fix: correct DT5 patch applyto from 'T' to 'DT' in PATCHES.xml

[nightcityblade]

Fixes #456

The DT5 patch in pdb2pqr/dat/PATCHES.xml had <applyto>T</applyto> instead of <applyto>DT</applyto>. This is inconsistent with all other deoxynucleotide patches (DA3/DA5 → DA, DC3/DC5 → DC, DG3/DG5 → DG, DT3 → DT) and causes a fatal error when processing PDB files containing DT5 residues.

[sobolevnrm]

This was probably implemented because T is not found in RNA. However, I agree that it would be good to fix. Can you please add a test that demonstrates the problem and then shows that your patch fixes the problem?

[nightcityblade]

Good catch. I added a focused regression test that loads the packaged definitions, verifies DT5 is built from the DT reference residue, and checks the expected terminal phosphate atoms are removed while H5T is present. I also ran pytest -q tests/patches_test.py locally and it passes.

[sobolevnrm]

@nightcityblade the tests are failing for Python 3.10. I've lost track of the various PRs but merged #453 which might help. Can you please try merging that into your branch and see if it addresses the failures?

[nightcityblade]

I merged the latest upstream main changes from #453 into this branch and pushed the result in 44ca79d so the Python 3.10 failures can be re-evaluated against the updated base. I was only able to do a lightweight local check here because the environment is missing project test dependencies (for example propka), but the branch is now rebased onto the newer upstream state you pointed me to.