class Convert(object):
def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer()
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager(self.docx.namespace)
self.images_manager = ImagesManager(self.oeb, self.docx.document_relationships)
self.fonts_manager = FontsManager(self.docx.namespace, self.oeb, self.opts)
self.blocks = Blocks(self.docx.namespace, self.styles_manager)
for item in self.oeb.spine:
self.process_item(item)
self.styles_manager.finalize(self.blocks.all_blocks)
self.write()
def process_item(self, item):
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile)
self.abshref = self.images_manager.abshref = item.abshref
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self, html_tag, stylizer, is_first_tag=False):
tagname = barename(html_tag.tag)
if tagname in {'script', 'style', 'title', 'meta'}:
return
tag_style = stylizer.style(html_tag)
if tag_style.is_hidden:
return
display = tag_style._get('display')
if display in {'inline', 'inline-block'} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
# TODO: Implement this
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.end_current_block()
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and tag_style['float'] in {'left', 'right'}:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
for child in html_tag.iterchildren('*'):
self.process_tag(child, stylizer)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
if display == 'table-row':
return # We ignore the tail for these tags
ignore_whitespace_tail = is_block or display.startswith('table')
if not is_first_tag and html_tag.tail and (not ignore_whitespace_tail or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.tail, stylizer.style(html_tag.getparent()), is_parent_style=True)
def add_block_tag(self, tagname, html_tag, tag_style, stylizer, is_table_cell=False):
block = self.blocks.start_new_block(html_tag, tag_style, is_table_cell=is_table_cell)
if tagname == 'img':
self.images_manager.add_image(html_tag, block, stylizer)
else:
if html_tag.text:
block.add_text(html_tag.text, tag_style, ignore_leading_whitespace=True, is_parent_style=True)
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(html_tag.getparent().iterchildren('*'))[-1]:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(tag_style['clear'], 'none'))
elif tagname == 'img':
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
self.images_manager.add_image(html_tag, block, stylizer)
else:
if html_tag.text:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.text, tag_style, is_parent_style=False)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts, self.docx.fonts)
class Convert(object):
def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer()
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager()
self.images_manager = ImagesManager(self.oeb, self.docx.document_relationships)
self.fonts_manager = FontsManager(self.oeb, self.opts)
self.blocks = Blocks(self.styles_manager)
for item in self.oeb.spine:
self.process_item(item)
self.styles_manager.finalize(self.blocks.all_blocks)
self.write()
def process_item(self, item):
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile)
self.abshref = self.images_manager.abshref = item.abshref
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self, html_tag, stylizer, is_first_tag=False):
tagname = barename(html_tag.tag)
if tagname in {'script', 'style', 'title', 'meta'}:
return
tag_style = stylizer.style(html_tag)
if tag_style.is_hidden:
return
display = tag_style._get('display')
if display in {'inline', 'inline-block'} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
# TODO: Implement this
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and tag_style['float'] in {'left', 'right'}:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
for child in html_tag.iterchildren('*'):
self.process_tag(child, stylizer)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
if display == 'table-row':
return # We ignore the tail for these tags
if not is_first_tag and html_tag.tail and (not is_block or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.tail, stylizer.style(html_tag.getparent()), is_parent_style=True)
def add_block_tag(self, tagname, html_tag, tag_style, stylizer, is_table_cell=False):
block = self.blocks.start_new_block(html_tag, tag_style, is_table_cell=is_table_cell)
if tagname == 'img':
self.images_manager.add_image(html_tag, block, stylizer)
else:
if html_tag.text:
block.add_text(html_tag.text, tag_style, ignore_leading_whitespace=True, is_parent_style=True)
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(html_tag.getparent().iterchildren('*'))[-1]:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(tag_style['clear'], 'none'))
elif tagname == 'img':
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
self.images_manager.add_image(html_tag, block, stylizer)
else:
if html_tag.text:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.text, tag_style, is_parent_style=False)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts, self.docx.fonts)
class Convert:
