简介
当听到样式表这个词时,您可能会想到 CSS 样式表。XSLT 样式表通常用于 XML 转换,比如在 Web 服务之间映射数据。因为 XSLT 非常适合此用途,所以创建了顶层元素 <stylesheet> 的 <xsl:transform> 别名,虽然这很少使用。这种 XSLT 转换的输入结构与输出结构有很大的不同。最重要的是,命名空间的不同。
XSLT 样式表的输入结构与输出结构相似,但却更简单些。其中已经扩充了一些标记,但大部分标记只是原样复制到输出。输入和输出的命名空间是相同的 (HTML)。输入文档也可以包含样式表指令(比如创建脚注),这些指令属于另一个命名空间,不会传递到输出中。
常用缩略语
- CSS:级联样式表
- XHTML:可扩展超文本标记语言
- XPath:XML 路径语言
- XSLT:可扩展样式表语言转换
在本文中,我们将学习如何使用 XSLT 样式表扩充 XHTML 文档。文中的示例展示了如何使用指令,如何引用其他源文档的部分,以及如何使用链接在主文档中导航。此外,我们还探索了页面的解释和编译之间的区别。
CSS 样式表的局限性
XSLT 样式表不会阻止您使用其他技术,比如 JavaScript 或 CSS。CSS 适用于字体、加粗、颜色、间距等。它不适合将来自不同位置的信息组合在一起,比如脚注、模块或生成一个目录。这正是 XSLT 的用武之地,它补充而不是替代了 CSS。
XSLT 用途示例
实际上,您可以将 XSLT 代码集中在一个文件中。为了简单起见,本文中的每个示例均位于一个独立的 XSLT 文件中,除了一些必要的代码。清单 1 给出了必需的代码。
清单 1. 必需的代码(位于 samples/common.xml 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
|
<
s:stylesheet
exclude-result-prefixes
=
"a h"
version
=
"1.0"
>
<
s:template
match
=
"h:head"
>
<
s:copy
>
<
s:apply-templates
select
=
"@*|node()"
/>
<
meta
http-equiv
=
"content-type"
content
=
"text/html;charset=UTF-8"
/>
<
link
href
=
"common.css"
rel
=
"stylesheet"
type
=
"text/css"
/>
</
s:copy
>
</
s:template
>
<
s:template
match
=
"*"
>
<
s:copy
>
<
s:copy-of
select
=
"@*"
/>
<
s:apply-templates
/>
</
s:copy
>
</
s:template
>
</
s:stylesheet
>
|
XHTML 的命名空间定义了两次:默认定义和 h:。默认命名空间用于编写输出 XHTML 标记,其中应该避免使用命名空间前缀。h: 用在 XPath 表达式中。
本文使用 XSLT 1.0 版本。目前,大部分浏览器都无法解释 XSLT 2.0。但是,如果 XSLT 运行在服务器上,那么它可能是一个实用的选择。XSLT 2.0 还提供了:
- XPATH 2.0(if…then…else 和许多内置的函数)
- 内置和用户编写的 XPATH 函数
- 分组
在 清单 1 中:
- s:template match="head" 扩充了源文档的 head 一节,添加了一个 CSS 样式表的链接。即使 UTF-8 是在 XML 中的默认编码,一些浏览器也需要内容类型才能呈现它。
- s:template match="*" 是默认的详细副本。原则上,所有内容都会复制到目标文档中。如果遗漏了此模板,只会将标记的文本内容复制到目标文档。不会复制处理指令节点。
本文中的所有其他示例都是导入 common.xsl 的独立文件。
扩充
通过扩充,添加了一个未在源文档中显式请求的特性。一个示例是 清单 1 中的 CSS 样式表的链接。尝试另一个示例,向每个内部链接添加一个小箭头 (^ v),指明目标在它之前还是之后。清单 2 给出了该样式表。
清单 2. 样式表(在 samples/linkUpDown.xsl 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
|
<
s:stylesheet
version
=
"1.0"
>
<
s:import
href
=
"common.xsl"
/>
<
s:template
match
=
"h:a[starts-with(@href,'#')]"
>
<
s:copy
>
<
s:copy-of
select
=
"@*"
/>
<
s:variable
name
=
"name"
select
=
"substring-after(@href,'#')"
/>
<
s:choose
>
<
s:when
test
=
"preceding::h:a[@name=$name]"
>
<
s:text
>^</
s:text
>
</
s:when
>
<
s:when
test
=
"following::h:a[@name=$name]"
>
<
s:text
>v</
s:text
>
</
s:when
>
</
s:choose
>
<
s:apply-templates
/>
</
s:copy
>
</
s:template
>
</
s:stylesheet
>
|
首先,导入清单 2 中的通用样式表。模板与内部链接(以 '#' 开头)相匹配。如果链接指向的锚点位于链接之前,那么使用一个向上箭头扩充该链接(如果情况相反,则使用向下箭头)。
s:copy-of 和 s:apply-templates 可确保不会沿途丢下任何内容。
清单 3 给出了一个示例文档(其中包含内部链接),它经过了清单 2 中的样式表进行扩充。
清单 3. 源文档(在 samples/linkUpDown.xml 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
|
<?
