Add industry benchmarkset

[jthorton]

This PR adds the industry benchmark dataset prepared here. Each dataset contains the following files:

• protein.pdb
• ligands.sdf
• cofactors.sdf: if present in the system

Future PRs will add other versions of the ligands using different charges and different network types.
Datasets should then be constructed via the combination of a network + ligands with a charge model.

Datasets:

• charge_annihilation_set
• fragments
• jacs set
• janssen_bace
• mcs_docking_set
• merck
• miscellaneous_set
• water set

[hannahbaumann]

Thanks @jthorton , lgtm!