# Word does not apply default styling to hyperlinks, so we ensure they get
# default styling (the conversion pipeline does not apply any styling to
# them).
base_css = '''
a[href] { text-decoration: underline; color: blue }
'''
def __init__(self, oeb, docx, mi, add_cover, add_toc):
self.oeb, self.docx, self.add_cover, self.add_toc = oeb, docx, add_cover, add_toc
self.log, self.opts = docx.log, docx.opts
self.mi = mi
self.cover_img = None
p = self.opts.output_profile
p.width_pts, p.height_pts = page_effective_area(self.opts)
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer(base_css=self.base_css)
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager(self.docx.namespace, self.log,
self.mi.language)
self.links_manager = LinksManager(self.docx.namespace,
self.docx.document_relationships,
self.log)
self.images_manager = ImagesManager(self.oeb,
self.docx.document_relationships,
self.opts)
self.lists_manager = ListsManager(self.docx)
self.fonts_manager = FontsManager(self.docx.namespace, self.oeb,
self.opts)
self.blocks = Blocks(self.docx.namespace, self.styles_manager,
self.links_manager)
self.current_link = self.current_lang = None
for item in self.oeb.spine:
self.log.debug('Processing', item.href)
self.process_item(item)
if self.add_toc:
self.links_manager.process_toc_links(self.oeb)
if self.add_cover and self.oeb.metadata.cover and str(
self.oeb.metadata.cover[0]) in self.oeb.manifest.ids:
cover_id = str(self.oeb.metadata.cover[0])
item = self.oeb.manifest.ids[cover_id]
self.cover_img = self.images_manager.read_image(item.href)
all_blocks = self.blocks.all_blocks
remove_blocks = []
for i, block in enumerate(all_blocks):
try:
nb = all_blocks[i + 1]
except IndexError:
break
block.resolve_skipped(nb)
if block.skipped:
remove_blocks.append((i, block))
for pos, block in reversed(remove_blocks):
self.blocks.delete_block_at(pos)
self.blocks.all_blocks[0].is_first_block = True
self.blocks.apply_page_break_after()
self.blocks.resolve_language()
if self.cover_img is not None:
self.cover_img = self.images_manager.create_cover_markup(
self.cover_img, self.opts.preserve_cover_aspect_ratio,
*page_size(self.opts))
self.lists_manager.finalize(all_blocks)
self.styles_manager.finalize(all_blocks)
self.write()
def process_item(self, item):
self.current_item = item
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data,
item.href,
self.oeb,
self.opts,
profile=self.opts.output_profile,
base_css=self.base_css)
self.abshref = self.images_manager.abshref = item.abshref
self.current_lang = lang_for_tag(
item.data) or self.styles_manager.document_lang
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
self.blocks.top_bookmark = self.links_manager.bookmark_for_anchor(
self.links_manager.top_anchor, self.current_item, body)
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self,
html_tag,
stylizer,
is_first_tag=False,
float_spec=None):
tagname = barename(html_tag.tag)
tag_style = stylizer.style(html_tag)
ignore_tag_contents = tagname in {'script', 'style', 'title', 'meta'
} or tag_style.is_hidden
display = tag_style._get('display')
is_block = False
if not ignore_tag_contents:
previous_link = self.current_link
if tagname == 'a' and html_tag.get('href'):
self.current_link = (self.current_item, html_tag.get('href'),
html_tag.get('title'))
previous_lang = self.current_lang
tag_lang = lang_for_tag(html_tag)
if tag_lang:
self.current_lang = tag_lang
is_float = tag_style['float'] in {'left', 'right'
} and not is_first_tag
if float_spec is None and is_float:
float_spec = FloatSpec(self.docx.namespace, html_tag,
tag_style)
if display in {
'inline', 'inline-block'
} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
if is_float and float_spec.is_dropcaps:
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
float_spec=float_spec)
float_spec = None
else:
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
is_list_item=True)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.end_current_block()
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and is_float:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
if tagname == 'hr':
for edge in 'right bottom left'.split():
tag_style.set('border-%s-style' % edge, 'none')
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
float_spec=float_spec)
for child in html_tag.iterchildren():
if isinstance(getattr(child, 'tag', None), string_or_bytes):
self.process_tag(child, stylizer, float_spec=float_spec)
else: # Comment/PI/etc.