xml-stylesheet
href
=
"linkUpDown.xsl"
type
=
"text/xsl"
?>
<
head
/>
<
body
>
<
a
name
=
"a"
/>
<
p
>This link goes <
a
href
=
"vb"
>downward.</
a
></
p
>
<
br
/>
<
p
>Reduce the size of the window to verify the link really works.</
p
>
<
br
/>
<
a
name
=
"b"
/>
<
p
>This link goes <
a
href
=
"^a"
>upward.</
a
>
</
p
>
</
body
>
</
html
>
|
目标文档看起来相同,除了清单 4 中的条目。
清单 4. 目标文档(在 samples/linkUpDown.html 中)
1
2
|
… <
a
href
=
"#b"
>v downwards.</
a
> …
… <
a
href
=
"#a"
>^ upwards.</
a
> …
|
指令
您可以在源文档中添加一些指令,告诉样式表执行何种操作。它们属于另一个命名空间(在本例中为前缀 a:),不会被复制到目标文档。
在清单 5 中,源文档中任何地方的指令标记 a:ref 都会创建一个脚注。
清单 5. 样式表(在 samples/footnote.xsl 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
|
<
s:stylesheet
version
=
"1.0"
>
<
s:import
href
=
"common.xsl"
/>
<
s:template
match
=
"h:body"
>
<
s:copy
>
<
s:apply-templates
select
=
"@*|node()"
/
<!-- put the footnotes at the end
if there is no a:references directive -->
<
s:if
test
=
"not(descendant::a:references)"
>
<
s:call-template
name
=
"references"
/>
</
s:if
>
</
s:copy
>
</
s:template
>
<!-- Create a footnote -->
<
s:template
match
=
"a:ref"
>
<
s:variable
name
=
"number"
select
=
"count(preceding::a:ref) + 1"
/>
<
a
name
=
"ref-{$number}"
></
a
>
<
a
class
=
"footnote"
href
=
"#reference-{$number}"
>
<
s:value-of
select
=
"concat('v ',$number)"
/>
</
a
>
</
s:template
>
<!-- if a:reference is missing, assume it at the end of the body -->
<
s:template
match
=
"a:references"
name
=
"references"
>
<
hr
/>
<
s:for-each
select
=
"//a:ref"
>
<
s:variable
name
=
"number"
select
=
"count(preceding::a:ref) + 1"
/>
<
p
>
<
a
name
=
"reference-{$number}"
></
a
>
<
a
class
=
"footnote"
href
=
"#ref-{$number}"
>
<
s:value-of
select
=
"concat(' ^',$number)"
/>
</
a
>
<
s:apply-templates
/>
</
p
>
</
s:for-each
>
</
s:template
>
</
s:stylesheet
>
|
使用源文档中的 a:references 指令,名为 references 的模板会在模板与该指令匹配的地方分配脚注。如果缺少这样一个指令,第一个与 body 匹配的模板会在 body 的末尾分配脚注,方法是调用名为 references 的相同模板。在两种情况下,都会列出脚注的内容,并生成一个由向上箭头表示的向上链接。
第二个模板(匹配 a:ref)使用向下箭头创建脚注的链接。脚注具有编号。这里忽略了它的内容。
class="footnote" 属性在 XSLT 转换之后由一个 CSS 样式表解析,该样式表链接在 XSLT 样式表 common.xsl 中。
清单 6 中的源文档使用 a:ref 指令创建脚注。
清单 6. 源文档(在 samples/footnote.xml 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
|
<?
