You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tika.apache.org by Otis Gospodnetic <ot...@yahoo.com> on 2010/08/08 07:06:04 UTC
Additional search for tika.a.o
Hello,
Over at http://search-lucene.com we index Tika's mailing lists, wiki, web site,
source code, javadoc, jira...
Would the community be interested in a patch that adds another search option to
the search box on tika.apache.org?
Oh, btw., this is what I just experienced with the current search on
tika.apache.org (note the times):
Found 36 results in 39.38 seconds. Displaying page 1 of 4, sorted by ...
Found 0 results in 11.713 seconds. Displaying page 0 of 0, sorted by ...
Found 1,177 results in 50.915 seconds. Displaying page 1 of 118, sorted by ...
This makes me think having an alternative option would be a good thing to have.
Assuming people are for this, any suggestions for how the search should function
by default or any specific instructions for how the search box should be
modified would be great!
Thanks,
Otis
Re: Additional search for tika.a.o
Posted by "Mattmann, Chris A (388J)" <ch...@jpl.nasa.gov>.
+1 to Jukka's comments.
Cheers,
Chris
On 8/9/10 3:13 PM, "Jukka Zitting" <ju...@gmail.com> wrote:
Hi,
On Sun, Aug 8, 2010 at 7:06 AM, Otis Gospodnetic
<ot...@yahoo.com> wrote:
> Would the community be interested in a patch that adds another search option to
> the search box on tika.apache.org?
Sure. I think the earlier consensus within the Lucene PMC was that any
reasonable Lucene-based search engines would be given equal standing
as our site search providers. It's a bit unfortunate that this may
lead to some inconsistency in the user interface, but I think we can
live with that until someone steps up to implement and maintain such a
search engine on Apache hardware.
> Assuming people are for this, any suggestions for how the search should function
> by default or any specific instructions for how the search box should be
> modified would be great!
Ideally the search box would allow the user to choose which provider
to use. Something like a cookie that remembers the user's selection
would be nice. If the user doesn't make an explicit selection of the
provider, then one should be selected randomly for the first search
and remembered afterwards for consistency.
It would be great if the required logic was implemented on the client
side using javascript, as otherwise we'd need to start messing up with
CGI scripts, etc.
The current site template can be found at [1].
[1] https://svn.apache.org/repos/asf/tika/site/src/site/site.vm
BR,
Jukka Zitting
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Email: Chris.Mattmann@jpl.nasa.gov
WWW: http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Re: Additional search for tika.a.o
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Sun, Aug 8, 2010 at 7:06 AM, Otis Gospodnetic
<ot...@yahoo.com> wrote:
> Would the community be interested in a patch that adds another search option to
> the search box on tika.apache.org?
Sure. I think the earlier consensus within the Lucene PMC was that any
reasonable Lucene-based search engines would be given equal standing
as our site search providers. It's a bit unfortunate that this may
lead to some inconsistency in the user interface, but I think we can
live with that until someone steps up to implement and maintain such a
search engine on Apache hardware.
> Assuming people are for this, any suggestions for how the search should function
> by default or any specific instructions for how the search box should be
> modified would be great!
Ideally the search box would allow the user to choose which provider
to use. Something like a cookie that remembers the user's selection
would be nice. If the user doesn't make an explicit selection of the
provider, then one should be selected randomly for the first search
and remembered afterwards for consistency.
It would be great if the required logic was implemented on the client
side using javascript, as otherwise we'd need to start messing up with
CGI scripts, etc.
The current site template can be found at [1].
[1] https://svn.apache.org/repos/asf/tika/site/src/site/site.vm
BR,
Jukka Zitting
Re: Additional search for tika.a.o
Posted by Otis Gospodnetic <og...@yahoo.com>.
Hi Chris,
Sorry, I didn't mean to say that the slowness was the reason for an
alternative. Providing an alternative search box was discussed (maybe on
private@lucene?) several months ago.
Regarding slow queries there - the slowness is gone now. Things were slow
yesterday when I sent the email. I just happened to try the search there and
when i saw the slowness the obvious thought happened - it's good to have a
backup, that's all.