tail = getattr(child, 'tail', None)
if tail:
block = self.create_block_from_parent(
html_tag, stylizer)
block.add_text(tail,
tag_style,
is_parent_style=False,
link=self.current_link,
lang=self.current_lang)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
self.current_link = previous_link
self.current_lang = previous_lang
# Now, process the tail if any
if display == 'table-row':
return # We ignore the tail for these tags
ignore_whitespace_tail = is_block or display.startswith('table')
if not is_first_tag and html_tag.tail and (
not ignore_whitespace_tail or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text(html_tag.tail,
stylizer.style(html_tag.getparent()),
is_parent_style=True,
link=self.current_link,
lang=self.current_lang)
def create_block_from_parent(self, html_tag, stylizer):
parent = html_tag.getparent()
block = self.blocks.current_or_new_block(parent,
stylizer.style(parent))
# Do not inherit page-break-before from parent
block.page_break_before = False
return block
def add_block_tag(self,
tagname,
html_tag,
tag_style,
stylizer,
is_table_cell=False,
float_spec=None,
is_list_item=False):
block = self.blocks.start_new_block(html_tag,
tag_style,
is_table_cell=is_table_cell,
float_spec=float_spec,
is_list_item=is_list_item)
anchor = html_tag.get('id') or html_tag.get('name')
if anchor:
block.bookmarks.add(self.bookmark_for_anchor(anchor, html_tag))
if tagname == 'img':
self.images_manager.add_image(html_tag,
block,
stylizer,
as_block=True)
else:
text = html_tag.text
is_list_item = tagname == 'li'
has_sublist = is_list_item and len(html_tag) and barename(
html_tag[0].tag) in ('ul', 'ol') and len(html_tag[0])
if text and has_sublist and not text.strip():
text = '' # whitespace only, ignore
if text:
block.add_text(text,
tag_style,
ignore_leading_whitespace=True,
is_parent_style=True,
link=self.current_link,
lang=self.current_lang)
elif has_sublist:
block.force_not_empty = True
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
anchor = html_tag.get('id') or html_tag.get('name') or None
bmark = None
if anchor:
bmark = self.bookmark_for_anchor(anchor, html_tag)
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(
html_tag.getparent().iterchildren('*'))[-1]:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_break(clear={
'both': 'all',
'left': 'left',
'right': 'right'
}.get(tag_style['clear'], 'none'),
bookmark=bmark)
elif tagname == 'img':
block = self.create_block_from_parent(html_tag, stylizer)
self.images_manager.add_image(html_tag,
block,
stylizer,
bookmark=bmark)
else:
if html_tag.text:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text(html_tag.text,
tag_style,
is_parent_style=False,
bookmark=bmark,
link=self.current_link,
lang=self.current_lang)
elif bmark:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text('',
tag_style,
is_parent_style=False,
bookmark=bmark,
link=self.current_link,
lang=self.current_lang)
def bookmark_for_anchor(self, anchor, html_tag):
return self.links_manager.bookmark_for_anchor(anchor,
self.current_item,
html_tag)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
if self.links_manager.toc:
self.links_manager.serialize_toc(
body, self.styles_manager.primary_heading_style)
if self.cover_img is not None:
self.images_manager.write_cover_block(body, self.cover_img)
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles,
self.docx.font_table,
self.docx.embedded_fonts, self.docx.fonts)
self.lists_manager.serialize(self.docx.numbering)
class Convert(object):