xml-stylesheet
href
=
"footnote.xsl"
type
=
"text/xsl"
?>
<
html
>
<
head
/>
<
body
>
<
p
>
This example looks a little scientific
<
a:ref
>
From Latin
<
em
>scientia</
em
>
</
a:ref
>
and academic
<
a:ref
>From Greek akademia</
a:ref
>.
</
p
>
<
p
>
Do you know why?
<
a:ref
>
It uses
<
em
>footnotes</
em
>.
</
a:ref
>
</
p
>
<
p
>Reduce size of window to verify links are generated.</
p
>
</
body
>
</
html
>
|
目标文档将脚注列表包含在底部,如清单 7 所示。
清单 7. 目标文档(在 samples/footnote.html 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
|
<
html
<
head
><
link
type
=
"text/css"
rel
=
"stylesheet"
href
=
"common.css"
/></
head
>
<
body
>
<
p
>This example looks a little scientific
<
a
name
=
"ref-1"
/><
a
href
=
"#reference-1"
class
=
"footnote"
>v 1</
a
>
and academic.
<
a
name
=
"ref-2"
/><
a
href
=
"#reference-2"
class
=
"footnote"
>v 2lt;/a>
</
p
>
<
p
>Do you know why?
<
a
name
=
"ref-3"
/><
a
href
=
"#reference-3"
class
=
"footnote"
>v 3</
a
>
</
p
>
<
p
>Reduce size of window to verify links are generated.</
p
>
br/><
br
/>
<
hr
/>
<
p
><
a
name
=
"reference-1"
/><
a
href
=
"#ref-1"
class
=
"footnote"
> ^1</
a
>
From Latin
<
em
>scientia</
em
>
</
p
>
<
p
><
a
name
=
"reference-2"
/>
<
a
href
=
"#ref-2"
class
=
"footnote"
> ^2</
a
>From Greek akademia</
p
>
<
p
><
a
name
=
"reference-3"
/><
a
href
=
"#ref-3"
class
=
"footnote"
> ^3</
a
>
It uses
<
em
>footnotes</
em
>.
</
p
>
</
body
>
</
html
>
|
突破源文档的边界
也可引用其他源文档的其中一些部分。a:include 指令包含一个可能属于另一个源文档的元素并转换它,如清单 8 所示。
清单 8. 样式表(在 samples/include.xsl 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
|
<
s:stylesheet
version
=
"1.0"
>
<
s:import
href
=
"common.xsl"
/>
<
s:template
match
=
"a:include"
>
<
s:choose
>
<
s:when
test
=
"0!=string-length(@src)"
>
<
s:apply-templates
select
=
"document(@src)//*[@id=current()/@refid]"
/>
</
s:when
>
<
s:when
test
=
"not(@src) and //a:default[1]/@src"
>
<
s:apply-templates
select
=
"document(//a:default[1]/@src)//*[@id=current()/@refid]"
/>
</
s:when
>
<
s:when
test
=
"0=string-length(@src) or not(//a:default[1]/@src)"
>
<
s:apply-templates
select
=
"//*[@id=current()/@refid]"
/>
</
s:when
>
</
s:choose
>
</
s:template
>
</
s:stylesheet
>
|
源文档中的一个 a:include 指令引用源元素的 id。包含该元素的文档可在一个 src 属性中命名。如果缺少该属性,将使用 a:default 指令的 src 属性。如果在任何地方都没有 src 属性,则使用同一个源文档。因此,refid 会引用 id 来避免无限的递归。
导入的元素可能具有一种复杂的类型,并在包含 (apply-templates)之后进行转换。清单 9、清单 10 和清单 11 给出了示例。
清单 9. 源文档(在 samples/include.xml 中)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
|
<?
