Is Flax Alive?

[Paul]

Hello,
Will be starting an eval shortly and was considering evaluating
Xapian via flax. Is this a good idea? The mailing list for this
project makes me think that not much is going on with it, but I do see
some commits from not too long ago.

If using flax, I guess I should use the trunk, since there have been
no new releases in the last year?

Please let me know your thoughts and if there is anything I should be
looking out for.

Thanks,
Paul

[Charlie Hull]

Hi Paul,

Flax is very much alive, for example we've recently committed flax.core (a thin interface to Xapian) and within a week there will be another module going in. You can read about flax.core on the Flax blog at www.flax.co.uk/blog .

If you could let me know a bit more about the aims of your project I'd be very happy to tell you how Flax might help. You could try Flax Basic, but it is a very simple and basic application and certainly doesn't demonstrate the full range of Xapian's capabilities.

Cheers

Charlie
--
Charlie Hull
Managing Director, Flax
Tel. +44 (0)8700 118334
www.flax.co.uk

--
You received this message because you are subscribed to the Google Groups "flax-discuss" group.
 To post to this group, send email to flax-d...@googlegroups.com
 To unsubscribe from this group, send email to flax-discuss...@googlegroups.com
 For more options, visit this group at http://groups.google.co.uk/group/flax-discuss?hl=en
For more about Flax, visit the homepage at http://www.flax.co.uk

[Paul]

Hi Charlie,
That is great to hear. We are currently in eval mode to replace an
aging proprietary search system that also has the capabilities to
build new advanced features on top of. We have ~19 million documents
of various types that range in size from less than 1K to 100K, with a
vast majority at the smaller end of the spectrum. Typically, 10-20
fields per document.

Most of the queries end up being large non-scored, non-stemmed,
boolean term filtering sorted by date. The query strings can get
pretty large, 30+ nested boolean filters in some extreme cases. Our
goal is to be able to swap in a new backend without updating any front
end code. We hope to be able to dynamically translate query strings to
the new search platforms syntax.

We are most concerned with verifying functionality, search
performance, and scaling with both number of documents and number/
variety of queries.

What are your thoughts on the most efficient way to eval xapian/flax
in this regard?

Thanks!
Paul

On 16 July, 15:06, Charlie Hull <char...@juggler.net> wrote:
> Hi Paul,
>
> Flax is very much alive, for example we've recently committed flax.core (a
> thin interface to Xapian) and within a week there will be another module
> going in. You can read about flax.core on the Flax blog atwww.flax.co.uk/blog.
>
> If you could let me know a bit more about the aims of your project I'd be
> very happy to tell you how Flax might help. You could try Flax Basic, but it
> is a very simple and basic application and certainly doesn't demonstrate the
> full range of Xapian's capabilities.
>
> Cheers
>
> Charlie
> --
> Charlie Hull
> Managing Director, Flax
> Tel. +44 (0)8700 118334www.flax.co.uk
>

[Paul]

Just wanted to post that Charlie and I had further discussions off
this thread.

My conclusion is that while flax looks like an interesting start,
there are still core features missing that prevent this from being an
out of the box solution. For example, it is not possible to search
across two databases with flax search service(at the moment). It is
opensource and one can add these features, if they wanted, but the
fact that they are missing to begin with makes me not consider flax as
a contender for my evaluation.

Thanks,
Paul

[Pierpaolo Baldan]

I Paul

my name is Pierpaolo, I love search engine and staff like that.

I'm sure that you and yours staff considering solr as search engine, but if you are looking at a very out of the box solution , for my experience, the google search appliance is a good ones.

Ciao

 

 

2010/8/24, Paul <ppe...@gmail.com>: