Open davidfarinajr opened 4 years ago
We need to load the RMG-database reaction family to get the recipe and the top node for the template. But we don't need the full tree. Probably hard to load the top nodes without loading the rest of the tree. But we can probably avoid doing the steps of adding in the training reactions and filling up by averaging, which often take a long time.
When estimating TS distances for a reaction, we currently load RMG-database and AutoTST database. For example, when we create an autotst H Abstraction reaction and want to generate a TS, AutoTST loads both the RMG-database H Abstraction kinetics tree and the AutoTST H Abstraction TS tree. I don't think its necessary to load the RMG H Abstraction kinetics tree.