Superstructure search

[JJ-Pineda]

Description Resolves #86 by introducing superstructure search functionality and CLI and API endpoints. This means that we are searching for compounds that are substructures of the query compound (i.e. smaller than the query compound). Similar to substructure search and identity search, superstructure search uses scaffolds to speed up searches.

Important Notes

• Indexing endpoints have been updated such that compounds with no assigned scaffolds are indexed with a scaffold ID of "-1". This makes it easier to distinguish between compounds that have scaffolds and those that don't.

[JJ-Pineda]

For now, this is fast as superstructure gets. Performing a comprehensive superstructure search for paclitaxel takes ~7 seconds, but this is an atypically complex compound.

I tried to add "NOT" queries for non-matching scaffolds in an attempt to narrow down search space, but the speed up was negligible (~0.2-0.3 seconds for paclitaxel).