# Word does not apply default styling to hyperlinks, so we ensure they get
# default styling (the conversion pipeline does not apply any styling to
# them).
base_css = '''
a[href] { text-decoration: underline; color: blue }
'''
def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer(base_css=self.base_css)
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager(self.docx.namespace)
self.links_manager = LinksManager(self.docx.namespace, self.docx.document_relationships)
self.images_manager = ImagesManager(self.oeb, self.docx.document_relationships)
self.lists_manager = ListsManager(self.docx)
self.fonts_manager = FontsManager(self.docx.namespace, self.oeb, self.opts)
self.blocks = Blocks(self.docx.namespace, self.styles_manager, self.links_manager)
self.current_link = None
for item in self.oeb.spine:
self.process_item(item)
all_blocks = self.blocks.all_blocks
remove_blocks = []
for i, block in enumerate(all_blocks):
try:
nb = all_blocks[i+1]
except IndexError:
break
block.resolve_skipped(nb)
if block.skipped:
remove_blocks.append((i, block))
for pos, block in reversed(remove_blocks):
self.blocks.delete_block_at(pos)
self.blocks.all_blocks[0].is_first_block = True
self.lists_manager.finalize(all_blocks)
self.styles_manager.finalize(all_blocks)
self.write()
def process_item(self, item):
self.current_item = item
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile, base_css=self.base_css)
self.abshref = self.images_manager.abshref = item.abshref
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
body.set('id', body.get('id', None) or self.links_manager.top_anchor)
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self, html_tag, stylizer, is_first_tag=False, float_spec=None):
tagname = barename(html_tag.tag)
if tagname in {'script', 'style', 'title', 'meta'}:
return
tag_style = stylizer.style(html_tag)
if tag_style.is_hidden:
return
previous_link = self.current_link
if tagname == 'a' and html_tag.get('href'):
self.current_link = (self.current_item, html_tag.get('href'), html_tag.get('title'))
display = tag_style._get('display')
is_float = tag_style['float'] in {'left', 'right'} and not is_first_tag
if float_spec is None and is_float:
float_spec = FloatSpec(self.docx.namespace, html_tag, tag_style)
if display in {'inline', 'inline-block'} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
if is_float and float_spec.is_dropcaps:
self.add_block_tag(tagname, html_tag, tag_style, stylizer, float_spec=float_spec)
float_spec = None
else:
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_list_item=True)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.end_current_block()
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and is_float:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
self.add_block_tag(tagname, html_tag, tag_style, stylizer, float_spec=float_spec)
for child in html_tag.iterchildren('*'):
self.process_tag(child, stylizer, float_spec=float_spec)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
self.current_link = previous_link
if display == 'table-row':
return # We ignore the tail for these tags
ignore_whitespace_tail = is_block or display.startswith('table')
if not is_first_tag and html_tag.tail and (not ignore_whitespace_tail or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.tail, stylizer.style(html_tag.getparent()), is_parent_style=True, link=self.current_link)
def add_block_tag(self, tagname, html_tag, tag_style, stylizer, is_table_cell=False, float_spec=None, is_list_item=False):
block = self.blocks.start_new_block(html_tag, tag_style, is_table_cell=is_table_cell, float_spec=float_spec, is_list_item=is_list_item)
anchor = html_tag.get('id') or html_tag.get('name')
if anchor:
block.bookmarks.add(self.bookmark_for_anchor(anchor, html_tag))
if tagname == 'img':
self.images_manager.add_image(html_tag, block, stylizer, as_block=True)
else:
if html_tag.text:
block.add_text(html_tag.text, tag_style, ignore_leading_whitespace=True, is_parent_style=True, link=self.current_link)
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
anchor = html_tag.get('id') or html_tag.get('name') or None
bmark = None
if anchor:
bmark = self.bookmark_for_anchor(anchor, html_tag)
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(html_tag.getparent().iterchildren('*'))[-1]:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(tag_style['clear'], 'none'), bookmark=bmark)
elif tagname == 'img':
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
self.images_manager.add_image(html_tag, block, stylizer, bookmark=bmark)
else:
if html_tag.text:
block = self.blocks.current_or_new_block(html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.text, tag_style, is_parent_style=False, bookmark=bmark, link=self.current_link)
def bookmark_for_anchor(self, anchor, html_tag):
return self.links_manager.bookmark_for_anchor(anchor, self.current_item, html_tag)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts, self.docx.fonts)
self.lists_manager.serialize(self.docx.numbering)
class Convert(object):
