📘 EvaluateSmilesEquivalence

The EvaluateSmilesEquivalence operator is designed to assess the equivalence of SMILES (Simplified Molecular-Input Line-Entry System) strings. For each row in the input data, it compares a list of "golden" (ground truth) SMILES against a list of synthesized SMILES. The comparison is done by canonicalizing both SMILES strings using RDKit and checking for equality. The operator then calculates and outputs detailed results, block-level scores, and an overall summary of the evaluation.

__init__

def __init__(self, llm_serving: LLMServingABC = None)

Parameter	Type	Default	Description
llm_serving	LLMServingABC	None	Large language model serving instance. Note: This parameter is not used by this operator.

Prompt Template Descriptions

run

def run(self, storage: DataFlowStorage, golden_key: str = "golden_label", synth_key: str = "synth_smiles", output_key: str = "final_result")

Parameter	Type	Default	Description
storage	DataFlowStorage	Required	Data flow storage instance for reading and writing data.
golden_key	str	"golden_label"	The column name in the input data that contains the ground truth SMILES information.
synth_key	str	"synth_smiles"	The column name in the input data that contains the synthesized SMILES to be evaluated.
output_key	str	"final_result"	The column name where the detailed list of comparison results will be stored.

🧠 Example Usage

🧾 Output Format

The run function adds the following columns to the DataFrame in the storage:

Field	Type	Description
final_result	list[dict]	A list of dictionaries, each containing the evaluation result for one SMILES string. Each dictionary includes abbreviation, full_name, smiles, and a score (1 for a match, 0 otherwise).
block_score	int	The total number of correct SMILES matches for the row.
block_total	int	The total number of ground truth SMILES entries for the row.
block_accuracy	float	The accuracy for the row, calculated as block_score / block_total.