rapodaca's cheminfbenchmark at master - GitHub - 0 views
-
Mike Chelen on 31 Jan 09Cheminformatics Benchmark Toolkit --------------------------------- Why? Because we want to: 1. measure if a patch improves performance 2. compare cheminformatics toolkits Project ------- This is an open project, and anyone can contribute or patch the tests, in order to attempt to create a balanced and fair test suite impartial to the toolkit being tested. Versioning ---------- When reporting results for a toolkit for a test, always report the version of ChemInfBenchmark release, as well as the test number as listed below. The tests --------- Toolkits should always to the bare minimum; any extra things will only make the outcome unfavourable, but toolkits may do that if they wish. 1. Convert an SD file to XYZ 2. Calculate the molecular mass of the entries in an SD file 3. Convert an SD file to CML 2.5 4. Perform a substructure match on all of the entries in an SD file 5. Parse a collection of SMILES - this will be a measure of performance and completeness 6. Detect rings in a collection of molecules. 7. Parse (not perform the search) for a collection of SMARTS patterns - this will be a measure of performance and completeness Running tests ------------- Each test is run via a configure file. A test can be run by $ ant -Dconfig=test-config.xml run