Remove usage of classic
similarity algorithm in elasticsearch
#7498
Labels
Milestone
classic
similarity algorithm in elasticsearch
#7498
We're using the
classic
similarity algorithm in elasticsearch at the moment, which isn't the default - it's there for compatibility with old elasticsearch versions, and we set that way back in #5785 to try to address issues we were having with search results after upgrading.That setting has been deprecated for a while, but now the documentation for 6.8 mentions this:
I've noticed the quality drop in my branch as part of the work to upgrade to 6.8. We should try to anticipate this and switch to the
default
algorithm and see if it's good enough for us, we've made a number of changes to search scoring since so my hope is that we don't actually need theclassic
algorithm anymore.The text was updated successfully, but these errors were encountered: