Remove molvs dependency

atomsci/ddm/test/unit/test_struct_utils.py

@@ -20,13 +20,13 @@ def test_get_rdkit_smiles():
 
 def test_rdkit_smiles_from_smiles():
     result = su.rdkit_smiles_from_smiles(test_smiles, useCanonicalTautomers=True)
-    assert result == ['Brc1cc(O[C@@H]2C[C@H]3CC[C@@H](C2)N3)cc(-c2ccccc2)c1', 
-        'Brc1ccc(N2CCN(CCCCOc3ccc4ccccc4c3)CC2)cc1', 
+    assert result == ['Brc1cc(O[C@@H]2C[C@H]3CC[C@@H](C2)N3)cc(-c2ccccc2)c1',
+        'Brc1ccc(N2CCN(CCCCOc3ccc4ccccc4c3)CC2)cc1',
         'Brc1ccc(N2CCN(CCCN3CCC(Cc4ccccc4)CC3)CC2)cc1']
 
     result = su.rdkit_smiles_from_smiles(test_smiles, useCanonicalTautomers=False)
     assert result == ['Brc1cc(O[C@@H]2C[C@H]3CC[C@@H](C2)N3)cc(-c2ccccc2)c1',
-        'Brc1ccc(N2CCN(CCCCOc3ccc4ccccc4c3)CC2)cc1', 
+        'Brc1ccc(N2CCN(CCCCOc3ccc4ccccc4c3)CC2)cc1',
         'Brc1ccc(N2CCN(CCCN3CCC(Cc4ccccc4)CC3)CC2)cc1']
 
     result = su.rdkit_smiles_from_smiles(test_smiles, useIsomericSmiles=False)
@@ -157,7 +157,7 @@ def test_mol_wt_from_smiles():
 
 def test_canonical_tautomers_from_smiles():
     canonical_tautomers = [su.canonical_tautomers_from_smiles(s) for s in ['asdf']+test_smiles]
-    assert canonical_tautomers==[[''], 
+    assert canonical_tautomers==[[''],
         ['Brc1cc(O[C@@H]2C[C@H]3CC[C@@H](C2)N3)cc(-c2ccccc2)c1'],
         ['Brc1ccc(N2CCN(CCCCOc3ccc4ccccc4c3)CC2)cc1'],
         ['Brc1ccc(N2CCN(CCCN3CCC(Cc4ccccc4)CC3)CC2)cc1']]
@@ -177,4 +177,4 @@ def test_canonical_tautomers_from_smiles():
     test_draw_structure()
     test_smiles_to_inchi_key()
     test_mol_wt_from_smiles()
-    test_canonical_tautomers_from_smiles()
+    test_canonical_tautomers_from_smiles()

atomsci/ddm/utils/struct_utils.py

@@ -7,15 +7,13 @@
 
 import re
 import numpy as np
-import molvs
 import logging
 
 from rdkit import Chem
 from rdkit.Chem import AllChem, Draw, Descriptors
 from rdkit.Chem.MolStandardize import rdMolStandardize
 
-stdizer = molvs.standardize.Standardizer(prefer_organic=True)
-uncharger = molvs.charge.Uncharger()
+uncharger = rdMolStandardize.Uncharger()
 
 
 def get_rdkit_smiles(orig_smiles, useIsomericSmiles=True):
@@ -34,13 +32,12 @@ def get_rdkit_smiles(orig_smiles, useIsomericSmiles=True):
     mol = Chem.MolFromSmiles(orig_smiles)
     if mol is None:
         return ""
-    else:
-        return Chem.MolToSmiles(mol, isomericSmiles=useIsomericSmiles)
+    return Chem.MolToSmiles(mol, isomericSmiles=useIsomericSmiles)
 
 
 def rdkit_smiles_from_smiles(orig_smiles, useIsomericSmiles=True, useCanonicalTautomers=False, workers=1):
     """Parallel version of get_rdkit_smiles. If orig_smiles is a list and workers is > 1, spawn 'workers'
-    threads to convert input SMILES strings to standardized RDKit format.
+    processes to convert input SMILES strings to standardized RDKit format.
 
