我的第一个想法是做一个否定匹配:
body.xpath( './node()[not(self::b)]' )
可悲的是,这将排除<b>
而不是忽略它:
> body = Nokogiri::HTML::DocumentFragment.parse %(<h3><b>foo</b></h3><h3>bar</h3>fooz<b>baz</b>whatever); true
> body.xpath( './node()[not(self::b)]' ).to_a
[
[0] <h3>
<b>foo</b>
</h3>,
[1] <h3>bar</h3>,
[2] fooz,
[3] whatever
]
因此,您别无选择,只能在这里使用缓冲区:我们可以先遍历节点,以填充有关是否应该有新行的缓冲区,然后迭代此缓冲区以将您的行添加到 pdf :
buffer = []
body.xpath( './node()' ).each do |node|
if %w[text b].include? node.name
# add to previous line or create one
buffer << [] unless buffer.count
buffer.last << { node: node }
else
# set content and create a new line
buffer << [ { node: node, title: node.name == 'h3' } ]
buffer << []
end
end
# Now, each first level item in buffer is a line,
# containing elements we just have to concatenate text of
# to pass to `pdf#text`
buffer.each do |line|
text = line.map do |part|
node = part[ :node ]
inner = node.text.to_s
# restore <b> tag if you want bold style in pdf
node.name == 'b' ? "<b>#{inner}</b>" : inner
end.join
if line.first
if line.first[ :title ]
pdf.fill_color colour
pdf.text text, :size => 16
pdf.move_down 5
else
pdf.fill_color '#444444'
# inline_format ensure basic html formating is used, <b> in our case
# See http://prawn.majesticseacreature.com/docs/0.11.1/Prawn/Text.html#method-i-text
pdf.text text, size: 10, leading: 1, inline_format: true
pdf.move_down 10
end
end
end
当然,这一切都是考虑到你不控制原始 html。否则,你应该把你的文本节点放在里面<p>
什么的,就不会再有问题了。