[NUTCH-255] Regular Expression for RegexUrlNormalizer to remove jsessionid

• Type: Improvement
• Status: Closed
• Priority: Trivial
• Resolution: Duplicate
• Affects Version/s: 0.8
• Component/s: fetcher

Environment

Windows XP Media Center 2005, 2 Gigs RAM, 3.0 Ghz Pentium 4 Hyperthreaded, Eclipse 3.2.0

• Assignee: Unassigned
• Reporter: Dennis Kubes
• Created: 26/Apr/06 00:42
• Updated: 22/Sep/08 15:12
• Resolved: 22/Sep/08 15:12

Description

Some URLs are filtered out by the crawl url filter for special characters (by default). One of these is the jsessionid urls such as:

http://www.somesite.com;jsessionid=A8D7D812B5EFD3099F099A760F779E3B?query=string

We want to get rid of the jessionid and keep everything else so that it looks like this:

http://www.somesite.com?query=string

Below is a regular expression for the regex-normalize.xml file used by the RegexUrlNormalizer that sucessfully removes jsessionid strings while leaving the hostname and querystring. I have also attached a patch for the regex-normalize.xml.template file that adds the following expression.

<regex>
<pattern>(.*)(;jsessionid=[a-zA-Z0-9]

{32}

)(.*)</pattern>
<substitution>$1$3</substitution>
</regex>

Issue Links

• incorporates: NUTCH-243 Some meta-refresh urls get ignored due to matching regular expression
• is part of: NUTCH-279 Additions for regex-normalize