Scanpy version 1.11.2 changed sc.pp.normalize_total()

Scanpy changed normalize_total() to work with Numba, with the release 1.11.2 (https://github.com/scverse/scanpy/pull/3571, https://github.com/scverse/scanpy/commit/e6e86f2bd4d97340557848a24275c08bdb20517e). This results in the loss of values when adata.X is really small. For our testdata adata this will result in the loss of values failing several unit tests such as test_embedding().

Solutions: Change the target_sum parameter of sc.pp.normalize_total() from None to a float value maybe even something expected for a full dataset as 1e3.