Hi Alan,

> It's worth noting, however - and perhaps more for other readers here, given 
> fSphinx is a Python solution, and Thinking Sphinx is aimed at Rubyists that 
> don't want to write their Sphinx configuration by hand

I'm not sure about that. Personally I was looking for a solution that
would allow me to keep the flexibility of Sphinx (using configuration
files) while doing faceted search regardless of the programming
language. I looked into Thinking Sphinx but thought it was too rigid
and would not allow me to use all the power of Sphinx under its hood.
So then I worked on a lower level module to facilitate faceted search
with Sphinx.

> that Thinking Sphinx only CRC's string values. This is because Sphinx 1.10 
> doesn't support grouping on string facets, and previous Sphinx versions 
> didn't support string facets at all.

Yes and fSphinx provides a solution to have string facets and that
without storing all the terms in memory. However, with the release of
Sphinx 2.0.1, string facets will also be natively supported in
fSphinx.

> Any non-string facets aren't CRC'd by Thinking Sphinx, as Sphinx understands 
> them natively. So, in a lot of cases, collisions are avoided.

According to the Thinking Sphinx doc: "If the field is a string (which
is the case in most situations), then the value is converted to a
CRC32 integer." Which will obviously create collisions which basically
means that at the moment you can really only use Thinking Sphinx for
non-string facets. Of course I trust that string attributes will be
quickly implemented.

> Now that there's the release of Sphinx 2.0.1-beta, it looks like we can get 
> some better string facet support happening, as has already been discussed on 
> this list between Clemens Kofler and myself. With a bit of luck, Clemens and 
> I will have some time while we're in the same city to pair program and allow 
> TS to degrade gracefully to the existing solution for those not using Sphinx 
> 2.0.1.

Yes as it will be in fSphinx.


So far fSphinx has been used on some pretty cool projects:

http://imdb.cloudmining.net (1M documents, each document with hundreds
of facet values)
http://medline.cloudmining.net (20M documents, each document with
hundreds of facet values, distributed on 4 cores)

In fact fSphinx is the backend of Cloud Mining and Cloud Mining is a
web interface to fSphinx.

Regards,

Alex

> > Hi everyone,
>
> > Thinking Sphinx is really cool. However its faceted search handling
> > seems pretty limited. Each facet value is converted to a CRC32
> > integer. This will create a lot of collisions once passed millions of
> > facet values.
>
> > If you are looking for more a flexible and scalable way of performing
> > faceted search, have a look at fSphinx:
>
> >https://github.com/alexksikes/fSphinx
>
> > Cheers,
>
> > Alex
>
> > --
> > You received this message because you are subscribed to the Google Groups 
> > "Thinking Sphinx" group.
> > To post to this group, send email to [email protected].
> > To unsubscribe from this group, send email to 
> > [email protected].
> > For more options, visit this group 
> > athttp://groups.google.com/group/thinking-sphinx?hl=en.

-- 
You received this message because you are subscribed to the Google Groups 
"Thinking Sphinx" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/thinking-sphinx?hl=en.

Reply via email to