Is your feature request related to a problem? Please describe.
It is somewhat difficult to describe relations among compounds, e.g. X is enantiomeric to/diastereomeric to/derivative of Y. Some of these relations seem important as they are used to defined measurements (e.g. selectivity).
Describe the solution you'd like
An unambiguous way to describe relations among compounds.
Describe alternatives you've considered
If compounds have unique identifiers (can different Compound have duplicate CompoundIdentifier?) then we can
record relations explicitly as triples, or
add CompoundRelation and a compound_relations field to Compound and ProductCompound.
Compound
...
compound_relations
type: is_derivative_product_of # can be an enum
value: <compound identifier>
This could help queries. E.g. what could be used to dissolve reagent X? The simple way is to find compounds with reaction_role==solvent for all reactions including X. But there are cases where X was first dissolved in Y to form a ReactionInput and Y is not the solvent for the reaction (like amalgam). If we have relations like X is_directly_dissolved_in Y then we can query against the relation directly. This could also give better solubility estimates compared to "the simple way" where solvent quantity is recorded for the entire reaction mixture.
Is your feature request related to a problem? Please describe. It is somewhat difficult to describe relations among compounds, e.g. X is enantiomeric to/diastereomeric to/derivative of Y. Some of these relations seem important as they are used to defined measurements (e.g. selectivity).
Describe the solution you'd like An unambiguous way to describe relations among compounds.
Describe alternatives you've considered If compounds have unique identifiers (can different
Compoundhave duplicateCompoundIdentifier?) then we canCompoundRelationand acompound_relationsfield toCompoundandProductCompound.This could help queries. E.g. what could be used to dissolve reagent X? The simple way is to find compounds with
reaction_role==solventfor all reactions including X. But there are cases where X was first dissolved in Y to form aReactionInputand Y is not the solvent for the reaction (like amalgam). If we have relations likeX is_directly_dissolved_in Ythen we can query against the relation directly. This could also give better solubility estimates compared to "the simple way" where solvent quantity is recorded for the entire reaction mixture.