Proper usage of JTidy to purify HTML
I am trying to use JTidy (jtidy-r938.jar) to sanitize an input HTML string, but I seem to have problems getting the default settings right. Often strings such as "hello world" end up as "helloworld" after tidying. I wanted to show what I'm doing here, and any pointers would be really appreciated:
Assume that rawHtml
is the String containing the input (real world) HTML. This is what I'm doing:
Tidy tidy = new Tidy();
tidy.setPrintBodyOnly(true);
ByteArrayOutputStream baos = new ByteArrayOutputStream();
PrintStream ps = new PrintStream(baos);
tidy.parse(new StringReader(rawHtml), ps);
return baos.toString("UTF8");
First off, does anything look fundamentally wrong with the above code? I seem to be getting weird results with this.
For example, consider the following input:
<p class="MsoNormal" style="text-autospace:none;"><font color="black"><span style="color:black;">???</span></font><b><font color="#7f0055"><span style="color:#7f0055;font-weight:bold;">private</span></font></b><font color="black"><span style="color:black;"> String parseDescription</span></font><font>
The output is:
<p class="MsoNor开发者_如何学Gomal" style="text-autospace:none;"><font color=
"black"><span style="color:black;"> </span></font>
<b><font color="#7F0055"><span style=
"color:#7f0055;font-weight:bold;">private</span></font></b><font
color="black"><span style="color:black;">String
parseDescription</span></font></p>
So,
"public String parseDescription" becomes "publicString parseDescription"
Thanks in advance!
Have a look at how JTidy is configured:
StringWriter writer = new StringWriter();
tidy.getConfiguration().printConfigOptions(writer, true);
System.out.println(writer.toString());
Maybe it then get clear what causes the problem.
What is weird? Little example, of actual output and expected... maybe ?
Well, this seems to be a bug in Jtidy. For the exact file which causes problems, refer here:
http://sourceforge.net/tracker/?func=detail&aid=2985849&group_id=13153&atid=113153
Thanks for all the help folks!
Here is how we are calling JTidy from Ant. You may infer the API call from it:
<tidy destdir="${build.dir.result}">
<fileset dir="${src}" includes="**/*.htm"/>
<parameter name="tidy-mark" value="false"/>
<parameter name="output-xml" value="no"/>
<parameter name="numeric-entities" value="yes"/>
<parameter name="indent-spaces" value="2"/>
<parameter name="indent-attributes" value="no"/>
<parameter name="markup" value="yes"/>
<parameter name="wrap" value="2000"/>
<parameter name="uppercase-tags" value="no"/>
<parameter name="uppercase-attributes" value="no"/>
<parameter name="quiet" value="no"/>
<parameter name="clean" value="yes"/>
<parameter name="show-warnings" value="yes"/>
<parameter name="break-before-br" value="yes"/>
<parameter name="hide-comments" value="yes"/>
<parameter name="char-encoding" value="latin1"/>
<parameter name="output-html" value="yes"/>
</tidy>
精彩评论