nutch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Markus Jelsma (JIRA)" <>
Subject [jira] [Commented] (NUTCH-1233) Rely on Tika for outlink extraction
Date Mon, 13 Aug 2012 17:00:38 GMT


Markus Jelsma commented on NUTCH-1233:

Hi Ken,

The rel attribute is extracted in Tika 1.1 so that's fine. I'd prefer the whitespace being
collapsed in Tika instead but it was for me quicker to fix in in Nutch. It also retrieves
more URL's in some cases, that's usually a good thing.

It seems odd indeed to not want it collapsed by Tika, but perhaps some funky academic research
would like to measure strange HTML on the web using Tika's outlink extractor. It could be
made configurable in Tika anyway.

Perhaps we should deliver a patch for Tika.
> Rely on Tika for outlink extraction
> -----------------------------------
>                 Key: NUTCH-1233
>                 URL:
>             Project: Nutch
>          Issue Type: Improvement
>            Reporter: Markus Jelsma
>            Assignee: Markus Jelsma
>             Fix For: 1.6
>         Attachments: NUTCH-1233-1.5-wip.patch, NUTCH-1233-1.6-1.patch, NUTCH-1233-1.6-2.patch
> Tika provides outlink extraction features that are not used in Nutch. To be able to use
it in Nutch we need Tika to return the rel attr value of each link, which it currently doesn't.
There's a patch for Tika 1.1. If that patch is included in Tika and we upgraded to that new
version this issue can be worked on. Here's preliminary code that does both Tika and current
outlink extraction. This also includes parts of the Boilerpipe code.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:!default.jspa
For more information on JIRA, see:


View raw message