Command-line interface

After installing chemFilters, the chemfilters command is available for filtering datasets from the terminal.

Usage

chemfilters -i INPUT -o OUTPUT [OPTIONS]

Required arguments

Flag	Description
-i, --input	Path to the input file containing SMILES strings
-o, --output	Path to the output CSV file

Optional arguments

Flag	Default	Description
-c, --col-name	None	Column name for CSV input. If not provided, input is treated as a text file with one SMILES per line
--rdkit-filter	enabled	Apply RDKit structural alert filters
--no-rdkit-filter		Disable RDKit filters
--rdkit-subset	"ALL"	Subset of RDKit filters to apply (e.g., PAINS, CHEMBL, BRENK, ALL)
--rdkit-valtype	"string"	Output format: "string" for descriptions, "bool" for boolean flags
--pep-filter	enabled	Apply peptide filters (PepSift)
--no-pep-filter		Disable peptide filters
--silly-filter	enabled	Apply silly molecule filters (molspotter)
--no-silly-filter		Disable silly molecule filters
--bloom-filter	enabled	Apply purchasability filters (molbloom)
--no-bloom-filter		Disable purchasability filters
--std-mols	enabled	Standardize molecules before filtering
--no-std-mols		Skip standardization
--std-method	"chembl"	Standardization method (chembl, canon, papyrus, molvs)
--n-jobs	8	Number of parallel jobs
--chunk-size	-1	Process SMILES in chunks of this size. Negative means process all at once. Reduce if memory is limited

Examples

Filter a text file of SMILES using all filters:

chemfilters -i molecules.txt -o results.csv

Filter a CSV file, disabling peptide and silly filters:

chemfilters -i data.csv -c smiles_column -o results.csv \
    --no-pep-filter --no-silly-filter

Use only PAINS filters with boolean output:

chemfilters -i molecules.txt -o results.csv \
    --rdkit-subset PAINS --rdkit-valtype bool \
    --no-pep-filter --no-silly-filter --no-bloom-filter