Closed jimmyjbling closed 1 month ago
Need to expand the list of CurationStep ranking to include label steps and a subset molecule step
new ranking will be:
1. RDKit mol creation (NEVER USE THIS! ONLY CurateValid SHOULD BE 1) 2. Subsets a molecule (e.g. mixture handling) 3. Changes/Standardizes a molecule (e.g. neutralize, flatten) 4. Excludes a molecule (e.g. inorganic) 5. Standardize labels (e.g. make numeric/categorical) 6. Modify labels (e.g. binarize continuous labels) 7. Duplicate handling (e.g. greedy duplicate handling 8. Duplicate handling that changes label values 9. Canonicalize SMILES (NEVER USE THIS! ONLY CurateCanonicalize SHOULD BE 9)
fixed this in 9fb1bf3
Need to expand the list of CurationStep ranking to include label steps and a subset molecule step
new ranking will be: