pref: lazy import modules

dpdata/__init__.py

@@ -1,36 +1,12 @@
-# monty needs lzma
-# See https://github.com/pandas-dev/pandas/pull/27882
-try:
-    import lzma  # noqa: F401
-except ImportError:
-
-    class fakemodule:
-        pass
-
-    import sys
-
-    sys.modules["lzma"] = fakemodule
-
 from . import lammps, md, vasp
+from .bond_order_system import BondOrderSystem
 from .system import LabeledSystem, MultiSystems, System
 
 try:
     from ._version import version as __version__
 except ImportError:
     from .__about__ import __version__
 
-# BondOrder System has dependency on rdkit
-try:
-    # prevent conflict with dpdata.rdkit
-    import rdkit as _  # noqa: F401
-
-    USE_RDKIT = True
-except ModuleNotFoundError:
-    USE_RDKIT = False
-
-if USE_RDKIT:
-    from .bond_order_system import BondOrderSystem
-
 __all__ = [
     "__version__",
     "lammps",

dpdata/amber/md.py

@@ -2,7 +2,6 @@
 import re
 
 import numpy as np
-from scipy.io import netcdf_file
 
 from dpdata.amber.mask import pick_by_amber_mask
 from dpdata.unit import EnergyConversion
@@ -44,6 +43,8 @@ def read_amber_traj(
     labeled : bool
         Whether to return labeled data
     """
+    from scipy.io import netcdf_file
+
     flag_atom_type = False
     flag_atom_numb = False
     amber_types = []

dpdata/ase_calculator.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING, List, Optional
 
-from ase.calculators.calculator import (
+from ase.calculators.calculator import (  # noqa: TID253
     Calculator,
     PropertyNotImplementedError,
     all_changes,

dpdata/bond_order_system.py

@@ -3,7 +3,6 @@
 from copy import deepcopy
 
 import numpy as np
-from rdkit.Chem import Conformer
 
 import dpdata.rdkit.utils
 from dpdata.rdkit.sanitize import Sanitizer
@@ -102,6 +101,8 @@ def from_fmt_obj(self, fmtobj, file_name, **kwargs):
         return self
 
     def to_fmt_obj(self, fmtobj, *args, **kwargs):
+        from rdkit.Chem import Conformer
+
         self.rdkit_mol.RemoveAllConformers()
         for ii in range(self.get_nframes()):
             conf = Conformer()

dpdata/deepmd/hdf5.py

@@ -3,16 +3,15 @@
 from __future__ import annotations
 
 import warnings
+from typing import TYPE_CHECKING
 
-try:
-    import h5py
-except ImportError:
-    pass
 import numpy as np
-from wcmatch.glob import globfilter
 
 import dpdata
 
+if TYPE_CHECKING:
+    import h5py
+
 __all__ = ["to_system_data", "dump"]
 
 
@@ -35,6 +34,8 @@
     labels : bool
         labels
     """
+    from wcmatch.glob import globfilter
+
     g = f[folder] if folder else f
 
     data = {}

dpdata/gaussian/gjf.py

@@ -10,16 +10,7 @@
 from typing import List, Optional, Tuple, Union
 
 import numpy as np
-from scipy.sparse import csr_matrix
-from scipy.sparse.csgraph import connected_components
 
-try:
-    from openbabel import openbabel
-except ImportError:
-    try:
-        import openbabel
-    except ImportError:
-        openbabel = None
 from dpdata.periodic_table import Element
 
 
@@ -53,10 +44,13 @@
     ImportError
         if Open Babel is not installed
     """
-    if openbabel is None:
-        raise ImportError(
-            "Open Babel (Python interface) should be installed to detect fragmentation!"
-        )
+    from scipy.sparse import csr_matrix
+    from scipy.sparse.csgraph import connected_components
+
+    try:
+        from openbabel import openbabel
+    except ImportError:
+        import openbabel
     atomnumber = len(symbols)
     # Use openbabel to connect atoms
     mol = openbabel.OBMol()

dpdata/periodic_table.py

@@ -1,9 +1,9 @@
+import json
 from pathlib import Path
 
-from monty.serialization import loadfn
-
-fpdt = str(Path(__file__).absolute().parent / "periodic_table.json")
-_pdt = loadfn(fpdt)
+fpdt = Path(__file__).absolute().parent / "periodic_table.json"
+with fpdt.open("r") as fpdt:
+    _pdt = json.load(fpdt)
 ELEMENTS = [
     "H",
     "He",

dpdata/plugins/ase.py

@@ -6,15 +6,9 @@
 from dpdata.driver import Driver, Minimizer
 from dpdata.format import Format
 
-try:
-    import ase.io
-    from ase.calculators.calculator import PropertyNotImplementedError
-    from ase.io import Trajectory
-
-    if TYPE_CHECKING:
-        from ase.optimize.optimize import Optimizer
-except ImportError:
-    pass
+if TYPE_CHECKING:
+    import ase
+    from ase.optimize.optimize import Optimizer
 
 
 @Format.register("ase/structure")
@@ -84,6 +78,8 @@
             ASE will raise RuntimeError if the atoms does not
             have a calculator
         """
+        from ase.calculators.calculator import PropertyNotImplementedError
+
         info_dict = self.from_system(atoms)
         try:
             energies = atoms.get_potential_energy(force_consistent=True)
@@ -137,6 +133,8 @@
         ase.Atoms
             ASE atoms in the file
         """
+        import ase.io
+
         frames = ase.io.read(file_name, format=ase_fmt, index=slice(begin, end, step))
         yield from frames
 
@@ -222,6 +220,8 @@
         dict_frames: dict
             a dictionary containing data of multiple frames
         """
+        from ase.io import Trajectory
+
         traj = Trajectory(file_name)
         sub_traj = traj[begin:end:step]
         dict_frames = ASEStructureFormat().from_system(sub_traj[0])
@@ -264,6 +264,8 @@
         dict_frames: dict
             a dictionary containing data of multiple frames
         """
+        from ase.io import Trajectory
+
         traj = Trajectory(file_name)
         sub_traj = traj[begin:end:step]
 

dpdata/plugins/deepmd.py

@@ -1,11 +1,8 @@
 from __future__ import annotations
 
 import os
+from typing import TYPE_CHECKING
 
-try:
-    import h5py
-except ImportError:
-    pass
 import numpy as np
 
 import dpdata
@@ -16,6 +13,9 @@
 from dpdata.driver import Driver
 from dpdata.format import Format
 
+if TYPE_CHECKING:
+    import h5py
+
 
 @Format.register("deepmd")
 @Format.register("deepmd/raw")
@@ -202,6 +202,8 @@
         TypeError
             file_name is not str or h5py.Group or h5py.File
         """
+        import h5py
+
         if isinstance(file_name, (h5py.Group, h5py.File)):
             return dpdata.deepmd.hdf5.to_system_data(
                 file_name, "", type_map=type_map, labels=labels
@@ -300,6 +302,8 @@
         **kwargs : dict
             other parameters
         """
+        import h5py
+
         if isinstance(file_name, (h5py.Group, h5py.File)):
             dpdata.deepmd.hdf5.dump(
                 file_name, "", data, set_size=set_size, comp_prec=comp_prec
@@ -330,6 +334,8 @@
         h5py.Group
             a HDF5 group in the HDF5 file
         """
+        import h5py
+
         with h5py.File(directory, "r") as f:
             for ff in f.keys():
                 yield f[ff]
@@ -353,6 +359,8 @@
         h5py.Group
             a HDF5 group with the name of formula
         """
+        import h5py
+
         with h5py.File(directory, "w") as f:
             for ff in formulas:
                 yield f.create_group(ff)

dpdata/plugins/rdkit.py

@@ -1,20 +1,18 @@
+import dpdata.rdkit.utils
 from dpdata.format import Format
 
-try:
-    import rdkit.Chem
-
-    import dpdata.rdkit.utils
-except ModuleNotFoundError:
-    pass
-
 
 @Format.register("mol")
 @Format.register("mol_file")
 class MolFormat(Format):
     def from_bond_order_system(self, file_name, **kwargs):
+        import rdkit.Chem
+
         return rdkit.Chem.MolFromMolFile(file_name, sanitize=False, removeHs=False)
 
     def to_bond_order_system(self, data, mol, file_name, frame_idx=0, **kwargs):
+        import rdkit.Chem
+
         assert frame_idx < mol.GetNumConformers()
         rdkit.Chem.MolToMolFile(mol, file_name, confId=frame_idx)
 
@@ -24,6 +22,8 @@
 class SdfFormat(Format):
     def from_bond_order_system(self, file_name, **kwargs):
         """Note that it requires all molecules in .sdf file must be of the same topology."""
+        import rdkit.Chem
+
         mols = [
             m
             for m in rdkit.Chem.SDMolSupplier(file_name, sanitize=False, removeHs=False)
@@ -35,6 +35,8 @@
         return mol
 
     def to_bond_order_system(self, data, mol, file_name, frame_idx=-1, **kwargs):
+        import rdkit.Chem
+
         sdf_writer = rdkit.Chem.SDWriter(file_name)
         if frame_idx == -1:
             for ii in range(mol.GetNumConformers()):

dpdata/pymatgen/molecule.py

@@ -1,13 +1,11 @@
-import numpy as np
-
-try:
-    from pymatgen.core import Molecule
-except ImportError:
-    pass
 from collections import Counter
 
+import numpy as np
+
 
 def to_system_data(file_name, protect_layer=9):
+    from pymatgen.core import Molecule
+
     mol = Molecule.from_file(file_name)
     elem_mol = list(str(site.species.elements[0]) for site in mol.sites)
     elem_counter = Counter(elem_mol)

dpdata/pymatgen/structure.py

@@ -1,10 +1,5 @@
 import numpy as np
 
-try:
-    from pymatgen.core import Structure  # noqa: F401
-except ImportError:
-    pass
-
 
 def from_system_data(structure) -> dict:
     symbols = [site.species_string for site in structure]