.oeb, self.opts, self.log = oeb, opts, oeb.log
self.title = opts.toc_title or DEFAULT_TITLE
self.at_start = opts.mobi_toc_at_start
self.generated_item = None
self.added_toc_guide_entry = False
self.has_toc = oeb.toc and oeb.toc.count() > 1
self.tocitem = tocitem = None
if replace_previous_inline_toc:
tocitem = self.tocitem = find_previous_calibre_inline_toc(oeb)
if ignore_existing_toc and 'toc' in oeb.guide:
oeb.guide.remove('toc')
if 'toc' in oeb.guide:
# Remove spurious toc entry from guide if it is not in spine or it
# does not have any hyperlinks
href = urlnormalize(oeb.guide['toc'].href.partition('#')[0])
if href in oeb.manifest.hrefs:
item = oeb.manifest.hrefs[href]
if (hasattr(item.data, 'xpath') and
XPath('//h:a[@href]')(item.data)):
if oeb.spine.index(item) < 0:
oeb.spine.add(item, linear=False)
return
elif self.has_toc:
oeb.guide.remove('toc')
else:
oeb.guide.remove('toc')
if (not self.has_toc or 'toc' in oeb.guide or opts.no_inline_toc or
getattr(opts, 'mobi_passthrough', False)):
return
self.log('\tGenerating in-line ToC')
embed_css = ''
s = getattr(oeb, 'store_embed_font_rules', None)
if getattr(s, 'body_font_family', None):
css = [css_text(x) for x in s.rules] + [
'body { font-family: %s }'%s.body_font_family]
embed_css = '\n\n'.join(css)
root = etree.fromstring(TEMPLATE.format(xhtmlns=XHTML_NS,
title=self.title, embed_css=embed_css,
extra_css=(opts.extra_css or '')))
parent = XPath('//h:ul')(root)[0]
parent.text = '\n\t'
for child in self.oeb.toc:
self.process_toc_node(child, parent)
if tocitem is not None:
href = tocitem.href
if oeb.spine.index(tocitem) > -1:
oeb.spine.remove(tocitem)
tocitem.data = root
else:
id, href = oeb.manifest.generate('contents', 'contents.xhtml')
tocitem = self.generated_item = oeb.manifest.add(id, href, XHTML_MIME,
data=root)
if self.at_start:
oeb.spine.insert(0, tocitem, linear=True)
else:
oeb.spine.add(tocitem, linear=False)
oeb.guide.add('toc', 'Table of Contents', href)
def subset_all_fonts(container, font_stats, report):
remove = set()
total_old = total_new = 0
changed = False
for name, mt in iter_subsettable_fonts(container):
chars = font_stats.get(name, set())
with container.open(name, 'rb') as f:
f.seek(0, os.SEEK_END)
total_old += f.tell()
if not chars:
remove.add(name)
report(_('Removed unused font: %s') % name)
continue
with container.open(name, 'r+b') as f:
raw = f.read()
try:
font_name = get_font_names(raw)[-1]
except Exception as e:
container.log.warning(
'Corrupted font: %s, ignoring. Error: %s' %
(name, as_unicode(e)))
continue
warnings = []
container.log('Subsetting font: %s' % (font_name or name))
try:
nraw, old_sizes, new_sizes = subset(raw,
chars,
warnings=warnings)
except UnsupportedFont as e:
container.log.warning(
'Unsupported font: %s, ignoring. Error: %s' %
(name, as_unicode(e)))
continue
for w in warnings:
container.log.warn(w)
olen = sum(itervalues(old_sizes))
nlen = sum(itervalues(new_sizes))
total_new += len(nraw)
if nlen == olen:
report(_('The font %s was already subset') % font_name)
else:
report(
_('Decreased the font {0} to {1} of its original size').
format(font_name, ('%.1f%%' % (nlen / olen * 100))))
changed = True
f.seek(0), f.truncate(), f.write(nraw)
for name in remove:
container.remove_item(name)
changed = True
if remove:
for name, mt in iteritems(container.mime_map):
if mt in OEB_STYLES:
sheet = container.parsed(name)
if remove_font_face_rules(container, sheet, remove, name):
container.dirty(name)
elif mt in OEB_DOCS:
for style in XPath('//h:style')(container.parsed(name)):
if style.get('type',
'text/css') == 'text/css' and style.text:
sheet = container.parse_css(style.text, name)
if remove_font_face_rules(container, sheet, remove,
name):
style.text = css_text(sheet)
container.dirty(name)
if total_old > 0:
report(
_('Reduced total font size to %.1f%% of original') %
(total_new / total_old * 100))
else:
report(_('No embedded fonts found'))
return changed
def collect_font_face_rules(self, container, processed, spine_name, sheet, sheet_name):
if sheet_name in processed:
sheet_rules = processed[sheet_name]
else:
sheet_rules = []
if sheet_name != spine_name:
processed[sheet_name] = sheet_rules
for rule, base_name, rule_index in iterrules(container, sheet_name, rules=sheet, rule_type='FONT_FACE_RULE'):
cssdict = {}
for prop in iterdeclaration(rule.style):
if prop.name == 'font-family':
cssdict['font-family'] = [icu_lower(x) for x in parse_font_family(css_text(prop.propertyValue))]
elif prop.name.startswith('font-'):
cssdict[prop.name] = prop.propertyValue[0].value
elif prop.name == 'src':
for val in prop.propertyValue:
x = val.value
fname = container.href_to_name(x, sheet_name)
if container.has_name(fname):
cssdict['src'] = fname
break
else:
container.log.warn('The @font-face rule refers to a font file that does not exist in the book: %s' % css_text(prop.propertyValue))
if 'src' not in cssdict:
continue
ff = cssdict.get('font-family')
if not ff or ff[0] in bad_fonts:
continue
normalize_font_properties(cssdict)
prepare_font_rule(cssdict)
sheet_rules.append(cssdict)
self.font_rule_map[spine_name].extend(sheet_rules)
def css(d):
return css_text(d).replace('\n', ' ')
def subset_all_fonts(container, font_stats, report):
remove = set()
total_old = total_new = 0
changed = False
for name, mt in iter_subsettable_fonts(container):
chars = font_stats.get(name, set())
with container.open(name, 'rb') as f:
f.seek(0, os.SEEK_END)
total_old += f.tell()
if not chars:
remove.add(name)
report(_('Removed unused font: %s')%name)
continue
with container.open(name, 'r+b') as f:
raw = f.read()
try:
font_name = get_font_names(raw)[-1]
except Exception as e:
container.log.warning(
'Corrupted font: %s, ignoring. Error: %s'%(
name, as_unicode(e)))
continue
warnings = []
container.log('Subsetting font: %s'%(font_name or name))
try:
nraw, old_sizes, new_sizes = subset(raw, chars,
warnings=warnings)
except UnsupportedFont as e:
container.log.warning(
'Unsupported font: %s, ignoring. Error: %s'%(
name, as_unicode(e)))
continue
for w in warnings:
container.log.warn(w)
olen = sum(itervalues(old_sizes))
nlen = sum(itervalues(new_sizes))
total_new += len(nraw)
if nlen == olen:
report(_('The font %s was already subset')%font_name)
else:
report(_('Decreased the font {0} to {1} of its original size').format(
font_name, ('%.1f%%' % (nlen/olen * 100))))
changed = True
f.seek(0), f.truncate(), f.write(nraw)
for name in remove:
container.remove_item(name)
changed = True
if remove:
for name, mt in iteritems(container.mime_map):
if mt in OEB_STYLES:
sheet = container.parsed(name)
if remove_font_face_rules(container, sheet, remove, name):
container.dirty(name)
elif mt in OEB_DOCS:
for style in XPath('//h:style')(container.parsed(name)):
if style.get('type', 'text/css') == 'text/css' and style.text:
sheet = container.parse_css(style.text, name)
if remove_font_face_rules(container, sheet, remove, name):
style.text = css_text(sheet)
container.dirty(name)
if total_old > 0:
report(_('Reduced total font size to %.1f%% of original')%(
total_new/total_old*100))
else:
report(_('No embedded fonts found'))
return changed
def css(d):
return css_text(d).replace('\n', ' ')
def cssText(self):
' This will return either a string or a tuple of strings '
if len(self) == 1:
return css_text(self[0])
return tuple(css_text(x) for x in self)
def cssText(self):
' This will return either a string or a tuple of strings '
if len(self) == 1:
return css_text(self[0])
return tuple(css_text(x) for x in self)