CompoundSearch produces different scores with subset library about maven HOT 7 CLOSED

Please see https://github.com/PMSeitzer/maven/issues/5 for more information.

from maven.

I have reanalyzed using 20190509-Calico-Lipids-neg.msp, and I can confirm that the bug still exists.

from maven.

The problem seems to be this:
a peakgroup matches to multiple compounds. One matches higher than all the others.
In the compoundSearch widget, this appears to be working correctly.
However, in the DB Search widget, this is incorrect - all candidates show the score
of the top-scoring match, instead of the score associated with that candidate.

The correct behavior would be to exclude the rows of all matches except for the top matches.
These matches can always be retrieved using the 'Compound Search' widget.

from maven.

Alternatively, we could take a "compound-centric" view, where we are interested in the top-scoring feature for a compound.

from maven.

Desired options in GUI for peakgroup <--> compound+adduct matches:

Offer options to

• return all compound+adduct matches for peak group
• return one top hit (in the event of a tie, choose alphabetical)
• return all compounds with highest score

Offer option to retain peak groups that do not match to any compounds

from maven.

This work will be done in issue-24.

from maven.

Waiting to close this bug to see if Eugene would like any adjustments.

from maven.