delayed imports reduce dependencies

.github/workflows/test_and_publish.yml

@@ -42,6 +42,8 @@ jobs:
         run: |
           mamba install -y -q --file extra_requirements.txt python=${{ matrix.python-version }} -c acellera -c conda-forge
           mamba install -y -q rdkit=2023.03.3 -c conda-forge
+          # scikit-learn, joblib, biopython are only used for rarely-used functions in smallmol
+          mamba install -y -q scikit-learn joblib biopython -c conda-forge
 
       - name: Install moleculekit
         run: |

extra_requirements.txt

@@ -1,11 +1,7 @@
 tqdm
-ipython
 matplotlib-base
-joblib
-scikit-learn
 rdkit>=2020.09.5
 mdtraj
 pdb2pqr>=3.5.2+18
 propka
-openbabel>=3.1.1
-biopython
+openbabel>=3.1.1

moleculekit/smallmol/tools/clustering.py

@@ -9,7 +9,6 @@
 import logging
 from moleculekit.smallmol.util import convertToString
 from tqdm import tqdm
-from joblib import Parallel, delayed  # Import delayed as well for other modules
 
 
 logger = logging.getLogger(__name__)
@@ -19,6 +18,7 @@ def ParallelExecutor(**joblib_args):
     """
     A wrapper for joblib.Parallel to allow custom progress bars.
     """
+    from joblib import Parallel
 
     def aprun(**tq_args):
         tqdm_f = lambda x, args: tqdm(x, **args)
@@ -91,7 +91,7 @@ def cluster(
     smallmol_list, method, distThresholds=0.2, returnDetails=True, removeHs=True
 ):
     """
-    Rreturn the SmallMol objects grouped in the cluster. It can also return the details of the clusters computed.
+    Return the SmallMol objects grouped in the cluster. It can also return the details of the clusters computed.
 
     Parameters
     ----------
@@ -117,8 +117,18 @@ def cluster(
     details: list
         A list with all the cluster details
     """
-
-    from sklearn.cluster import DBSCAN
+    try:
+        from sklearn.cluster import DBSCAN
+    except ImportError:
+        raise ImportError(
+            "Please install scikit-learn to use the clustering methods. You can install it with `conda install scikit-learn`"
+        )
+    try:
+        from joblib import delayed
+    except ImportError:
+        raise ImportError(
+            "Please install joblib to use the clustering methods. You can install it with `conda install joblib`"
+        )
 
     import sys
 
@@ -136,9 +146,7 @@ def cluster(
 
     if method not in _methods:
         raise ValueError(
-            "The method provided {} does not exists. The ones available are the following: {}".format(
-                method, _methods
-            )
+            f"The method provided '{method}' does not exist. The ones available are the following: {_methods}"
         )
 
     smallmol_list = np.array([sm.copy() for sm in smallmol_list])
@@ -202,6 +210,7 @@ def _maccsClustering(rdkit_mols):
         The numpy array containing the tanimoto matrix
     """
     from rdkit.Chem import MACCSkeys  # calcola MACCS keys
+    from joblib import delayed
 
     fps = []
     for m in tqdm(rdkit_mols):
@@ -230,6 +239,7 @@ def _pathFingerprintsClustering(rdkit_mols):
         The numpy array containing the tanimoto matrix
     """
     from rdkit.Chem.Fingerprints import FingerprintMols  # calcola path fingerprints
+    from joblib import delayed
 
     fps = []
     for m in tqdm(rdkit_mols):
@@ -258,6 +268,7 @@ def _atomsFingerprintsClustering(rdkit_mols):
         The numpy array containing the dice matrix
     """
     from rdkit.Chem.AtomPairs import Pairs  # Atom pairs
+    from joblib import delayed
 
     fps = []
     for m in tqdm(rdkit_mols):
@@ -286,6 +297,7 @@ def _torsionsFingerprintsClustering(rdkit_mols):
         The numpy array containing the dice matrix
     """
     from rdkit.Chem.AtomPairs import Torsions  # Topological Torsions
+    from joblib import delayed
 
     fps = []
     for m in tqdm(rdkit_mols):
@@ -318,6 +330,7 @@ def _circularFingerprintsClustering(rdkit_mols, radius=2):
         The numpy array containing the dice matrix
     """
     from rdkit.Chem import AllChem  # calcola circular fingerprints
+    from joblib import delayed
 
     fps = []
     for m in rdkit_mols:

moleculekit/smallmol/util.py

@@ -212,7 +212,6 @@ def getChemblSimilarLigandsBySmile(smi, threshold=85, returnSmiles=False):
 
 
 def convertToString(arr):
-
     if isinstance(arr, list):
         arr_str = " ".join([str(i) for i in arr])
     elif isinstance(arr, tuple):
@@ -256,7 +255,6 @@ def _depictMol(
     from os.path import splitext
     from rdkit.Chem import Kekulize
     from rdkit.Chem.Draw import rdMolDraw2D
-    from IPython.display import SVG
 
     if highlightAtoms is not None and not isinstance(highlightAtoms, list):
         raise ValueError(
@@ -319,6 +317,8 @@ def _depictMol(
 
     # activate jupiter-notebook rendering
     if ipython:
+        from IPython.display import SVG
+
         svg = svg.replace("svg:", "")
         return SVG(svg)
     else:
@@ -358,7 +358,6 @@ def depictMultipleMols(
 
     """
     from rdkit.Chem.Draw import MolsToGridImage
-    from IPython.display import SVG
     from os.path import splitext
 
     sel_atoms = []
@@ -382,9 +381,10 @@ def depictMultipleMols(
 
     from rdkit.Chem.Draw import IPythonConsole as CDIPythonConsole
 
-    if MolsToGridImage == CDIPythonConsole.ShowMols:
-        CDIPythonConsole.UninstallIPythonRenderer()
-        from rdkit.Chem.Draw import MolsToGridImage
+    if ipython:
+        if MolsToGridImage == CDIPythonConsole.ShowMols:
+            CDIPythonConsole.UninstallIPythonRenderer()
+            from rdkit.Chem.Draw import MolsToGridImage
 
     svg = MolsToGridImage(
         mols_list,
@@ -404,6 +404,8 @@ def depictMultipleMols(
         f.close()
 
     if ipython:
+        from IPython.display import SVG
+
         _svg = SVG(svg)
         return _svg
     else:

moleculekit/tools/atomtyper.py

@@ -453,7 +453,13 @@ def getFeatures(mol):
 
 def parallel(func, listobj, n_cpus=-1, *args):
     from tqdm import tqdm
-    from joblib import Parallel, delayed
+
+    try:
+        from joblib import Parallel, delayed
+    except ImportError:
+        raise ImportError(
+            "Please install joblib to use the parallel function with `conda install joblib`."
+        )
 
     results = Parallel(n_jobs=n_cpus)(delayed(func)(ob, *args) for ob in tqdm(listobj))
     return results
@@ -472,7 +478,9 @@ def test_preparation(self):
 
         assert mol_equal(mol2, ref, exceptFields=("coords",))
 
-    @unittest.skipIf(sys.platform.startswith("win"), "Windows OBabel fails at atom typing")
+    @unittest.skipIf(
+        sys.platform.startswith("win"), "Windows OBabel fails at atom typing"
+    )
     def test_obabel_atomtyping(self):
         from moleculekit.home import home
         from moleculekit.molecule import Molecule