CIF file support

[mihaitodor]

I'm not sure if I'm misusing this tool, but, at a quick glance, I believe it has support for CIF files. However, if I try to run it with https://www.ebi.ac.uk/pdbe/entry-files/download/4mvj.cif:

$ pdb2pqr --log-level=DEBUG 4mvj.cif 4mvj.pqr

It fails with the following error:

...
INFO:Loading molecule: 4mvj.cif
Traceback (most recent call last):
  File "/Users/todor/Projects/grisell/apbs/.venv/bin/pdb2pqr", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/todor/Projects/grisell/apbs/.venv/lib/python3.13/site-packages/pdb2pqr/main.py", line 841, in main
    if main_driver(args) == 1:
       ~~~~~~~~~~~^^^^^^
  File "/Users/todor/Projects/grisell/apbs/.venv/lib/python3.13/site-packages/pdb2pqr/main.py", line 777, in main_driver
    pdblist, is_cif = io.get_molecule(args.input_path)
                      ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
  File "/Users/todor/Projects/grisell/apbs/.venv/lib/python3.13/site-packages/pdb2pqr/io.py", line 463, in get_molecule
    pdblist, errlist = cif.read_cif(input_file)
                       ~~~~~~~~~~~~^^^^^^^^^^^^
  File "/Users/todor/Projects/grisell/apbs/.venv/lib/python3.13/site-packages/pdb2pqr/cif.py", line 1043, in read_cif
    src_pdb, src_err = source(block)
                       ~~~~~~^^^^^^^
  File "/Users/todor/Projects/grisell/apbs/.venv/lib/python3.13/site-packages/pdb2pqr/cif.py", line 582, in source
    "ORGANISM_COMMON: "
    ~~~~~~~~~~~~~~~~~~~
    + src_obj.get_value("gene_src_common_name", i)
    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
TypeError: can only concatenate str (not "NoneType") to str

I tried another file https://www.ebi.ac.uk/pdbe/entry-files/download/1hel.cif, but this one fails silently and produces an empty pqr file:

out.log

Running the same commands with PDB files works as expected.

[bamattsson]

First of all, thank you to Electrostatics for maintaining this open source repo. I looked for a while for code that can assign hydrogens to proteins reliably, most things are closed source or not very performant. This code seems to do a good job. Thanks!

I have also noticed the issue with cif files. I have tried with a few different cif files. It works for quite a few, but it fails for some. Here are some examples of cif files for which it fails: 1FKI.cif, 9X2O.cif, or any cif file generated with co-folding algorithms such as Boltz-2.

I have looked through the code and I believe the reasons why it fails falls into two categories:

1. Many functions that converts different cif objects into pdb lines lack a mechanism in case those cif objects don't exist. E.g. here. If database_PDB_matrix doesn't exist this function will return None, which will fail a few rows below when we try to do or_obj.get_value("origx[1][1]", 0)). This is the type of issue that @mihaitodor reported above, and there are a few closed issues that report this problem as well with different parts of the code.
2. The code that converts cif x/y/z-coordinates to pdb ones doesn't take into account that cif coordinates can have more than 3 significant digits whereas pdb files need to have exactly 3 significant digits (e.g. here). This sometimes leads to too many characters, and hence a too long and missaligned pdb string which makes the cif reading fail.

I will raise a PR that aims to fix these errors :) But I am posting the issues here for visibility.