May I suggest a sub-problem with which you will have to grapple anyhow, and will help you stage your work?
This is to "canonicate" a chemical graph. The procedure accepts as input a description of a specific structure in
arbitrary form, and outputs one in accordance with the canons of Dendral order. This is a vital part of the structure-generator,
but does not yet require a lot of book-keeping detail needed for stepping from one to the next radial in sequence. Please
do not hesitate to let me know of your problems and successes.