# Word does not apply default styling to hyperlinks, so we ensure they get
# default styling (the conversion pipeline does not apply any styling to
# them).
base_css = '''
a[href] { text-decoration: underline; color: blue }
'''
def __init__(self, oeb, docx, mi, add_cover, add_toc):
self.oeb, self.docx, self.add_cover, self.add_toc = oeb, docx, add_cover, add_toc
self.log, self.opts = docx.log, docx.opts
self.mi = mi
self.cover_img = None
p = self.opts.output_profile
p.width_pts, p.height_pts = page_effective_area(self.opts)
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer(base_css=self.base_css)
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager(self.docx.namespace, self.log, self.mi.language)
self.links_manager = LinksManager(self.docx.namespace, self.docx.document_relationships, self.log)
self.images_manager = ImagesManager(self.oeb, self.docx.document_relationships, self.opts)
self.lists_manager = ListsManager(self.docx)
self.fonts_manager = FontsManager(self.docx.namespace, self.oeb, self.opts)
self.blocks = Blocks(self.docx.namespace, self.styles_manager, self.links_manager)
self.current_link = self.current_lang = None
for item in self.oeb.spine:
self.log.debug('Processing', item.href)
self.process_item(item)
if self.add_toc:
self.links_manager.process_toc_links(self.oeb)
if self.add_cover and self.oeb.metadata.cover and unicode(self.oeb.metadata.cover[0]) in self.oeb.manifest.ids:
cover_id = unicode(self.oeb.metadata.cover[0])
item = self.oeb.manifest.ids[cover_id]
self.cover_img = self.images_manager.read_image(item.href)
all_blocks = self.blocks.all_blocks
remove_blocks = []
for i, block in enumerate(all_blocks):
try:
nb = all_blocks[i+1]
except IndexError:
break
block.resolve_skipped(nb)
if block.skipped:
remove_blocks.append((i, block))
for pos, block in reversed(remove_blocks):
self.blocks.delete_block_at(pos)
self.blocks.all_blocks[0].is_first_block = True
self.blocks.apply_page_break_after()
self.blocks.resolve_language()
if self.cover_img is not None:
self.cover_img = self.images_manager.create_cover_markup(self.cover_img, self.opts.preserve_cover_aspect_ratio, *page_size(self.opts))
self.lists_manager.finalize(all_blocks)
self.styles_manager.finalize(all_blocks)
self.write()
def process_item(self, item):
self.current_item = item
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, profile=self.opts.output_profile, base_css=self.base_css)
self.abshref = self.images_manager.abshref = item.abshref
self.current_lang = lang_for_tag(item.data) or self.styles_manager.document_lang
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
self.blocks.top_bookmark = self.links_manager.bookmark_for_anchor(self.links_manager.top_anchor, self.current_item, body)
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self, html_tag, stylizer, is_first_tag=False, float_spec=None):
tagname = barename(html_tag.tag)
tag_style = stylizer.style(html_tag)
ignore_tag_contents = tagname in {'script', 'style', 'title', 'meta'} or tag_style.is_hidden
display = tag_style._get('display')
is_block = False
if not ignore_tag_contents:
previous_link = self.current_link
if tagname == 'a' and html_tag.get('href'):
self.current_link = (self.current_item, html_tag.get('href'), html_tag.get('title'))
previous_lang = self.current_lang
tag_lang = lang_for_tag(html_tag)
if tag_lang:
self.current_lang = tag_lang
is_float = tag_style['float'] in {'left', 'right'} and not is_first_tag
if float_spec is None and is_float:
float_spec = FloatSpec(self.docx.namespace, html_tag, tag_style)
if display in {'inline', 'inline-block'} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
if is_float and float_spec.is_dropcaps:
self.add_block_tag(tagname, html_tag, tag_style, stylizer, float_spec=float_spec)
float_spec = None
else:
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_list_item=True)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.end_current_block()
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and is_float:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
if tagname == 'hr':
for edge in 'right bottom left'.split():
tag_style.set('border-%s-style' % edge, 'none')
self.add_block_tag(tagname, html_tag, tag_style, stylizer, float_spec=float_spec)
for child in html_tag.iterchildren():
if isinstance(getattr(child, 'tag', None), basestring):
self.process_tag(child, stylizer, float_spec=float_spec)
else: # Comment/PI/etc.