xml-stylesheet
href
=
"include.xsl"
type
=
"text/xsl"
?>
<
html
<
head
>
<
a:default
src
=
"includedY.xml"
/>
</
head
>
<
body
>
<
p
>The following text is included:</
p
>
<
a:include
refid
=
"x"
src
=
"includedX.xml"
/>
<
a:include
refid
=
"y1"
/>
<
p
id
=
"i"
>double</
p
>
<
a:include
refid
=
"y2"
/>
<
a:include
refid
=
"i"
src
=
""
/>
</
body
>
</
html
>
|
清单 10. 源文档的部分(在 samples/includeY.xml 中)
1
2
|
<
h2
id
=
"y2"
>I'm the <
em
>included</
em
> h2</
h2
>
<
h1
id
=
"y1"
>I'm the <
em
>included</
em
> h1</
h1
>
|
清单 11. 目标文档(在 samples/include.html 中)
1
2
3
4
5
6
7
8
9
|
<
body
>
<
p
>The following text is included:</
p
>
<
p
id
=
"x"
>I'm the <
em
>included</
em
> paragraph.</
p
>
<
h1
id
=
"y1"
>I'm the <
em
>included</
em
> h1</
h1
>
<
p
id
=
"i"
>double</
p
>
<
h2
id
=
"y2"
>I'm the <
em
>included</
em
> h2</
h2
>
<
p
id
=
"i"
>double</
p
>
</
body
>
</
html
>
|
主文档和导航
如果您有一个包含多个页面的演示,有一个主文档包含页面标题及其链接。您可以生成完整的导航,从每个页面到任何其他页面,以及到前一个和后一个页面。这些细节不属于本文的介绍范围,但 参考资料 中提供了使用主文档的 HTML 演示的链接。可将 .xml 替换为 .html 来获得编译后的版本。让浏览器向您显示 .xml 的整洁源代码。您会对它生成的源代码量感到惊奇。
解释与编译的对比
解释意味着页面为 XML 格式(其文件扩展名为 .xml,其内容类型为文本/xml 或应用程序/xml),并且处理指令所引用的 XSLT 样式表可在浏览器中执行。
编译意味着浏览器看到的是 HTML(其文件扩展名为 .html,内容类型为文本/html),它是在请求页面之前从您的开发环境中或服务器上的 XML 转换而来的。Xalan 和 Saxon 都是著名的 XSLT 处理器。
解释是未来的发展方向。所有现代浏览器都支持 XSLT,并且它具有一些优点:
- 当测试时,您会立即获得结果。只需在您测试的每个浏览器中按下 F5,即可反映源页面、CSS 和 XSLT 样式表的更改。
- 要传递给客户端的信息量减少了。
- 客户端看到的是一个干净、整洁的网页,因为还未生成扩充内容。
但也要注意一些缺点:
- 有一些旧浏览器可能不支持 XSLT。如果向一个受控的环境(内部网)发布页面,就不会出现问题。
- 一些现代浏览器禁止 XSLT 样式表引用另一个目录中的另一个样式表。
- 将 XSLT 域其他功能相结合(比如 SVG 或 iframe)可能在一些浏览器中导致问题。
- 因为大部分浏览器都不支持 XSLT 2.0 或即将推出的 3.0,所以您无法使用新功能。没有 XPath 2.0
if () then … else
,也没有用户编写的 XPath 功能。
无论进行编译还是解释,页面的其他转换 (CSS、JavaScript) 都会在 XSLT 转换之后执行。
结束语
在本文中,您学习了如何使用 XSLT 样式表来扩充 XHTML 文档。您可以使用本文中的示例作为起点,构建您自己的 XSLT 样式表。
下载
关于作者
Jürgen M. Regel 是位于德国汉诺威的 TUI InfoTec GmbH 的 Architecture Management & Software Engineering 部门的一名高级软件工程师。他主要研究旅游行业中的企业应用程序集成 (EAI)。