The new Erlwood nodes add some very nice functionality, and it is nice to see an implementation of the reaction vectors algorithm. It would be really nice if the vector creation node could also be used to generate the vectors in an output table, and then be used as input into the reaction generator node, removing the need for a database external to Knime.
I originally did implement a node that generated a table of reaction vectors to be used as an input to the node. However since then, a colleague implemented a much faster method to search the DB and quickly apply a suitable RV using a previously stored fragmentation path. This is much more efficient and extremely fast compared to my original algorithm. Unfortunately this requires an external DB for fast access of the data.
Thanks - would it be possible to release the alternative DB-free nodes to the community too? Obviously, they may be less efficient but for proof-of-principle workflows they would be a real benefit!