tail = getattr(child, 'tail', None)
if tail:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text(tail, tag_style, is_parent_style=False, link=self.current_link, lang=self.current_lang)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
self.current_link = previous_link
self.current_lang = previous_lang
# Now, process the tail if any
if display == 'table-row':
return # We ignore the tail for these tags
ignore_whitespace_tail = is_block or display.startswith('table')
if not is_first_tag and html_tag.tail and (not ignore_whitespace_tail or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text(html_tag.tail, stylizer.style(html_tag.getparent()), is_parent_style=True, link=self.current_link, lang=self.current_lang)
def create_block_from_parent(self, html_tag, stylizer):
parent = html_tag.getparent()
block = self.blocks.current_or_new_block(parent, stylizer.style(parent))
# Do not inherit page-break-before from parent
block.page_break_before = False
return block
def add_block_tag(self, tagname, html_tag, tag_style, stylizer, is_table_cell=False, float_spec=None, is_list_item=False):
block = self.blocks.start_new_block(
html_tag, tag_style, is_table_cell=is_table_cell, float_spec=float_spec, is_list_item=is_list_item)
anchor = html_tag.get('id') or html_tag.get('name')
if anchor:
block.bookmarks.add(self.bookmark_for_anchor(anchor, html_tag))
if tagname == 'img':
self.images_manager.add_image(html_tag, block, stylizer, as_block=True)
else:
if html_tag.text:
block.add_text(html_tag.text, tag_style, ignore_leading_whitespace=True, is_parent_style=True, link=self.current_link, lang=self.current_lang)
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
anchor = html_tag.get('id') or html_tag.get('name') or None
bmark = None
if anchor:
bmark = self.bookmark_for_anchor(anchor, html_tag)
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(html_tag.getparent().iterchildren('*'))[-1]:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(tag_style['clear'], 'none'), bookmark=bmark)
elif tagname == 'img':
block = self.create_block_from_parent(html_tag, stylizer)
self.images_manager.add_image(html_tag, block, stylizer, bookmark=bmark)
else:
if html_tag.text:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text(html_tag.text, tag_style, is_parent_style=False, bookmark=bmark, link=self.current_link, lang=self.current_lang)
elif bmark:
block = self.create_block_from_parent(html_tag, stylizer)
block.add_text('', tag_style, is_parent_style=False, bookmark=bmark, link=self.current_link, lang=self.current_lang)
def bookmark_for_anchor(self, anchor, html_tag):
return self.links_manager.bookmark_for_anchor(anchor, self.current_item, html_tag)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
if self.links_manager.toc:
self.links_manager.serialize_toc(body, self.styles_manager.primary_heading_style)
if self.cover_img is not None:
self.images_manager.write_cover_block(body, self.cover_img)
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts, self.docx.fonts)
self.lists_manager.serialize(self.docx.numbering)
class Convert(object):
# Word does not apply default styling to hyperlinks, so we ensure they get
# default styling (the conversion pipeline does not apply any styling to
# them).
base_css = '''
a[href] { text-decoration: underline; color: blue }
'''
def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer(base_css=self.base_css)
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager(self.docx.namespace)
self.links_manager = LinksManager(self.docx.namespace,
self.docx.document_relationships)
self.images_manager = ImagesManager(self.oeb,
self.docx.document_relationships)
self.lists_manager = ListsManager(self.docx)
self.fonts_manager = FontsManager(self.docx.namespace, self.oeb,
self.opts)
self.blocks = Blocks(self.docx.namespace, self.styles_manager,
self.links_manager)
self.current_link = None
for item in self.oeb.spine:
self.process_item(item)
all_blocks = self.blocks.all_blocks
remove_blocks = []
for i, block in enumerate(all_blocks):
try:
nb = all_blocks[i + 1]
except IndexError:
break
block.resolve_skipped(nb)
if block.skipped:
remove_blocks.append((i, block))
for pos, block in reversed(remove_blocks):
self.blocks.delete_block_at(pos)
self.blocks.all_blocks[0].is_first_block = True
self.lists_manager.finalize(all_blocks)
self.styles_manager.finalize(all_blocks)
self.write()
def process_item(self, item):
self.current_item = item
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data,
item.href,
self.oeb,
self.opts,
self.opts.output_profile,
base_css=self.base_css)
self.abshref = self.images_manager.abshref = item.abshref
for i, body in enumerate(XPath('//h:body')(item.data)):