     Args:
         orig_smiles (list or str): List of SMILES strings to canonicalize.
@@ -59,7 +56,7 @@ def rdkit_smiles_from_smiles(orig_smiles, useIsomericSmiles=True, useCanonicalTa
 
     if isinstance(orig_smiles, list):
         from functools import partial
-        func = partial(rdkit_smiles_from_smiles, useIsomericSmiles=useIsomericSmiles, 
+        func = partial(rdkit_smiles_from_smiles, useIsomericSmiles=useIsomericSmiles,
                        useCanonicalTautomers=useCanonicalTautomers)
         if workers > 1:
             from multiprocessing import pool
@@ -85,7 +82,7 @@ def rdkit_smiles_from_smiles(orig_smiles, useIsomericSmiles=True, useCanonicalTa
 
 def mols_from_smiles(orig_smiles, workers=1):
     """Parallel function to create RDKit Mol objects for a list of SMILES strings. If orig_smiles is a list
-    and workers is > 1, spawn 'workers' threads to convert input SMILES strings to Mol objects.
+    and workers is > 1, spawn 'workers' processes to convert input SMILES strings to Mol objects.
 
     Args:
         orig_smiles (list or str): List of SMILES strings to convert to Mol objects.
@@ -115,7 +112,7 @@ def mols_from_smiles(orig_smiles, workers=1):
     return mols
 
 
-def base_smiles_from_smiles(orig_smiles, useIsomericSmiles=True, removeCharges=False, 
+def base_smiles_from_smiles(orig_smiles, useIsomericSmiles=True, removeCharges=False,
                             useCanonicalTautomers=False, workers=1):
     """Generate standardized SMILES strings for the largest fragments of each molecule specified by
     orig_smiles. Strips salt groups and replaces any rare isotopes with the most common ones for each element.
@@ -225,9 +222,14 @@ def base_mol_from_smiles(orig_smiles, useIsomericSmiles=True, removeCharges=Fals
     cmpd_mol = Chem.MolFromSmiles(orig_smiles)
     if cmpd_mol is None:
         return None
-    std_mol = stdizer.isotope_parent(stdizer.fragment_parent(cmpd_mol), skip_standardize=True)
-    if removeCharges:
-        std_mol = uncharger(std_mol)
+    try:
+        std_mol = rdMolStandardize.IsotopeParent(
+            rdMolStandardize.FragmentParent(cmpd_mol),
+            skipStandardize=True)
+        if removeCharges:
+            std_mol = uncharger.uncharge(std_mol)
+    except Exception:
+        std_mol = None
     return std_mol
 
 
@@ -294,12 +296,17 @@ def base_mol_from_inchi(inchi_str, useIsomericSmiles=True, removeCharges=False):
     cmpd_mol = Chem.inchi.MolFromInchi(inchi_str)
     if cmpd_mol is None:
         return None
-    std_mol = stdizer.isotope_parent(stdizer.fragment_parent(cmpd_mol), skip_standardize=True)
-    if removeCharges:
-        std_mol = uncharger(std_mol)
+    try:
+        std_mol = rdMolStandardize.IsotopeParent(
+            rdMolStandardize.FragmentParent(cmpd_mol),
+            skipStandardize=True)
+        if removeCharges:
+            std_mol = uncharger.uncharge(std_mol)
+    except Exception as e:
+        print(f"Error standardizing InChI {inchi_str}: {e}")
+        std_mol = None
     return std_mol
 
-
 def draw_structure(smiles_str, image_path, image_size=500):
     """Draw structure for the compound with the given SMILES string as a PNG file.
 
@@ -418,7 +425,7 @@ def _merge_dataframes_by_smiles(dataframes, smiles_col='rdkit_smiles', id_col='c
                 new_df = new_df.drop([lCol, rCol], axis=1)
         left_df = new_df
 
-    return new_df
+    return left_df
 
 
 def smiles_to_inchi_key(smiles):
@@ -514,6 +521,4 @@ def canonical_tautomers_from_smiles(smiles):
         smiles = [smiles]
     mols = [Chem.MolFromSmiles(smi) for smi in smiles]
     canon_tautomers = [taut_enum.Canonicalize(m) if m is not None else None for m in mols]
-    return [Chem.MolToSmiles(m) if m is not None else '' for m in canon_tautomers]
-
-
+    return [Chem.MolToSmiles(m) if m is not None else '' for m in canon_tautomers]

install-ampl/pyproject.toml

@@ -1,40 +1,21 @@
 
 [project]
-name = "ampl"
-version = "1.0"
+name = "atomsci-ampl"
+description = "ATOM Modeling PipeLine"
+version = "1.8.0"
+requires-python = ">=3.9,<3.12"
 
 # NOTE:
 # * deepchem only works with python <= 3.10 (and possibly 3.11)
 #   - seems to install in python 3.11 but website says suports only 
 #     python <= 3.10
-#   - does not work with python 3.12+ due to tensorflow-addons dependency
+#   - does NOT work with python 3.12+ due to tensorflow-addons dependency
 #     which is no longer supported and only has support up to python 3.11
 # * ampl only works with python 3.9, 3.10, 3.11 (due to deepchem)
-#   - encoders / decoders need python 3.9
-#   - everthing else works with python 3.9 and 3.10 (and possibly 3.11)
-# * this will install on python 3.12 but won't have deepchem
-requires-python = ">=3.9,<3.13"
 
 # Dependencies for all Python versions
 dependencies = [
-    "setuptools",
-    "pip",
-
-    # core dependencies
-    "torch_cluster",
-    "torch_scatter",
-    "torch_sparse",
-    "torch_spline_conv",
     "torch-geometric",
-
-    # deepchem 2.8.0 deps [NOTE: may not work on python >= 3.11]
-    "deepchem[torch,tensorflow]==2.8.0; python_version >= '3.10' and python_version < '3.12'",
-
-    # deepchem[tensorflow] deps
-    "tensorflow-cpu==2.14.0; python_version >= '3.10' and python_version < '3.12'",
-
-    # atomsci-ampl deps [need to be added to atomsci-ampl]
-    "molvs",
     "xgboost",
     "umap-learn",
     "imbalanced-learn",
@@ -43,28 +24,41 @@ dependencies = [
     "pytest",
     "seaborn",
     "matplotcheck",
-
-    # code repos
-    "atomsci-ampl; python_version >= '3.10' and python_version < '3.12'",
+    "numpy<2",
     "atomsci-clients",
 
-    # rdkit
-    "rdkit>=2024.3.5; python_version >= '3.10'",
+    # deepchem 2.7.1 for Python 3.9 version
+    "deepchem[torch,tensorflow]==2.7.1; python_version < '3.10'",
+    # deepchem 2.8.0 deps [NOTE: may not work on python 3.11]
+    "deepchem[torch,tensorflow]==2.8.0; python_version >= '3.10'",
 
-    "numpy<2",
+    # deepchem[tensorflow] deps
+    "tensorflow-cpu==2.14.0",
 
-    # python specific versions [dgl / pyg dependent]
+    "rdkit>=2024.3.5; python_version >= '3.10'",
     "dgl==2.1.0; python_version >= '3.10'",
 
-    "pyg_lib==0.3.1; python_version >= '3.10' and python_version < '3.12'",
-    "pyg_lib==0.4.0; python_version >= '3.12'",
+    "pyg_lib==0.2.0; python_version < '3.10'",
+    "pyg_lib==0.3.1; python_version >= '3.10'",
 
-    "torch==2.1.2+cpu; python_version >= '3.10' and python_version < '3.12'",
-    "torch==2.3.1+cpu; python_version >= '3.12'"
+    "torch==2.0.1+cpu; python_version < '3.10'",
+    "torch==2.1.2+cpu; python_version >= '3.10'",
+    "torch_cluster",
+    "torch_scatter",
+    "torch_sparse",
+    "torch_spline_conv",
 ]
 
+[build-system]
+requires = ["setuptools>=61.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[tool.setuptools.packages.find]
+include = ["atomsci*"]
+namespaces = true
 
 [tool.uv]
+managed = true
 allow-insecure-host = [
     "download.pytorch.org",
     "data.dgl.ai",
@@ -82,9 +76,8 @@ url = "https://data.dgl.ai/wheels/repo.html"
 format = "flat"
 
 [[tool.uv.index]]
-# NOTE: pyg libraries for torch >= 2.2 seem to have GLIBC incompatibilities on LC
-name = "pyg_lib_py312"
-url = "https://data.pyg.org/whl/torch-2.3.1+cpu.html"
+name = "pyg_lib_py39"
+url = "https://data.pyg.org/whl/torch-2.0.1+cpu.html"
 format = "flat"
 
 [[tool.uv.index]]
@@ -94,34 +87,31 @@ format = "flat"
 
 [tool.uv.sources]
 # code repos
-atomsci-ampl = [
-    { git = "https://github.com/ATOMScience-org/AMPL.git", rev = "py310",  marker = "python_version >= '3.10'"}
-]
 atomsci-clients = { git = "ssh://git@czgitlab.llnl.gov:7999/atom/clients.git", rev = "master" }
 
 # dgl
 dgl = { index = "dgl" }
 
 # torch
 torch = { index = "pytorch" }
+
 pyg_lib = [
-    { index = "pyg_lib_py310", marker = "python_version >= '3.10' and python_version < '3.12'" },
-    { index = "pyg_lib_py312", marker = "python_version >= '3.12'" },
+    { index = "pyg_lib_py39", marker = "python_version < '3.10'" },
+    { index = "pyg_lib_py310", marker = "python_version >= '3.10'" },
 ]
 torch_cluster = [
-    { index = "pyg_lib_py310", marker = "python_version >= '3.10' and python_version < '3.12'" },
-    { index = "pyg_lib_py312", marker = "python_version >= '3.12'" },
+    { index = "pyg_lib_py39", marker = "python_version < '3.10'" },
+    { index = "pyg_lib_py310", marker = "python_version >= '3.10'" },
 ]
 torch_scatter = [
-    { index = "pyg_lib_py310", marker = "python_version >= '3.10' and python_version < '3.12'" },
-    { index = "pyg_lib_py312", marker = "python_version >= '3.12'" },
+    { index = "pyg_lib_py39", marker = "python_version < '3.10'" },
+    { index = "pyg_lib_py310", marker = "python_version >= '3.10'" },
 ]
 torch_sparse = [
-    { index = "pyg_lib_py310", marker = "python_version >= '3.10' and python_version < '3.12'" },
-    { index = "pyg_lib_py312", marker = "python_version >= '3.12'" },
+    { index = "pyg_lib_py39", marker = "python_version < '3.10'" },
+    { index = "pyg_lib_py310", marker = "python_version >= '3.10'" },
 ]
 torch_spline_conv = [
-    { index = "pyg_lib_py310", marker = "python_version >= '3.10' and python_version < '3.12'" },
-    { index = "pyg_lib_py312", marker = "python_version >= '3.12'" },
+    { index = "pyg_lib_py39", marker = "python_version < '3.10'" },
+    { index = "pyg_lib_py310", marker = "python_version >= '3.10'" },
 ]
-

pip/cpu_requirements.txt

@@ -93,7 +93,6 @@ maestrowf
 
 deepchem==2.8
 rdkit==2024.3.5
-MolVS
 mordred
 
 # ==========================

pip/cuda_requirements.txt

@@ -39,7 +39,6 @@ imblearn
 # pycairo
 
 maestrowf
-MolVS
 mordred
 
 pytest

pip/docker_requirements.txt

@@ -36,7 +36,6 @@ bravado
 # pycairo
 
 maestrowf
-MolVS
 mordred
 pytest
 ipykernel

pip/mchip_requirements.txt

@@ -38,7 +38,6 @@ imblearn
 # pycairo
 
 maestrowf
-MolVS
 mordred
 
 pytest

pip/readthedocs_requirements.txt

@@ -4,7 +4,6 @@ hyperopt
 IPython
 joblib
 matplotlib
-molvs
 numpy
 pandas
 pygments

pip/rocm_requirements.txt

@@ -27,7 +27,6 @@ imblearn
 
 
 maestrowf
-MolVS
 mordred
 
 pytest