Need a check to ensure children are actually children of the parent node

[shamelmerchant]

With the introduction of nitrogen and sulfur we have missed adding the hetroatoms in the parent nodes but define children nodes specific to N and S. This then results in the rate estimation just falling up to the head node and never descending down the tree.

Connie also sees other examples where this happens.

[bbuesser]

Where does it happen for nitrogen?

[shamelmerchant]

An example which I found was in H abstraction

I was trying to add a rate rule for L5: Cd/H2/NonDeN, H_rad but it kept on hitting X_H, H_rad because the parent two level up L3: Cd_H was only defined for C=C and not {C,N}.

[rwest]

I suggest:

1. get Database.isWellFormed() working
2. call isWellFormed() after loading (or importing) each database (maybe append to the load functions)
3. add a check to the end of isWellFormed() that ensures each node's parent is indeed a parent. You should be able to use the isSubgraphIsomorphic() method on each entry.
4. create some unit tests for all of the above!

[bbuesser]

@rwest In the comment of Database.isWellFormed() I have found a sentence I don't understand:

If no tree is present (e.g. the primary libraries), then every entry in the library must have an entry in the dictionary.

Is it referring to a Java-style library?

[bbuesser]

Oh, now I see how it works.

[connie]

This is now part of a unit test suit within RMG-Py/databaseTest.py