You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Testing of MOPAC is based on simple regression tests that compare MOPAC output files between reference output files and the output files presently being generated by MOPAC. The comparisons are implemented in Python (tests/compare_output.py), orchestrated by CTest (tests/CMakeLists.txt), and monitored for coverage by CodeCov, as indicated by the coverage badge on the repository's main README file. Right now, the coverage is relatively low (68% at the time of writing this), and it should probably be in the high 80's or low 90's.
The submission of new tests that expand testing coverage is very welcome, particularly with an emphasis on any common use cases that are not presently represented in the existing tests. The tests in the main tests directory are a combination of historical regression tests used by Jimmy Stewart and validation tests of the INDO feature used by Rebecca Gieseking, and the keywords subdirectory contains keyword-specific tests I've written in the process of migrating keyword documentation between MOPAC's old website and the new website that is still in development.
Testing is particularly important for MOPAC because of its dual role as a popular, widely-used, and venerable electronic structure package and as the long-time primary codebase for the computational chemistry research work of Jimmy Stewart. In its first role, the core functionality of MOPAC has been heavily tested in practice by its large user base. However, there are many features and options in MOPAC produced over the course of Dr. Stewart's scientific career and many conceptually valid combinations of these features and options, not all of which see regular use. Some of these features may not be as robust as the core functionality, and if problems are encountered, it will be either by developers during the process of expanding test coverage or by users who are trying to apply MOPAC in uncommon and novel ways. The former scenario is much preferred to the latter.
The text was updated successfully, but these errors were encountered:
Here I attached a handful of tests, which are interesting to me (and cover some missing A-F keywords). @godotalgorithm let me know if these kind of files are useful.
Testing of MOPAC is based on simple regression tests that compare MOPAC output files between reference output files and the output files presently being generated by MOPAC. The comparisons are implemented in Python (
tests/compare_output.py
), orchestrated by CTest (tests/CMakeLists.txt
), and monitored for coverage by CodeCov, as indicated by the coverage badge on the repository's main README file. Right now, the coverage is relatively low (68% at the time of writing this), and it should probably be in the high 80's or low 90's.The submission of new tests that expand testing coverage is very welcome, particularly with an emphasis on any common use cases that are not presently represented in the existing tests. The tests in the main
tests
directory are a combination of historical regression tests used by Jimmy Stewart and validation tests of the INDO feature used by Rebecca Gieseking, and thekeywords
subdirectory contains keyword-specific tests I've written in the process of migrating keyword documentation between MOPAC's old website and the new website that is still in development.Testing is particularly important for MOPAC because of its dual role as a popular, widely-used, and venerable electronic structure package and as the long-time primary codebase for the computational chemistry research work of Jimmy Stewart. In its first role, the core functionality of MOPAC has been heavily tested in practice by its large user base. However, there are many features and options in MOPAC produced over the course of Dr. Stewart's scientific career and many conceptually valid combinations of these features and options, not all of which see regular use. Some of these features may not be as robust as the core functionality, and if problems are encountered, it will be either by developers during the process of expanding test coverage or by users who are trying to apply MOPAC in uncommon and novel ways. The former scenario is much preferred to the latter.
The text was updated successfully, but these errors were encountered: