Mailing-List: contact dev-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@jackrabbit.apache.org
Received-SPF: pass (nike.apache.org: domain of monkiki@gmail.com designates
 209.85.218.214 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:from:date:message-id:subject:to
         :content-type;
        b=k8BKTqp9UvBmkFIUm02ZM8nnqmC9h4uuleD0rAfHLEmqBazfYJEpjUlRSDAxARxmRl
         Cgf0Sl+7+JefzjGvBhbyF4Xs99qehWZuzyeXP1wSZoVvxfYaj7hdzoI+rSGW+qQP+msg
         SSNGP/gJPExy9z54yE9aStsi0qbWkF2kcjG4E=
MIME-Version: 1.0
In-Reply-To: <510143ac0911250526t4cf0d440t2557850d2badd067@mail.gmail.com>
References: <8f70390911241153l56758814pd0023bf5e4dba738@mail.gmail.com>
	<510143ac0911250526t4cf0d440t2557850d2badd067@mail.gmail.com>
From: Paco Avila <monkiki@gmail.com>
Date: Wed, 25 Nov 2009 18:00:04 +0100
Message-ID: <8f70390911250900w24daa737m12636cf80e69b177@mail.gmail.com>
Subject: Re: detect a failed text extraction?
To: dev@jackrabbit.apache.org
Content-Type: text/plain; charset=ISO-8859-1

On Wed, Nov 25, 2009 at 2:26 PM, Jukka Zitting <jukka.zitting@gmail.com> wrote:
> Hi,
>
> On Tue, Nov 24, 2009 at 8:53 PM, Paco Avila <monkiki@gmail.com> wrote:
>> There is any way to detect a failed text extraction ? I know, I can
>> see the log but the failure it not associated to a file or path.
>> [...]
>> I have posted this question in the user list, but I think it is
>> interesting talking about how it can be achieved.
>
> Could we solve this by improving the level of logging in the indexer?
>
> Alternatively, if you don't have easy access to the log files, we
> could possibly inject some special unique term to the index as a
> marker of failed text extraction. That way you could query for all
> nodes for which text extraction failed.

Increasing the log level can be a goog approach: the objective is link
a failed text extraction with a node path. This way, I can see if the
submitted document has failed in the text extraction process. The
other approach (injecting a special term) also is very cute because I
can get a list of failed indexed document from a XPath query. Both
solutions can be combined to improve the jackrabbit experience: the
XPath query give a list of unindexed document and the log can hep to
know what failed in the text extraction.

> Finally, as a debugging tool we could add a feature to the Jackrabbit
> webapp that allows you to download the extracted text content of a
> binary instead of the binary itself. We'd simply run a new text
> extraction pass on the stored binary and return the extracted text or
> any encountered errors to he client.

This also can be interesting.

>
> BR,
>
> Jukka Zitting
>

-- 
Paco Avila
OpenKM
http://www.openkm.com
http://www.guia-ubuntu.org