fixed alphanumeric character mismatches in parse_observation_longcode with latest fragalysis uploads

[BenEmeryCMD]

This PR fixes parsing of newer Fragalysis observation longcodes in Fragalysis.py.

Some newer fragalysis uploads can contain alphanumeric characters where the existing parser expects only numeric characters.

Example new upload format in target lb36049-14 (NXT1-NXF1) :

observation = x5064b
longcode = NXT1A-x5064_B_204_B_v1

vs. old format

observation = x5064c
longcode = NXT1A-x5064_B_205_0_v1

[cvallee]

Hi @BenEmeryCMD, I've tested your changes locally. When running add_hits with an xchem-hippo from pip install I get the following error:

UnsupportedFragalysisLongcodeError: NXT1A-x5155_B_303_0_1_NXT1A-x5194+B+303+C+1__LIG

Then, when running add_hits from your branch, it goes past the error, which means your fix worked. However I get a SanitisationError:

│   155 │   if mol:                                                                                │
│   156 │   │   smiles = MolToSmiles(mol, True)                                                    │
│   157 │   elif sanitisation_failed == "error":                                                   │
│ ❱ 158 │   │   raise SanitisationError                                                            │
│   159 │   elif sanitisation_failed == "warning":                                                 │
│   160 │   │   mrich.warning(f"sanitisation failed for {smiles=}")                                │
│   161                                                                                            │
│                                                                                                  │
│ ╭──────────────────────────── locals ─────────────────────────────╮                              │
│ │                 key = '[STB]'                                   │                              │
│ │                 mol = None                                      │                              │
│ │         orig_smiles = 'O=C([O-])C1CCN([S@TB2](=O)(=O)C2CC2)CC1' │                              │
│ │             radical = 'error'                                   │                              │
│ │                   s = 'O=C([O-])C1CCN([S@TB2](=O)(=O)C2CC2)CC1' │                              │
│ │ sanitisation_failed = 'error'                                   │                              │
│ │              smiles = 'O=C([O-])C1CCN([STB2](=O)(=O)C2CC2)CC1'  │                              │
│ │       stereo_smiles = 'O=C([O-])C1CCN([S@TB2](=O)(=O)C2CC2)CC1' │                              │
│ │           verbosity = False                                     │                              │
│ ╰─────────────────────────────────────────────────────────────────╯                              │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
SanitisationError

The issue seems to come from TB2 in the SMILES of a fragment with a a SO2 group