Similarity metrics fail if the SMILES have problems when loading to RDKit

IBM / Hestia-OOD

Independent evaluation set construction for trustworthy ML models in biochemistry

https://ibm.github.io/Hestia-OOD/

MIT License

7 stars 1 forks source link

Closed RaulFD-creator closed 1 month ago

RaulFD-creator commented 1 month ago

Describe the bug Similarity metrics fail if the SMILES have problems when loading to RDKit

To Solve Introduce method to sanitize the SMILES before going into the calculation.

RaulFD-creator commented 1 month ago

It should be done by user beforehand. I've included a santization to curve the easiest errors.