Fast full-text searching in Python (job for Whoosh?)

rbowman bowman at montana.com
Tue Mar 7 20:04:28 EST 2023


On Tue, 7 Mar 2023 07:33:01 -0500, Dino wrote:

> Played a little bit with both approaches in my little application.
> Re-requesting from the server seems to win hands down in my case.

That's necessary for a non-trivial data set. Assume you get 10 suggestions 
after the user type 'to'. 

today
tomorrow
tomato
tonsil
torque
totem
toad
toque
toward
touch

If the user type 'l' next and is trying for 'tolerance' you'll need a new 
set. You'll need a little refinement. If the user is a proficient typist 
and wants to type 'tolerance' they may get ahead of you. 

Another consideration is a less proficient typist or someone who can't 
spell. Again, play with maps.google.com. They're good at it. Put '123 
thomd' in the search bar. YMMV but I get 5 variations on 123 Thomas. When 
they were working down 'thompd' had zero matches so they backed up to 
'thom'.

If you play with their search they're using some more magic too.  Try '123 
ellekt'.  They may be using a variation on soundex or something more 
sophisticated. 


More information about the Python-list mailing list