Evaluation

MDEvaluator — post-hoc metric computation for a single mockdock benchmark run.

Usage (CLI):

python -m mockdock.evaluator results.csv –benchmark CHK1 [–output eval_metrics.json]

Usage (Python API):

from mockdock import MDEvaluator ev = MDEvaluator(“CHK1”) metrics = ev.compute_metrics(Path(“results.csv”))

class mockdock.evaluator.MDEvaluator(benchmark_name, scratch_dir=None)

Bases: object

Post-hoc evaluator for a single mockdock benchmark run.

Delegates all config/bioactivity loading to BenchmarkLoader — no docking engine is instantiated, making this lightweight and safe to run on a login node.

Parameters:

• benchmark_name (str)

• scratch_dir (Path | str | None)

compute_metrics(results_csv, output_path=None)

Compute all metrics for one results.csv file and optionally write JSON.

Parameters:

• results_csv (Path | str)

• output_path (Path | str | None)

Return type:

dict

mockdock.evaluator.main()