Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Refactor HTML tag processing in preparation for tables and lists

Browse Source
This commit is contained in:
Kovid Goyal 2015-03-31 16:29:57 +05:30
parent a2319a1338
commit 6d03706aa3
1 changed files with 95 additions and 70 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

165
src/calibre/ebooks/docx/writer/from_html.py
View File

@ -8,8 +8,6 @@ __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
import re import re
from lxml import etree
from calibre.ebooks.docx.writer.container import create_skeleton from calibre.ebooks.docx.writer.container import create_skeleton
from calibre.ebooks.docx.writer.styles import w, StylesManager from calibre.ebooks.docx.writer.styles import w, StylesManager
from calibre.ebooks.docx.writer.images import ImagesManager from calibre.ebooks.docx.writer.images import ImagesManager
@ -154,13 +152,42 @@ class Block(object):
return False return False
return True return True
class Blocks(object):
def __init__(self):
self.all_blocks = []
self.pos = 0
self.current_block = None
def start_new_block(self, styles_manager, html_block, style, is_first_tag=False):
if self.current_block is not None:
self.all_blocks.append(self.current_block)
self.current_block = Block(styles_manager, html_block, style, is_first_block=is_first_tag)
return self.current_block
def serialize(self, body):
for block in self.all_blocks:
block.serialize(body)
def __enter__(self):
self.pos = len(self.all_blocks)
def __exit__(self, *args):
if self.current_block is not None:
self.all_blocks.append(self.current_block)
self.current_block = None
if len(self.all_blocks) > self.pos and self.all_blocks[self.pos].is_empty():
# Delete the empty block corresponding to the <body> tag when the
# body tag has no text content before its first sub-block
del self.all_blocks[self.pos]
class Convert(object): class Convert(object):
def __init__(self, oeb, docx): def __init__(self, oeb, docx):
self.oeb, self.docx = oeb, docx self.oeb, self.docx = oeb, docx
self.log, self.opts = docx.log, docx.opts self.log, self.opts = docx.log, docx.opts
self.blocks = [] self.blocks = Blocks()
def __call__(self): def __call__(self):
from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
@ -174,7 +201,7 @@ class Convert(object):
for item in self.oeb.spine: for item in self.oeb.spine:
self.process_item(item) self.process_item(item)
self.styles_manager.finalize(self.blocks) self.styles_manager.finalize(self.blocks.all_blocks)
self.write() self.write()
def process_item(self, item): def process_item(self, item):
@ -183,80 +210,78 @@ class Convert(object):
stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile) stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.opts.output_profile)
self.abshref = self.images_manager.abshref = item.abshref self.abshref = self.images_manager.abshref = item.abshref
is_first_block = True for i, body in enumerate(XPath('//h:body')(item.data)):
for body in XPath('//h:body')(item.data): with self.blocks:
b = Block(self.styles_manager, body, stylizer.style(body), is_first_block=is_first_block) self.process_tag(body, stylizer, is_first_tag=i == 0)
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): def process_tag(self, html_tag, stylizer, is_first_tag=False):
block_style = stylizer.style(html_block) tagname = barename(html_tag.tag)
if block_style.is_hidden: if tagname in {'script', 'style', 'title', 'meta'}:
return return
if html_block.tag.endswith('}img'): tag_style = stylizer.style(html_tag)
self.images_manager.add_image(html_block, docx_block, stylizer) if tag_style.is_hidden:
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 return
if tag == 'br': display = tag_style._get('display')
if html_child.tail or html_child is not html_child.getparent()[-1]: inlined = True
docx_block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(style['clear'], 'none')) if display in {'inline', 'inline-block'} or tagname == 'br': # <br> has display:block but we dont want to start a new paragraph
elif tag == 'img': if self.blocks.current_block is not None:
self.images_manager.add_image(html_child, docx_block, stylizer) self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
elif display == 'list-item':
# TODO: Implement this
inlined = False
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
elif display.startswith('table') or display == 'inline-table':
inlined = False
# TODO: implement this
if display == 'table-cell':
self.add_block_tag(tagname, html_tag, tag_style, stylizer)
else: else:
if html_child.text: if tagname == 'img' and tag_style['float'] in {'left', 'right'}:
docx_block.add_text(html_child.text, style, html_parent=html_child) # Image is floating so dont start a new paragraph for it
for child in html_child.iterchildren(etree.Element): self.add_inline_tag(tagname, html_tag, tag_style, stylizer)
style = stylizer.style(child) else:
display = style.get('display', 'inline') self.add_block_tag(tagname, html_tag, tag_style, stylizer, is_first_tag=is_first_tag)
if display == 'block': inlined = False
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: for child in html_tag.iterchildren('*'):
self.blocks[-1].add_text(html_child.tail, stylizer.style(html_child.getparent()), html_parent=html_child.getparent(), is_parent_style=True) self.process_tag(child, stylizer)
if not is_first_tag and html_tag.tail:
if inlined:
self.add_text_to_current_block(html_tag.tail, stylizer.style(html_tag.getparent()))
elif html_tag.tail.strip():
self.blocks.start_new_block(self.styles_manager, html_tag.getparent(), stylizer.style(html_tag.getparent()))
self.add_text_to_current_block(html_tag.tail, stylizer.style(html_tag.getparent()))
def add_text_to_current_block(self, text, tag_style, ignore_leading_whitespace=False, html_parent=None, is_parent_style=False):
block = self.blocks.current_block
if block is not None:
block.add_text(text, tag_style, ignore_leading_whitespace=ignore_leading_whitespace, html_parent=html_parent, is_parent_style=is_parent_style)
def add_block_tag(self, tagname, html_tag, tag_style, stylizer, is_first_tag=False):
block = self.blocks.start_new_block(self.styles_manager, html_tag, tag_style, is_first_tag=is_first_tag)
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)
if tag_style['page-break-after'] == 'avoid':
block.keep_next = 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]:
self.blocks.current_block.add_break(clear={'both':'all', 'left':'left', 'right':'right'}.get(tag_style['clear'], 'none'))
elif tagname == 'img':
self.images_manager.add_image(html_tag, self.blocks.current_block, stylizer)
else:
if html_tag.text:
self.add_text_to_current_block(html_tag.text, tag_style, html_parent=html_tag)
def write(self): def write(self):
self.docx.document, self.docx.styles, body = create_skeleton(self.opts) self.docx.document, self.docx.styles, body = create_skeleton(self.opts)
for block in self.blocks: self.blocks.serialize(body)
block.serialize(body) body.append(body[0]) # Move <sectPr> to the end
self.styles_manager.serialize(self.docx.styles) self.styles_manager.serialize(self.docx.styles)
self.images_manager.serialize(self.docx.images) 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.fonts_manager.serialize(self.styles_manager.text_styles, self.docx.font_table, self.docx.embedded_fonts, self.docx.fonts)