Baunsgaard commented on PR #2103: URL: https://github.com/apache/systemds/pull/2103#issuecomment-2352797328
> Thank you for continuously providing valuable feedback. > > As I'm encountering code coverage for the first time, I have some questions. I've summarized the current situation and requests below as I understand them. > > Could you please confirm if this is correct? > > * Situation Summary: > > * Java Test Rejection: > > * Attempted to compare the Java test of the roll function with NumPy, but it was rejected because we cannot install Python packages in the Docker image. > * Limitations of the Python Test: > > * Verified correct implementation on dense and sparse matrices through Python tests comparing with NumPy. > * However, since it uses the statically built Python API, it is not included in code coverage via Codecov. > * Code Coverage Issue: > > * To improve code coverage, a Java-based test is necessary. > > * Request: > > * Add Java Test: > > * Need to write a Java test for the roll function without using NumPy. > * The test should generate sparse and dense matrices, apply the roll function to each, and verify that the results are identical. > * The test code should be located at `src/test/java/org/apache/sysds/test/component/matrix/libMatrixReorg/RollTest.java` (or `src/test/java/org/apache/sysds/test/functions/reorg/FullRollTest.java`). Yes this sounds correct. Therefore: 1. keep the python tests that you made, because they ensure compatibility with numpy, 2. make internal tests (component tests) that, as you say, verify the roll function without using numpy and verify the results correctly. Thanks. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: dev-unsubscr...@systemds.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
