Re: /select results different between 5.4 and 6.1

2016-08-19 Thread John Bickerstaff
Many thanks. On Fri, Aug 19, 2016 at 4:22 PM, Anshum Gupta wrote: > The default similarity changed from TF-IDF to BM25 in 6.0. > > On Fri, Aug 19, 2016 at 3:00 PM John Bickerstaff > > wrote: > > > Bump! > > > > TL;DR Question: Are scores (and

Re: /select results different between 5.4 and 6.1

2016-08-19 Thread Anshum Gupta
The default similarity changed from TF-IDF to BM25 in 6.0. On Fri, Aug 19, 2016 at 3:00 PM John Bickerstaff wrote: > Bump! > > TL;DR Question: Are scores (and debug output) *expected* to be different > between 5.4 and 6.1? > > On Thu, Aug 18, 2016 at 2:44 PM, John

Re: /select results different between 5.4 and 6.1

2016-08-19 Thread John Bickerstaff
Bump! TL;DR Question: Are scores (and debug output) *expected* to be different between 5.4 and 6.1? On Thu, Aug 18, 2016 at 2:44 PM, John Bickerstaff wrote: > Hi all, > > TL:DR - > Is it expected that the /select endpoint would produce different > scores/result order

/select results different between 5.4 and 6.1

2016-08-18 Thread John Bickerstaff
Hi all, TL:DR - Is it expected that the /select endpoint would produce different scores/result order between versions 5.4 and 6.1? (I'm aware that it's certainly possible I've done something different to these environments, although at this point I can't see any difference in configs etc... and