Add ELL view tests

[KyleVaughn]

Description:

Added by @S-A-Almohri

This issue involves adding a suite of tests for the ELL view within the sparse matrix library to ensure its correctness and performance. The tests should verify that the ELL format handles various scenarios correctly, including edge cases, and performs as expected. An additional check with Python might be implemented using SciPy functions to validate results.

Tasks:

Added by @S-A-Almohri

Review Existing ELL Implementation:

• Study peer's implementation of the ELL view to identify key features and possible edge cases.
• Understand the existing methods and their expected behavior.

Define Test Cases:

Basic functionality: Verify that simple, small matrices are correctly represented in ELL format. Edge cases: Empty matrices, matrices with a single element, symmetric matrices, and very sparse matrices. Performance: Measure the performance of SpMV using ELL format for larger matrices.

Implement Unit Tests:

• Write unit tests for each identified scenario.
• Ensure tests cover methods for viewing.
• Basic Tests: Create matrices in ELL format and verify that all values and indices are correctly stored.
• Edge Case Tests: Test for empty matrices, single-element matrices, and symmetirc matrices.
• Performance Tests: Measure and verify the performance of SpMV operations against expected benchmarks from scipy.

Run and Validate Tests:

• Execute the tests and validate the results.
• Raise any issues (if any)

Definition of done:

Added by @S-A-Almohri

• A comprehensive suite of unit tests for the ELL view is implemented.
• Tests cover all identified scenarios, including basic functionality, edge cases, and performance.
• All tests are executed and pass successfully (if the implementer did their work properly)

[NiloyGupta-UM]

I don't think performance tests fall under the scope of this issue; all these tests need to do is check that the implemented methods work and produce the expected behavior. Furthermore, assuming you are referring to sparse matrix-vector multiplication operations when writing "SpMV", that falls under another issue and should not be part of this one.

Also, specifically because ELL requires padding, making sure that the padded values are properly printed/written would be a good test. What "properly" means in this case might be defined in another issue concerning the ELL format or may need to be discussed by the group.

I think a good edge test case to add would be a matrix with a row of all zeros. Furthermore, a test should be added to check if matrix values accidentally get modified during the execution of the viewing methods. Otherwise, the cases you mentioned are a good set. It might also be beneficial to add a requirement that explicitly states that the library's testing framework should be utilized, but, depending on the group's opinion, this may already be an implicit one.

[S-A-Almohri]

Thanks for the thorough review. Sorry for not being clear, when I say SpMV I refer to the script, and by performance check against SciPy, I mean checking for time it takes to print and view each. The second bullet point just had that extra part, I changed it.