with self.blocks:
body.set('id',
body.get('id', None) or self.links_manager.top_anchor)
self.process_tag(body, stylizer, is_first_tag=i == 0)
def process_tag(self,
html_tag,
stylizer,
is_first_tag=False,
float_spec=None):
tagname = barename(html_tag.tag)
if tagname in {'script', 'style', 'title', 'meta'}:
return
tag_style = stylizer.style(html_tag)
if tag_style.is_hidden:
return
previous_link = self.current_link
if tagname == 'a' and html_tag.get('href'):
self.current_link = (self.current_item, html_tag.get('href'),
html_tag.get('title'))
display = tag_style._get('display')
is_float = tag_style['float'] in {'left', 'right'} and not is_first_tag
if float_spec is None and is_float:
float_spec = FloatSpec(self.docx.namespace, html_tag, tag_style)
if display in {
'inline', 'inline-block'
} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
if is_float and float_spec.is_dropcaps:
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
float_spec=float_spec)
float_spec = None
else:
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
is_list_item=True)
elif display.startswith('table') or display == 'inline-table':
if display == 'table-cell':
self.blocks.start_new_cell(html_tag, tag_style)
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
is_table_cell=True)
elif display == 'table-row':
self.blocks.start_new_row(html_tag, tag_style)
elif display in {'table', 'inline-table'}:
self.blocks.end_current_block()
self.blocks.start_new_table(html_tag, tag_style)
else:
if tagname == 'img' and is_float:
# Image is floating so dont start a new paragraph for it
self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
else:
self.add_block_tag(tagname,
html_tag,
tag_style,
stylizer,
float_spec=float_spec)
for child in html_tag.iterchildren('*'):
self.process_tag(child, stylizer, float_spec=float_spec)
is_block = html_tag in self.blocks.open_html_blocks
self.blocks.finish_tag(html_tag)
if is_block and tag_style['page-break-after'] == 'avoid':
self.blocks.all_blocks[-1].keep_next = True
self.current_link = previous_link
if display == 'table-row':
return # We ignore the tail for these tags
ignore_whitespace_tail = is_block or display.startswith('table')
if not is_first_tag and html_tag.tail and (
not ignore_whitespace_tail or not html_tag.tail.isspace()):
# Ignore trailing space after a block tag, as otherwise it will
# become a new empty paragraph
block = self.blocks.current_or_new_block(
html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.tail,
stylizer.style(html_tag.getparent()),
is_parent_style=True,
link=self.current_link)
def add_block_tag(self,
tagname,
html_tag,
tag_style,
stylizer,
is_table_cell=False,
float_spec=None,
is_list_item=False):
block = self.blocks.start_new_block(html_tag,
tag_style,
is_table_cell=is_table_cell,
float_spec=float_spec,
is_list_item=is_list_item)
anchor = html_tag.get('id') or html_tag.get('name')
if anchor:
block.bookmarks.add(self.bookmark_for_anchor(anchor, html_tag))
if tagname == 'img':
self.images_manager.add_image(html_tag,
block,
stylizer,
as_block=True)
else:
if html_tag.text:
block.add_text(html_tag.text,
tag_style,
ignore_leading_whitespace=True,
is_parent_style=True,
link=self.current_link)
def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
anchor = html_tag.get('id') or html_tag.get('name') or None
bmark = None
if anchor:
bmark = self.bookmark_for_anchor(anchor, html_tag)
if tagname == 'br':
if html_tag.tail or html_tag is not tuple(
html_tag.getparent().iterchildren('*'))[-1]:
block = self.blocks.current_or_new_block(
html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_break(clear={
'both': 'all',
'left': 'left',
'right': 'right'
}.get(tag_style['clear'], 'none'),
bookmark=bmark)
elif tagname == 'img':
block = self.blocks.current_or_new_block(
html_tag.getparent(), stylizer.style(html_tag.getparent()))
self.images_manager.add_image(html_tag,
block,
stylizer,
bookmark=bmark)
else:
if html_tag.text:
block = self.blocks.current_or_new_block(
html_tag.getparent(), stylizer.style(html_tag.getparent()))
block.add_text(html_tag.text,
tag_style,
is_parent_style=False,
bookmark=bmark,
link=self.current_link)
def bookmark_for_anchor(self, anchor, html_tag):
return self.links_manager.bookmark_for_anchor(anchor,
self.current_item,
html_tag)
def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
self.blocks.serialize(body)
body.append(body[0]) # Move <sectPr> to the end
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles,
self.docx.font_table,
self.docx.embedded_fonts, self.docx.fonts)
self.lists_manager.serialize(self.docx.numbering)
class Convert(object):
def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts
self.blocks = []
