假设您有表示以下两个文档中的<a>
元素的Nokogiri节点:
<r xmlns:x="foo"><a foo="bar" jim="jam" x:oh="no"><x:b>Hello</x:b></a></r>
<r xmlns:i="foo"><a jim="jam" i:oh="no" foo="bar"><i:b>Hello</i:b></a></r>
从DOM的角度来看,两者是等价的。我想有效地检测这个,但是Nokogiri::XML::Node#==
只是检查对象相等。由于Nokogiri 1.5.0还不支持规范化,所以我不能只序列化节点并比较字符串。
比较两个节点以确保它们的名称、属性和内容在标准意义上相等的最快方法是什么?
答案可能依赖于仅在Ruby 1.9.2+中可用的特性,如果需要的话。
测试用例ORIG1 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
ORIG2 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b jim='jam' foo='bar'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NOTEXT = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c/></a3>
</a>"
EXTRATEXT1 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foobar</a3c></a3>
</a>"
EXTRATEXT2 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b>hi</a3b><a3c>foo</a3c></a3>
</a>"
MISSINGNODE = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/></a3>
</a>"
EXTRANODE = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c><a3d/></a3>
</a>"
SWAPNODE = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3x/><a3b/><a3c>foo</a3c></a3>
</a>"
MISSINGATTRIB = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
EXTRAATTRIB1 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam' kits='meow'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
EXTRAATTRIB2 = "<a>
<a1><a1a/><a1b/><a1c kits='meow'/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
SWAPATTRIB1 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='zzz'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
SWAPATTRIB2 = "<a>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' zzz='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE1 = "<a xmlns:x='foo'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><x:a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE1B = "<a xmlns:z='foo'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><z:a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE1C = "<a xmlns:x='bar'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' jim='jam'/><a2c/></a2>
<a3><x:a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE2 = "<a xmlns:x='foo'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' x:jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE2B= "<a xmlns:z='foo'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' z:jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
NAMESPACE2C= "<a xmlns:x='bar'>
<a1><a1a/><a1b/><a1c/></a1>
<a2><a2a/><a2b foo='bar' x:jim='jam'/><a2c/></a2>
<a3><a3a/><a3b/><a3c>foo</a3c></a3>
</a>"
require 'nokogiri'
require 'minitest/autorun'
class NodeEquivalence < MiniTest::Unit::TestCase
def setup
@o1 = Nokogiri::XML(ORIG1,&:noblanks).root
end
def test_equivalence
o2 = Nokogiri::XML(ORIG2,&:noblanks).root
assert @o1 =~ o2, "Equivalent nodes should be equivalent"
assert o2 =~ @o1, "Equivalent nodes should be equivalent"
end
def test_textnodes
no_text = Nokogiri::XML(NOTEXT,&:noblanks).root
extra1 = Nokogiri::XML(EXTRATEXT1,&:noblanks).root
extra2 = Nokogiri::XML(EXTRATEXT2,&:noblanks).root
refute @o1 =~ no_text, "Notice missing text node child"
refute no_text =~ @o1, "Notice missing text node child"
refute @o1 =~ extra1, "Notice different text in text node"
refute extra1 =~ @o1, "Notice different text in text node"
refute @o1 =~ extra2, "Notice extra text node"
refute extra2 =~ @o1, "Notice extra text node"
end
def test_nodes
missing = Nokogiri::XML(MISSINGNODE,&:noblanks).root
extra = Nokogiri::XML(EXTRANODE,&:noblanks).root
changed = Nokogiri::XML(SWAPNODE,&:noblanks).root
refute @o1 =~ missing, "Notice missing node"
refute missing =~ @o1, "Notice missing node"
refute @o1 =~ extra, "Notice extra node"
refute extra =~ @o1, "Notice extra node"
refute @o1 =~ changed, "Notice renamed node"
refute changed =~ @o1, "Notice renamed node"
end
def test_attributes
missing = Nokogiri::XML(MISSINGATTRIB,&:noblanks).root
extra1 = Nokogiri::XML(EXTRAATTRIB1,&:noblanks).root
extra2 = Nokogiri::XML(EXTRAATTRIB2,&:noblanks).root
swap1 = Nokogiri::XML(SWAPATTRIB1,&:noblanks).root
swap2 = Nokogiri::XML(SWAPATTRIB2,&:noblanks).root
refute @o1 =~ missing, "Notice missing attribute"
refute missing =~ @o1, "Notice missing attribute"
refute @o1 =~ extra1, "Notice extra attribute"
refute extra1 =~ @o1, "Notice extra attribute"
refute @o1 =~ extra2, "Notice new attribute"
refute extra2 =~ @o1, "Notice new attribute"
refute @o1 =~ swap1, "Notice changed attribute value"
refute swap1 =~ @o1, "Notice changed attribute value"
refute @o1 =~ swap2, "Notice changed attribute name"
refute swap2 =~ @o1, "Notice changed attribute name"
end
def test_namespaces
ns1 = Nokogiri::XML(NAMESPACE1,&:noblanks).root
ns2 = Nokogiri::XML(NAMESPACE2,&:noblanks).root
ns1b = Nokogiri::XML(NAMESPACE1B,&:noblanks).root
ns2b = Nokogiri::XML(NAMESPACE2B,&:noblanks).root
ns1c = Nokogiri::XML(NAMESPACE1C,&:noblanks).root
ns2c = Nokogiri::XML(NAMESPACE2C,&:noblanks).root
refute @o1 =~ ns1, "Notice added node namespace"
refute ns1 =~ @o1, "Notice removed node namespace"
refute @o1 =~ ns2, "Notice added attribute namespace"
refute ns2 =~ @o1, "Notice removed attribute namespace"
assert ns1 =~ ns1b, "Different namespace names on nodes don't matter"
assert ns2 =~ ns2b, "Different namespace names on attributes don't matter"
refute ns1 =~ ns1c, "Notice different namespace hrefs on nodes"
refute ns2 =~ ns2c, "Notice different namespace hrefs on attributes"
end
end
这是我目前的实现。不是现在命名空间感知:
class Nokogiri::XML::Node
# Return true if this node is content-equivalent to other, false otherwise
def =~(other)
return true if self == other
return false unless name == other.name
stype = node_type; otype = other.node_type
return false unless stype == otype
sa = attributes; oa = other.attributes
return false unless sa.length == oa.length
sa = sa.sort.map{ |n,a| [n,a.value,a.namespace && a.namespace.href] }
oa = oa.sort.map{ |n,a| [n,a.value,a.namespace && a.namespace.href] }
return false unless sa == oa
skids = children; okids = other.children
return false unless skids.length == okids.length
return false if stype == TEXT_NODE && (content != other.content)
sns = namespace; ons = other.namespace
return false if !sns ^ !ons
return false if sns && (sns.href != ons.href)
skids.to_enum.with_index.all?{ |ski,i| ski =~ okids[i] }
end
end
下面是我的基准代码(使用上面测试用例中的常量):
require 'benchmark'
Benchmark.bm(10) do |x|
N = 1000
NODES = [
ORIG1, ORIG2, NOTEXT, EXTRATEXT1, EXTRATEXT2,
MISSINGNODE, EXTRANODE, SWAPNODE,
MISSINGATTRIB, EXTRAATTRIB1, EXTRAATTRIB2, SWAPATTRIB1, SWAPATTRIB2,
NAMESPACE1, NAMESPACE2
].map{ |xml| Nokogiri::XML(xml,&:noblanks).root }
MAIN = NODES.shift
x.report("Phrogz"){ N.times{
NODES.each{ |other| MAIN =~ other }
}}
end