Mark unit tests with @pytest.mark.benchmark part 2 #2924
Conversation
weiji14
added
maintenance
Specifically test_virtual_file, test_virtualfile_from_data_required_z_matrix, test_virtualfile_from_vectors, and test_virtualfile_from_matrix
CodSpeed Performance ReportMerging #2924 will not alter performanceFalling back to comparing Summary
Benchmarks breakdown
|
| @pytest.mark.benchmark | ||
| def test_load_remote_dataset_benchmark_with_region(): | ||
| """ | ||
| Benchmark loading a remote dataset with 'region'. | ||
| """ | ||
| data = load_remote_dataset_wrapper(resolution="01d", region=[-10, 10, -5, 5]) | ||
| assert data.name == "seafloor_age" | ||
| assert data.attrs["long_name"] == "age of seafloor crust" | ||
| assert data.attrs["cpt"] == "@earth_age.cpt" | ||
| assert data.attrs["units"] == "Myr" | ||
| assert data.attrs["horizontal_datum"] == "WGS84" | ||
| assert data.gmt.registration == 0 | ||
| assert data.shape == (11, 21) |
There was a problem hiding this comment.
Instead of benchmarking all the various load_earth_* functions, I just made created a new one that tests the _load_remote_dataset function, and it should cover the refactoring happening at #2917. Note that this test is somewhat duplicating test_earth_age_01d_with_region, but should be ok I hope.
pygmt/tests/test_figure.py
Outdated
| @pytest.mark.benchmark | ||
| def test_figure_repr(): | ||
| """ | ||
| Make sure that figure output's PNG and HTML printable representations look |
There was a problem hiding this comment.
Was a little unsure which test to benchmark for figure.py, so I picked this one that tests __repr_png__, since it goes through the savefig and _preview calls.
| @pytest.mark.benchmark | ||
| def test_earth_relief_holes(): | ||
| """ | ||
| Check that the @earth_relief_20m_holes.grd dataset loads without errors. |
There was a problem hiding this comment.
Only benchmarking test_earth_relief_holes since it calls both pygmt.which and pygmt.io.load_dataarray. The other sample datasets call pygmt.which and pd.read_csv only.
Description of proposed changes
Follow-up to #2908, this is where we mark the unit tests for any PyGMT functions we want to benchmark with the @pytest.mark.benchmark decorator.
Benchmark report at https://codspeed.io/GenericMappingTools/pygmt/branches/pytest/mark-benchmark-part2
See #2924 (comment) below on which new tests have been marked for benchmarking. Will prioritize PyGMT functions that use temporary files, and low-level clib functions that should be benchmarked. General discussion on which tests to mark can go in #2910.
Part 2 of addressing #2910. Covers test functions from test_binstats to test_filter1d according to the list in #2910 (comment).
Reminders
make formatandmake checkto make sure the code follows the style guide.doc/api/index.rst.Slash Commands
You can write slash commands (
/command) in the first line of a comment to performspecific operations. Supported slash commands are:
/format: automatically format and lint the code/test-gmt-dev: run full tests on the latest GMT development version