You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@nutch.apache.org by Andrzej Bialecki <ab...@getopt.org> on 2011/03/08 21:57:40 UTC
Re: will nutch-2 be able to index image files
On 3/8/11 9:09 PM, alxsss@aim.com wrote:
> Hello,
>
> I wondered if nutch version 2 be able to index image files?
In what way? Extract metadata and index image metadata as text? Sure, if
we implement a plugin for it. Tika already supports EXIF, so this
shouldn't be complicated, perhaps it's a tweak to the parse-tika
configuration. Or did you mean the image content-based retrieval (e.g.
using wavelets)?
--
Best regards,
Andrzej Bialecki <><
___. ___ ___ ___ _ _ __________________________________
[__ || __|__/|__||\/| Information Retrieval, Semantic Web
___|||__|| \| || | Embedded Unix, System Integration
http://www.sigram.com Contact: info at sigram dot com
Re: will nutch-2 be able to index image files
Posted by al...@aim.com.
Hello,
Looks like I will have some spare time in the next month, so I may work on writing this image indexing plugin. I wondered if there is a similar plugin to leverage code from or follow it?
Thanks.
Alex.
-----Original Message-----
From: Andrzej Bialecki <ab...@getopt.org>
To: user <us...@nutch.apache.org>
Sent: Wed, Mar 9, 2011 12:24 am
Subject: Re: will nutch-2 be able to index image files
On 3/8/11 10:50 PM, alxsss@aim.com wrote:
> I meant to extract image title, src link and alt from<img tags and not store
image files. For a keyword search in must display link, which automatically
displays image itself in the search page.
> Not sure what do you mean image content-based retrieval? Do image files have
tags like mp3 ones?
Yes, for example http://en.wikipedia.org/wiki/Exchangeable_image_file_format
> Must a parse plugin be written in both cases?
Yes - most data is already available either in the DOM tree, or can be
obtained from a Tika image parser, it just needs to be wrapped in a plugin.
--
Best regards,
Andrzej Bialecki <><
___. ___ ___ ___ _ _ __________________________________
[__ || __|__/|__||\/| Information Retrieval, Semantic Web
___|||__|| \| || | Embedded Unix, System Integration
http://www.sigram.com Contact: info at sigram dot com
Re: will nutch-2 be able to index image files
Posted by Andrzej Bialecki <ab...@getopt.org>.
On 3/8/11 10:50 PM, alxsss@aim.com wrote:
> I meant to extract image title, src link and alt from<img tags and not store image files. For a keyword search in must display link, which automatically displays image itself in the search page.
> Not sure what do you mean image content-based retrieval? Do image files have tags like mp3 ones?
Yes, for example http://en.wikipedia.org/wiki/Exchangeable_image_file_format
> Must a parse plugin be written in both cases?
Yes - most data is already available either in the DOM tree, or can be
obtained from a Tika image parser, it just needs to be wrapped in a plugin.
--
Best regards,
Andrzej Bialecki <><
___. ___ ___ ___ _ _ __________________________________
[__ || __|__/|__||\/| Information Retrieval, Semantic Web
___|||__|| \| || | Embedded Unix, System Integration
http://www.sigram.com Contact: info at sigram dot com
Re: will nutch-2 be able to index image files
Posted by al...@aim.com.
I meant to extract image title, src link and alt from <img tags and not store image files. For a keyword search in must display link, which automatically displays image itself in the search page.
Not sure what do you mean image content-based retrieval? Do image files have tags like mp3 ones?
Must a parse plugin be written in both cases?
Thanks.
Alex.
-----Original Message-----
From: Andrzej Bialecki <ab...@getopt.org>
To: user <us...@nutch.apache.org>
Sent: Tue, Mar 8, 2011 12:58 pm
Subject: Re: will nutch-2 be able to index image files
On 3/8/11 9:09 PM, alxsss@aim.com wrote:
> Hello,
>
> I wondered if nutch version 2 be able to index image files?
In what way? Extract metadata and index image metadata as text? Sure, if
we implement a plugin for it. Tika already supports EXIF, so this
shouldn't be complicated, perhaps it's a tweak to the parse-tika
configuration. Or did you mean the image content-based retrieval (e.g.
using wavelets)?
--
Best regards,
Andrzej Bialecki <><
___. ___ ___ ___ _ _ __________________________________
[__ || __|__/|__||\/| Information Retrieval, Semantic Web
___|||__|| \| || | Embedded Unix, System Integration
http://www.sigram.com Contact: info at sigram dot com