Maximilian Scheurer

Some of the test failures are of course related to "more iterations" in the Davidson procedure... The mismatches wrt properties/energies are harder to fix probably... 😦

Should be updated against current `master`, then the pipeline will run again.

@apapapostolou the data are uploaded, you need to change the download URL to `0.6.0` and get the hopefully correct SHA256SUMS file here: https://q-chem.de/adcc_testdata/0.6.0/SHA256SUMS

I think you need to rebase against `master` before being able to merge, @apapapostolou.

Thanks for the heads-up! Will be fixed via #44 hopefully soon 👍🏻

I might be wrong about this, but since you're using density fitting, you probably need to set `hess.auxbasis_response=2` to be able to compare with your numerical Hessian.

> The same structure Hessian calculation with B3LYP; no negative frequencies are observed. If the structure is not optimized at the level of theory you're using for the Hessian, you...

I tried to reproduce the error, and it seems to be related to NumPy. I get the same error in a conda environment with `numpy>=2.0`. Could you try to downgrade...