Prerequisites for searching

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Prerequisites for searching

Fredrik Andersson-2-2
Hey guys.

Here's probably a quite odd question. If I manually create a new
segment (also manually constructing the ParseText, ParseData and
Content objects for the write/append), and run the indexer on that
segment - what information am I missing for performing a successful
search (from the usual web interface) on those entries that I wrote ?
That is, what is needed, other than a ./segments directory for the
actual search? A 'db' directory?

It would take me an hour to explain what I am doing, but the segment
and index looks great after I created them. If I 'grep' after the
content that I manually wrote in to the segments, it is found in both
in the segment and in the index (binary form, grep recognizes that
also).

Fuzzy question, I know, but any input would be extremely appreciated.

Greetings,
Fredrik
Reply | Threaded
Open this post in threaded view
|

Re: Prerequisites for searching

Andrzej Białecki-2
Fredrik Andersson wrote:

> Hey guys.
>
> Here's probably a quite odd question. If I manually create a new
> segment (also manually constructing the ParseText, ParseData and
> Content objects for the write/append), and run the indexer on that
> segment - what information am I missing for performing a successful
> search (from the usual web interface) on those entries that I wrote ?
> That is, what is needed, other than a ./segments directory for the
> actual search? A 'db' directory?
>

You shouldn't be missing any information. Only this data is needed for
searching. WebDB is not needed for searching.

Hint: use SegmentWriter to hide some of the details of writing to a segment.


--
Best regards,
Andrzej Bialecki     <><
  ___. ___ ___ ___ _ _   __________________________________
[__ || __|__/|__||\/|  Information Retrieval, Semantic Web
___|||__||  \|  ||  |  Embedded Unix, System Integration
http://www.sigram.com  Contact: info at sigram dot com

Reply | Threaded
Open this post in threaded view
|

Re: Prerequisites for searching

Fredrik Andersson-2-2
Thanks Andrzej for the speedy reply.

I am using the SegmentWriter (is there another way?). Still no
resolvment though... I just tried merging a "functional" (searchable)
segment with my manually created segment, rebuilt the index and popped
the search - neither page could be found (there's only one page in
each segment, for debugging).
I tried opening my indices with Luke, and it could open both my
automatically generated index and the one create manually. So the
segment/index should be ok, it's just that isn't search-compatbile : )
If I search my indices from within Luke, they show up as well, I just
noticed.

Really strange. I'll hack it some more tomorrow, and if you come to
think of any rookie error I might have made, don't hesitate to notify
me.

Greetings,
Fredrik

On 7/18/05, Andrzej Bialecki <[hidden email]> wrote:

> Fredrik Andersson wrote:
> > Hey guys.
> >
> > Here's probably a quite odd question. If I manually create a new
> > segment (also manually constructing the ParseText, ParseData and
> > Content objects for the write/append), and run the indexer on that
> > segment - what information am I missing for performing a successful
> > search (from the usual web interface) on those entries that I wrote ?
> > That is, what is needed, other than a ./segments directory for the
> > actual search? A 'db' directory?
> >
>
> You shouldn't be missing any information. Only this data is needed for
> searching. WebDB is not needed for searching.
>
> Hint: use SegmentWriter to hide some of the details of writing to a segment.
>
>
> --
> Best regards,
> Andrzej Bialecki     <><
>   ___. ___ ___ ___ _ _   __________________________________
> [__ || __|__/|__||\/|  Information Retrieval, Semantic Web
> ___|||__||  \|  ||  |  Embedded Unix, System Integration
> http://www.sigram.com  Contact: info at sigram dot com
>
>
Reply | Threaded
Open this post in threaded view
|

Re: Prerequisites for searching

Fredrik Andersson-2-2
For archive follow-up: Problem semi-resolved. Search can be done with
the IndexSearcher class, but not through the Tomcat/web interface, so
there's probably just some odd configuration issue with that. I'll
append the answer to this configuration issue if and when I need to
figure it out.

On 7/19/05, Fredrik Andersson <[hidden email]> wrote:

> Thanks Andrzej for the speedy reply.
>
> I am using the SegmentWriter (is there another way?). Still no
> resolvment though... I just tried merging a "functional" (searchable)
> segment with my manually created segment, rebuilt the index and popped
> the search - neither page could be found (there's only one page in
> each segment, for debugging).
> I tried opening my indices with Luke, and it could open both my
> automatically generated index and the one create manually. So the
> segment/index should be ok, it's just that isn't search-compatbile : )
> If I search my indices from within Luke, they show up as well, I just
> noticed.
>
> Really strange. I'll hack it some more tomorrow, and if you come to
> think of any rookie error I might have made, don't hesitate to notify
> me.
>
> Greetings,
> Fredrik
>
> On 7/18/05, Andrzej Bialecki <[hidden email]> wrote:
> > Fredrik Andersson wrote:
> > > Hey guys.
> > >
> > > Here's probably a quite odd question. If I manually create a new
> > > segment (also manually constructing the ParseText, ParseData and
> > > Content objects for the write/append), and run the indexer on that
> > > segment - what information am I missing for performing a successful
> > > search (from the usual web interface) on those entries that I wrote ?
> > > That is, what is needed, other than a ./segments directory for the
> > > actual search? A 'db' directory?
> > >
> >
> > You shouldn't be missing any information. Only this data is needed for
> > searching. WebDB is not needed for searching.
> >
> > Hint: use SegmentWriter to hide some of the details of writing to a
> segment.
> >
> >
> > --
> > Best regards,
> > Andrzej Bialecki     <><
> >   ___. ___ ___ ___ _ _   __________________________________
> > [__ || __|__/|__||\/|  Information Retrieval, Semantic Web
> > ___|||__||  \|  ||  |  Embedded Unix, System Integration
> > http://www.sigram.com  Contact: info at sigram dot com
> >
> >
>