Re: Xapian-devel Digest, Vol 151, Issue 7

Uppinder Chugh <[email protected]> Tue, 27 Feb 2018 19:01:23 +0530
Newsgroups gmane.comp.search.xapian.devel
Message-ID <CAL3if+D7Sduu_0rOGwe_Nof89Gb8jEkViMTys9wMUh=jSk2u0Q@mail.gmail.com>
Thanks for the timely reply, James.

> You should probably look at how our letor and clustering APIs work, since
both of them need an MSet to get started. (Neither is yet part of a
release, so you'll need to check the source code rather than online API
documentation.)

I checked out the clustering and letor code; they seem to be working
directly with the mset after it has been generated by first ranking. So,
I'll be mentioning this in my proposal.

> You'll need to make a recommendation in your proposal. Will LC be useful
for clustering? Will the other clustering algorithms be useful for
diversification?

I think my query got misinterpreted; what I meant was: should I include the
LC clustering algorithm in xapian-core/cluster, or include it in the code
of my implementation of diversification. I believe this can be sorted out
later, but I just wanted to be thorough in my application in terms of
implementation.

> I'd read some of the tests we have written already, as well as a basic
introduction to testing. (This is a good start, based on python which is
fairly readable even if you haven't worked with it before:
http://www.diveintopython3.net/unit-testing.html)

Thanks for sharing the resource, I will have a look at it.

l believe I can get to writing the application now; I definitely will go
through the guides on Xapian GSOC pertaining to writing a proposal. If you
feel I'm missing something or should look into a bit more, please let me
know.

Thanks,
Uppinder