def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
self.svg_rasterizer = SVGRasterizer()
self.svg_rasterizer(self.oeb, self.opts)
self.styles_manager = StylesManager()
self.images_manager = ImagesManager(self.oeb, self.docx.document_relationships)
self.fonts_manager = FontsManager(self.oeb)
for item in self.oeb.spine:
self.process_item(item)
self.styles_manager.finalize(self.blocks)
self.write()
def process_item(self, item):
stylizer = self.svg_rasterizer.stylizer_cache.get(item)
if stylizer is None:
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile)
self.abshref = self.images_manager.abshref = item.abshref
is_first_block = True
for body in XPath('//h:body')(item.data):
b = Block(self.styles_manager, body, stylizer.style(body), is_first_block=is_first_block)
self.blocks.append(b)
is_first_block = False
self.process_block(body, b, stylizer, ignore_tail=True)
if self.blocks and self.blocks[0].is_empty():
del self.blocks[0]
def process_block(self, html_block, docx_block, stylizer, ignore_tail=False):
block_style = stylizer.style(html_block)
if block_style.is_hidden:
return
if html_block.tag.endswith('}img'):
self.images_manager.add_image(html_block, docx_block, stylizer)
else:
if html_block.text:
docx_block.add_text(html_block.text, block_style, ignore_leading_whitespace=True, is_parent_style=True)
for child in html_block.iterchildren(etree.Element):
tag = barename(child.tag)
style = stylizer.style(child)
display = style._get('display')
if display == 'block' and tag != 'br':
if tag == 'img' and style['float'] in {'left', 'right'}:
# Image is floating so dont start a new paragraph for
# it
self.process_inline(child, self.blocks[-1], stylizer)
else:
b = Block(self.styles_manager, child, style)
self.blocks.append(b)
self.process_block(child, b, stylizer)
else:
self.process_inline(child, self.blocks[-1], stylizer)
if block_style['page-break-after'] == 'avoid':
self.blocks[-1].keep_next = True
if ignore_tail is False and html_block.tail and html_block.tail.strip():
style = stylizer.style(html_block.getparent())
b = Block(self.styles_manager, html_block.getparent(), style)
self.blocks.append(b)
b.add_text(html_block.tail, style, is_parent_style=True)
def process_inline(self, html_child, docx_block, stylizer):
tag = barename(html_child.tag)
style = stylizer.style(html_child)
if style.is_hidden:
return
if tag == 'br':
if html_child.tail or html_child is not html_child.getparent()[-1]:
docx_block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(style['clear'], 'none'))
elif tag == 'img':
self.images_manager.add_image(html_child, docx_block, stylizer)
else:
if html_child.text:
docx_block.add_text(html_child.text, style, html_parent=html_child)
for child in html_child.iterchildren(etree.Element):
style = stylizer.style(child)
display = style.get('display', 'inline')
if display == 'block':
b = Block(self.styles_manager, child, style)
self.blocks.append(b)
self.process_block(child, b, stylizer)
else:
self.process_inline(child, self.blocks[-1], stylizer)
if html_child.tail:
self.blocks[-1].add_text(html_child.tail, stylizer.style(html_child.getparent()), html_parent=html_child.getparent(), is_parent_style=True)
def write(self):
dn = {k:v for k, v in namespaces.iteritems() if k in {'w', 'r', 'm', 've', 'o', 'wp', 'w10', 'wne', 'a', 'pic'}}
E = ElementMaker(namespace=dn['w'], nsmap=dn)
self.docx.document = doc = E.document()
body = E.body()
doc.append(body)
for block in self.blocks:
block.serialize(body)
width, height = PAPER_SIZES[self.opts.docx_page_size]
if self.opts.docx_custom_page_size is not None:
width, height = map(float, self.opts.docx_custom_page_size.partition('x')[0::2])
width, height = int(20 * width), int(20 * height)
def margin(which):
return w(which), str(int(getattr(self.opts, 'margin_'+which) * 20))
body.append(E.sectPr(
E.pgSz(**{w('w'):str(width), w('h'):str(height)}),
E.pgMar(**dict(map(margin, 'left top right bottom'.split()))),
E.cols(**{w('space'):'720'}),
E.docGrid(**{w('linePitch'):"360"}),
))
dn = {k:v for k, v in namespaces.iteritems() if k in tuple('wra') + ('wp',)}
E = ElementMaker(namespace=dn['w'], nsmap=dn)
self.docx.styles = E.styles(
E.docDefaults(
E.rPrDefault(
E.rPr(
E.rFonts(**{w('asciiTheme'):"minorHAnsi", w('eastAsiaTheme'):"minorEastAsia", w('hAnsiTheme'):"minorHAnsi", w('cstheme'):"minorBidi"}),
E.sz(**{w('val'):'22'}),
E.szCs(**{w('val'):'22'}),
E.lang(**{w('val'):'en-US', w('eastAsia'):"en-US", w('bidi'):"ar-SA"})
)
),
E.pPrDefault(
E.pPr(
E.spacing(**{w('after'):"0", w('line'):"276", w('lineRule'):"auto"})
)
)
)
)
self.docx.images = {}
self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images)
self.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts)