Windows + RDKit: BaseFeatures.fdef and .sdf reading errors

[jacrossley]

Hi,

I'm running into some issues on Windows after following the installation instructions. See the output of pytest below:

`(automartiniM3) C:\Users\Joel\Automartini_M3>pytest -v
================================================= test session starts =================================================
platform win32 -- Python 3.8.20, pytest-8.3.4, pluggy-1.5.0 -- C:\Users\Joel\anaconda3\envs\automartiniM3\python.exe
cachedir: .pytest_cache
rootdir: C:\Users\Joel\Automartini_M3
configfile: pyproject.toml
testpaths: auto_martiniM3/tests
collected 0 items / 2 errors

======================================================= ERRORS ========================================================
_________________________________ ERROR collecting auto_martiniM3/tests/test_basic.py _________________________________
auto_martiniM3\tests\test_basic.py:10: in <module>
    import auto_martiniM3
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\__init__.py:30: in <module>
    from . import solver, topology
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\solver.py:30: in <module>
    from . import optimization, output, topology
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\optimization.py:33: in <module>
    from . import topology # AutoM3 change
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\topology.py:40: in <module>
    factory = ChemicalFeatures.BuildFeatureFactory(fdefName)
E   OSError: File: C:/Users/Joel/anaconda3/envs/automartiniM3/Library\share/RDKit\Data\BaseFeatures.fdef could not be opened.
________________________________ ERROR collecting auto_martiniM3/tests/test_version.py ________________________________
auto_martiniM3\tests\test_version.py:5: in <module>
    import auto_martiniM3
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\__init__.py:30: in <module>
    from . import solver, topology
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\solver.py:30: in <module>
    from . import optimization, output, topology
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\optimization.py:33: in <module>
    from . import topology # AutoM3 change
..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\topology.py:40: in <module>
    factory = ChemicalFeatures.BuildFeatureFactory(fdefName)
E   OSError: File: C:/Users/Joel/anaconda3/envs/automartiniM3/Library\share/RDKit\Data\BaseFeatures.fdef could not be opened.
=============================================== short test summary info ===============================================
ERROR auto_martiniM3/tests/test_basic.py - OSError: File: C:/Users/Joel/anaconda3/envs/automartiniM3/Library\share/RDKit\Data\BaseFeatures.fdef could not be o...
ERROR auto_martiniM3/tests/test_version.py - OSError: File: C:/Users/Joel/anaconda3/envs/automartiniM3/Library\share/RDKit\Data\BaseFeatures.fdef could not be o...
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: 2 errors during collection !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
================================================== 2 errors in 0.63s ==================================================`

After some googling it seems the BaseFeatures.fdef could not be opened might be specific to the Windows conda-forge build of RDKit (see rdkit/rdkit#7803). I downgraded from 2024.03.5 to 2024.03.4 and managed to fix the issue, but now i get the following:

(automartiniM3) C:\Users\Joel\Automartini_M3>pytest -v auto_martiniM3\tests
============================================================================= test session starts =============================================================================
platform win32 -- Python 3.8.20, pytest-8.3.4, pluggy-1.5.0 -- C:\Users\Joel\anaconda3\envs\automartiniM3\python.exe
cachedir: .pytest_cache
rootdir: C:\Users\Joel\Automartini_M3
configfile: pyproject.toml
collected 7 items

auto_martiniM3/tests/test_basic.py::test_auto_martini_imported PASSED                                                                                                    [ 14%]
auto_martiniM3/tests/test_basic.py::test_connection_to_ALOGPS[CC(=O)OC1=CC=CC=C1C(=O)O] PASSED                                                                           [ 28%]
auto_martiniM3/tests/test_basic.py::test_auto_martini_run_smiles[CC(=O)OC1=CC=CC=C1C(=O)O-ASP-5] PASSED                                                                  [ 42%]
auto_martiniM3/tests/test_basic.py::test_auto_martini_run_smiles[CCC-PRO-1] PASSED                                                                                       [ 57%]
auto_martiniM3/tests/test_basic.py::test_auto_martini_run_sdf[sdf_file0-BENZ-3] FAILED                                                                                   [ 71%]
auto_martiniM3/tests/test_basic.py::test_auto_martini_run_sdf[sdf_file1-IBUP-6] FAILED                                                                                   [ 85%]
auto_martiniM3/tests/test_version.py::test_auto_martini_version PASSED                                                                                                   [100%]

================================================================================== FAILURES ===================================================================================
_________________________________________________________________ test_auto_martini_run_sdf[sdf_file0-BENZ-3] _________________________________________________________________

sdf_file = WindowsPath('tests/files/benzene.sdf'), name = 'BENZ', num_beads = 3

    @pytest.mark.parametrize(
        "sdf_file,name,num_beads",
        [
            (dpath / "benzene.sdf", "BENZ", 3),
            (dpath / "ibuprofen.sdf", "IBUP", 6)
        ],
    )

    def test_auto_martini_run_sdf(sdf_file: str, name:str, num_beads: int):
>       mol = auto_martiniM3.topology.gen_molecule_sdf(str(sdf_file))

auto_martiniM3\tests\test_basic.py:55:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

sdf = 'tests\\files\\benzene.sdf'

    def gen_molecule_sdf(sdf):
        """Generate mol object from SD file"""
        logger.debug("Entering gen_molecule_sdf()")
>       suppl = Chem.SDMolSupplier(sdf)
E       OSError: File error: Bad input file tests\files\benzene.sdf

..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\topology.py:120: OSError
_________________________________________________________________ test_auto_martini_run_sdf[sdf_file1-IBUP-6] _________________________________________________________________

sdf_file = WindowsPath('tests/files/ibuprofen.sdf'), name = 'IBUP', num_beads = 6

    @pytest.mark.parametrize(
        "sdf_file,name,num_beads",
        [
            (dpath / "benzene.sdf", "BENZ", 3),
            (dpath / "ibuprofen.sdf", "IBUP", 6)
        ],
    )

    def test_auto_martini_run_sdf(sdf_file: str, name:str, num_beads: int):
>       mol = auto_martiniM3.topology.gen_molecule_sdf(str(sdf_file))

auto_martiniM3\tests\test_basic.py:55:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

sdf = 'tests\\files\\ibuprofen.sdf'

    def gen_molecule_sdf(sdf):
        """Generate mol object from SD file"""
        logger.debug("Entering gen_molecule_sdf()")
>       suppl = Chem.SDMolSupplier(sdf)
E       OSError: File error: Bad input file tests\files\ibuprofen.sdf

..\anaconda3\envs\automartiniM3\lib\site-packages\auto_martiniM3\topology.py:120: OSError
=========================================================================== short test summary info ===========================================================================
FAILED auto_martiniM3/tests/test_basic.py::test_auto_martini_run_sdf[sdf_file0-BENZ-3] - OSError: File error: Bad input file tests\files\benzene.sdf
FAILED auto_martiniM3/tests/test_basic.py::test_auto_martini_run_sdf[sdf_file1-IBUP-6] - OSError: File error: Bad input file tests\files\ibuprofen.sdf
========================================================================= 2 failed, 5 passed in 2.90s =========================================================================

Which i think is a result of RDKit now not being able to read the .sdf files. Im not familiar with the package, so I thought I'd open this issue before proceeding to see what you would advise and ask if you've seen similar issues.

Many thanks in advance,
Joel