AB3 Developer Notes

From DISI
Jump to navigation Jump to search

UNDER DEVELOPMENT

The back-end and the front-end of this tool can be accessed at ABBB and AB3FrontEnd repositories at https://gitlab.docking.org, respectively.

Backend

The backend runs on Flask and has two main scripts running that do all the main calculations, which are BespokeAnaloging.py and arthorquery.py. The Bespoke Analoging script gets called in the routes section with the call_ABBB function. There we use a JSON file with all our reaction information as the first input and a smiles string as the second input which we get from the front-end. This script takes those inputs and runs the smiles against all those reactions and spits out building blocks if any of the reactions worked on the smiles. Those building blocks as well as a Tanimoto coefficient can then be used in the following script: arthorquery.py. The script takes all the building blocks that were derived from the first script and uses Arthor to find analogs of the building blocks using similarity against the Tanimoto coefficient and inclusion and exclusion rules as a basis.

Restarting

AB3 runs on the Epyc node. If AB3 is down first start by ssh-ing into Epyc. Use sudo docker logs --since 30m ab3 to see logs of AB3 and check to see if there was a reason for the crash. Then you can use the following commands "sudo docker restart redis" and "sudo docker restart AB3" to get AB3 back up and running. If that doesn't work replace "restart" with "start" and run those commands again.