You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The personal wiki I'm powering with Pepperminty Wiki has grown tot he point where searching is beginning to take ~1.1s. We should do some profiling to figure out if there's some optimising we can do to bring this down a bit.
The text was updated successfully, but these errors were encountered:
After doing some profiling with XDebug, a few things have come to light.
Looks like optimising the page id system has helped a lot! Next we need to utilise the inverted index in the search context extractor instead of running a regex. It's not the regex that's the issue curiously enough - it's the transliteration of the entire page content that's slowing us down.
Yas! that did teh trick. Now context generation takes a much more respectable 59ms on average - a ton better than the ~1100m s from before.
Also, turns out getting page ids is comparatively expensive - so spending some time both optimising it and minimising the number of calls to it has done wonders.
The personal wiki I'm powering with Pepperminty Wiki has grown tot he point where searching is beginning to take ~1.1s. We should do some profiling to figure out if there's some optimising we can do to bring this down a bit.
The text was updated successfully, but these errors were encountered: