Finding atoms/residues with the right properties

13 views
Skip to first unread message

Mateusz Bieniek

unread,
Jul 17, 2020, 2:16:26 PM7/17/20
to Whoosh
Hi everyone, 

I am playing with whoosh for looking up information about atoms. In one system there is anywhere from 100k to millions of atoms. In this case, I am adding documents with fieldname such as residue name, atom name, etc. 

I've noticed that looking up a small number of atoms by name works very fast. For example a specific residue name. Better than MDAnalysis. However, when I request a huge number of atoms, for example all oxygens in the water molecules, the performance degrades significantly and becomes slower than MDAnalysis. I attach a small script showing an example. 

Is there anything I could change in the configuration to make it faster? Is there any bottleneck related to returning/searching a huge number of documents?


Thanks, 
Mat
reuse_search_submit.py
Reply all
Reply to author
Forward
0 new messages