Re: Weighting Schemes: Implementing Piv+ Normalization
Vivek Pal <[email protected]>
| Newsgroups | gmane.comp.search.xapian.devel |
|---|---|
| Message-ID | <CAGxcUfwV7pWVUUhXArrUFJQTvr2eAoH0hpbbT0-GiEaFaHwwdA@mail.gmail.com> |
> I can't comment on compare views, due to how github works. You can > open a pull request, which does allow commenting, and just note it > isn't ready to merge yet Opened a pull request: https://github.com/xapian/xapian/pull/115 I have made quite of changes since those previous commits. Like I mentioned in previous email, I tried using another constructor to pass parameters s and delta that control the weights assigned by Piv+ normalization. That prevented issues faced earlier. All previous tests pass without any serialisation errors and I've added new tests as well; all passing. Will try to improve test coverage further now. Thanks, Vivek On Sun, Jul 31, 2016 at 10:10 PM, James Aylett <[email protected]> wrote: > On Sat, Jul 30, 2016 at 01:55:11AM +0530, Vivek Pal wrote: > > > I've pushed all the changes I made so far > > > https://github.com/xapian/xapian/compare/master...ivmarkp:piv+?diff=split&name=piv%2B > > > > Can you please add some comments on it? Support for normalization > weighting > > is complete -- just these issues with serialisation. > > I can't comment on compare views, due to how github works. You can > open a pull request, which does allow commenting, and just note it > isn't ready to merge yet. (That will also get it built by Travis CI, > meaning people don't have to dig up the build failure from the mailing > list to figure out what's going on.) > > J > > -- > James Aylett, occasional trouble-maker > xapian.org > >