So please think about the email without thinking about that slowness. :)
Otis
P.S.
The reply to Grant was:
The queries I tried were really basic, actually. Just query text + project
facet selected. The query text was simple stuff, like the word "parse" or
"search" or "solr".
It was slow regardless of the project I selected and I think regardless of the
query string, but was fast when the same query was executed twice - caching.
Probably some disk IO issues. Everything looks healthy now! :)
----- Original Message ----
> From: "Mattmann, Chris A (388J)" <ch...@jpl.nasa.gov>
> To: "dev@tika.apache.org" <de...@tika.apache.org>
> Sent: Mon, August 9, 2010 4:47:11 PM
> Subject: Re: Additional search for tika.a.o
>
> Hi Otis,
>
> What was your reply to Grant? I think his question was on topic because I
>experienced a similar behavior after seeing your email. That is, my queries to
>the existing Tika search box came back pretty quick, with little or no
>noticeable delay. Your original email thread seemed to cite that a potential
>reason for having an alternative search box might be that the existing one is
>slow?
>
> Cheers,
> Chris
>
>
>
> On 8/9/10 1:43 PM, "Otis Gospodnetic" <ot...@yahoo.com> wrote:
>
> Hello,
>
> I replied to Grant directly in order to keep this thread on topic.
>
> Otis
>
>
>
>
>
> ----- Original Message ----
> > From: Grant Ingersoll <gs...@apache.org>
> > To: dev@tika.apache.org
> > Sent: Mon, August 9, 2010 11:28:47 AM
> > Subject: Re: Additional search for tika.a.o
> >
> >
> > On Aug 8, 2010, at 1:06 AM, Otis Gospodnetic wrote:
> >
> > > Hello,
> > >
> > > Over at http://search-lucene.com we index Tika's mailing lists, wiki, web
> >site,
> >
> > > source code, javadoc, jira...
> > >
> > > Would the community be interested in a patch that adds another search
>option
> >
> >to
> >
> > > the search box on tika.apache.org?
> > >
> > > Oh, btw., this is what I just experienced with the current search on
> > > tika.apache.org (note the times):
> > >
> > > Found 36 results in 39.38 seconds. Displaying page 1 of 4, sorted by
...
> > > Found 0 results in 11.713 seconds. Displaying page 0 of 0, sorted by
...
> > > Found 1,177 results in 50.915 seconds. Displaying page 1 of 118, sorted
>by
>
> >...
> > >
> > > This makes me think having an alternative option would be a good thing
to
> >have.
> > > Assuming people are for this, any suggestions for how the search should
> >function
> >
> > > by default or any specific instructions for how the search box should be
> > > modified would be great!
> >
> > I'm all for alternatives, and it obviously up to the community to decide,
>but
>
> >might I ask what queries you are running? All of my searches are subsecond
>for
> >
> >the existing search for queries that I ran.
>
>
>
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> Chris Mattmann, Ph.D.
> Senior Computer Scientist
> NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
> Office: 171-266B, Mailstop: 171-246
> Email: Chris.Mattmann@jpl.nasa.gov
> WWW: http://sunset.usc.edu/~mattmann/
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> Adjunct Assistant Professor, Computer Science Department
> University of Southern California, Los Angeles, CA 90089 USA
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
>
Re: Additional search for tika.a.o
Posted by "Mattmann, Chris A (388J)" <ch...@jpl.nasa.gov>.
Hi Otis,
What was your reply to Grant? I think his question was on topic because I experienced a similar behavior after seeing your email. That is, my queries to the existing Tika search box came back pretty quick, with little or no noticeable delay. Your original email thread seemed to cite that a potential reason for having an alternative search box might be that the existing one is slow?
Cheers,
Chris
On 8/9/10 1:43 PM, "Otis Gospodnetic" <ot...@yahoo.com> wrote:
Hello,
I replied to Grant directly in order to keep this thread on topic.
Otis
----- Original Message ----
> From: Grant Ingersoll <gs...@apache.org>
> To: dev@tika.apache.org
> Sent: Mon, August 9, 2010 11:28:47 AM
> Subject: Re: Additional search for tika.a.o
>
>
> On Aug 8, 2010, at 1:06 AM, Otis Gospodnetic wrote:
>
> > Hello,
> >
> > Over at http://search-lucene.com we index Tika's mailing lists, wiki, web
>site,
>
> > source code, javadoc, jira...
> >
> > Would the community be interested in a patch that adds another search option
>
>to
>
> > the search box on tika.apache.org?
> >
> > Oh, btw., this is what I just experienced with the current search on
> > tika.apache.org (note the times):
> >
> > Found 36 results in 39.38 seconds. Displaying page 1 of 4, sorted by ...
> > Found 0 results in 11.713 seconds. Displaying page 0 of 0, sorted by ...
> > Found 1,177 results in 50.915 seconds. Displaying page 1 of 118, sorted by
>...
> >
> > This makes me think having an alternative option would be a good thing to
>have.
> > Assuming people are for this, any suggestions for how the search should
>function
>
> > by default or any specific instructions for how the search box should be
> > modified would be great!
>
> I'm all for alternatives, and it obviously up to the community to decide, but
>might I ask what queries you are running? All of my searches are subsecond for
>
>the existing search for queries that I ran.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Email: Chris.Mattmann@jpl.nasa.gov
WWW: http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Re: Additional search for tika.a.o
Posted by Otis Gospodnetic <ot...@yahoo.com>.
Hello,
I replied to Grant directly in order to keep this thread on topic.
Otis
----- Original Message ----
> From: Grant Ingersoll <gs...@apache.org>
> To: dev@tika.apache.org
> Sent: Mon, August 9, 2010 11:28:47 AM
> Subject: Re: Additional search for tika.a.o
>
>
> On Aug 8, 2010, at 1:06 AM, Otis Gospodnetic wrote:
>
> > Hello,
> >
> > Over at http://search-lucene.com we index Tika's mailing lists, wiki, web
>site,
>
> > source code, javadoc, jira...
> >
> > Would the community be interested in a patch that adds another search option
>
>to
>
> > the search box on tika.apache.org?
> >
> > Oh, btw., this is what I just experienced with the current search on
> > tika.apache.org (note the times):
> >
> > Found 36 results in 39.38 seconds. Displaying page 1 of 4, sorted by ...
> > Found 0 results in 11.713 seconds. Displaying page 0 of 0, sorted by ...
> > Found 1,177 results in 50.915 seconds. Displaying page 1 of 118, sorted by
>...
> >
> > This makes me think having an alternative option would be a good thing to
>have.
> > Assuming people are for this, any suggestions for how the search should
>function
>
> > by default or any specific instructions for how the search box should be
> > modified would be great!
>
> I'm all for alternatives, and it obviously up to the community to decide, but
>might I ask what queries you are running? All of my searches are subsecond for
>
>the existing search for queries that I ran.
Re: Additional search for tika.a.o
Posted by Grant Ingersoll <gs...@apache.org>.
On Aug 8, 2010, at 1:06 AM, Otis Gospodnetic wrote:
> Hello,
>
> Over at http://search-lucene.com we index Tika's mailing lists, wiki, web site,
> source code, javadoc, jira...
>
> Would the community be interested in a patch that adds another search option to
> the search box on tika.apache.org?
>
> Oh, btw., this is what I just experienced with the current search on
> tika.apache.org (note the times):
>
> Found 36 results in 39.38 seconds. Displaying page 1 of 4, sorted by ...
> Found 0 results in 11.713 seconds. Displaying page 0 of 0, sorted by ...
> Found 1,177 results in 50.915 seconds. Displaying page 1 of 118, sorted by ...
>
> This makes me think having an alternative option would be a good thing to have.
> Assuming people are for this, any suggestions for how the search should function
> by default or any specific instructions for how the search box should be
> modified would be great!
I'm all for alternatives, and it obviously up to the community to decide, but might I ask what queries you are running? All of my searches are subsecond for the existing search for queries that I ran.