dmt_transform behaving strangely (?) in most recent version

[sophiarubens]

This problem sounds far-fetched, but I and O(10) classmates in PHYS 641: Observational Techniques of Modern Astrophysics at McGill University are using sigpyproc to facilitate a multi-part project involving analysis of CHIME/pulsar data. Some of us updated certain environments and needed to reinstall sigpyproc between parts of the project, only to find that an imshow of dmt_transform (for the same preprocessed data and function call) exhibited a lot of unphysical structure when using this most recently pulled version of the package, while things looked as anticipated using a slightly older version of the package. We were all eventually able to recover the intended behaviour by installing an older commit (somewhat strangely, we had to go to about two commits ago), https://github.com/FRBs/sigpyproc3@f1e14f66b7d81c5835cc609789c8028c9d40f7aa . After reading the documentation, we were all able to use sigpyproc as intended, but it seems like dmt_transform (and this is pure speculation) may have been accidentally corrupted by some unrelated edits made during the pulls to main sometime between Jan-Feb 2025. We noticed this behaviour at the end of February, but I only remembered to post this note now.