> >I'm tempted to write a python wrapper for
> > the library. Before I start doing this (and don't hold your breath), do
> > you know of any other try in this direction? (I couldn't find anything on
> > the swish-e website or searching the web :-)
> I've been hoping someone would ask that. I think it would be great to get
> a Python wrapper. Make sure you use the development version of the API it
> was redesigned for an object approach. Look at the perl/API.xs module for
Great, I'll have a look (but it might take some time until I'll really come
around doing it)
> >While I'm extreamly happy with swish-e so far, I'm missing the possibility
> >build an index incrementaly. Is there something planned in this direction?
> Yes. There is work in this direction. (Jose, can you say more?) One
> unsolved problem is how to *remove* a file from the index.
> For now, reindexing is the way to go.
> >Is it planned to expose the indexing functionality as well to a C library?
> No plan for that. In fact I spent a lot of time removing the indexing code
> from the library in the current dev version.
> I could see it helpful for incremental indexing, but for normal indexing I
> don't see it being much better than just forking and running the binary.
> But I haven't thought much about it. Why do you ask?
Thinking about a python module, it would be nice to be able to do everything
from within this module (without running external programs, at least in some
final version). Especially, when indexing something other than good old files
(yes, I've seen the new feature with external programs that feed swish-e)
Anyway, all functionality in a lib would be great, but since all the nice
features can be accessed und run by the shell commands, it's not really a
My the way, everybody from my company who got into contact with the swish-e
search results, was quite impressed (they were used to MS indexing server,
that was probably not configured in the right way).
You are really doing a quite impressing work.
Received on Thu Nov 14 18:14:14 2002