environment.yml Unblocking Python 3.11 - rdkit to conda-forge channel, remove chemprop for now

.conda/meta.yaml

@@ -30,7 +30,6 @@ requirements:
     - cairocffi
     - cantera >=2.3.0
     - cclib >=1.6.3
-    - chemprop
     - coolprop
     - coverage
     - cython >=0.25.2

arkane/common.py

@@ -191,11 +191,11 @@ def update_species_attributes(self, species=None):
             self.multiplicity = species.molecule[0].multiplicity
             self.formula = species.molecule[0].get_formula()
             try:
-                inchi = to_inchi(species.molecule[0], backend='try-all', aug_level=0)
+                inchi = to_inchi(species.molecule[0], backend='openbabel-first', aug_level=0)
             except ValueError:
                 inchi = ''
             try:
-                inchi_key = to_inchi_key(species.molecule[0], backend='try-all', aug_level=0)
+                inchi_key = to_inchi_key(species.molecule[0], backend='openbabel-first', aug_level=0)
             except ValueError:
                 inchi_key = ''
             self.inchi = inchi

environment.yml

@@ -14,6 +14,8 @@
 # Changelog:
 # - May 15, 2023 Added this changelog, added inline documentation,
 #   made dependency list more explicit (@JacksonBurns).
+# - October 16, 2023 Switched RDKit and descripatastorus to conda-forge,
+#   moved diffeqpy to pip and (temporarily) removed chemprop
 #
 name: rmg_env
 channels:
@@ -44,6 +46,7 @@ dependencies:
   - conda-forge::mopac
   - conda-forge::cclib >=1.6.3,!=1.8.0
   - conda-forge::openbabel >= 3
+  - conda-forge::rdkit >=2022.09.1
 
 # general-purpose external software tools
   - conda-forge::julia=1.9.1
@@ -54,7 +57,7 @@ dependencies:
   - coverage
   - cython >=0.25.2
   - scikit-learn
-  - scipy
+  - scipy <1.11
   - numpy >=1.10.0
   - pydot
   - jinja2
@@ -94,14 +97,8 @@ dependencies:
   # rather than ours (which is only made so that we can get it from conda)
   # It is only on pip, so we will need to do something like:
   # https://stackoverflow.com/a/35245610
-
-  - rmg::chemprop
-  # Our build of this is version 0.0.1 (!!) and we are using parts
-  # of the API that are now gone. Need a serious PR to fix this.
-
-  - rmg::rdkit >=2020.03.3.0
-  # We should use the official channel, not sure how difficult this
-  # change will be.
+  # Note that _some other_ dep. in this list requires diffeqpy in its recipe
+  # which will cause it to be downloaded from the rmg conda channel
 
 # conda mutex metapackage
   - nomkl

rmgpy/ml/estimator.py

@@ -32,17 +32,27 @@
 from argparse import Namespace
 from typing import Callable, Union
 
+chemprop = None
 try:
     import chemprop
 except ImportError as e:
-    chemprop = None
     chemprop_exception = e
 import numpy as np
 
 from rmgpy.molecule import Molecule
 from rmgpy.species import Species
 from rmgpy.thermo import ThermoData
 
+ADMONITION = """
+Support for predicting thermochemistry using chemprop has been temporarily removed
+from RMG, pending official chemprop support for Python 3.11 and newer.
+
+To use chemprop and RMG, install a previous version of RMG (3.1.1 or earlier).
+
+See the link below for status of re-integration of chemprop:
+https://github.com/ReactionMechanismGenerator/RMG-Py/issues/2559
+"""
+
 
 class MLEstimator:
     """
@@ -118,7 +128,7 @@ def load_estimator(model_dir: str) -> Callable[[str], np.ndarray]:
     if chemprop is None:
         # Delay chemprop ImportError until we actually try to use it
         # so that RMG can load successfully without chemprop.
-        raise chemprop_exception
+        raise RuntimeError(ADMONITION + "\nOriginal Exception:\n" + str(chemprop_exception))
 
     args = Namespace()  # Simple class to hold attributes
 

rmgpy/molecule/molecule.pxd

@@ -224,14 +224,14 @@ cdef class Molecule(Graph):
                               bint raise_charge_exception=?, bint check_consistency=?)
 
     cpdef from_xyz(self, np.ndarray atomic_nums, np.ndarray coordinates, float critical_distance_factor=?, bint raise_atomtype_exception=?)
-
-    cpdef str to_inchi(self)
 
-    cpdef str to_augmented_inchi(self)
+    cpdef str to_inchi(self, str backend=?)
+
+    cpdef str to_augmented_inchi(self, str backend=?)
 
-    cpdef str to_inchi_key(self)
+    cpdef str to_inchi_key(self, str backend=?)
 
-    cpdef str to_augmented_inchi_key(self)
+    cpdef str to_augmented_inchi_key(self, str backend=?)
 
     cpdef str to_smiles(self)
 

rmgpy/molecule/molecule.py

@@ -1768,9 +1768,13 @@ def _repr_png_(self):
         os.unlink(temp_file_name)
         return png
 
-    def from_inchi(self, inchistr, backend='try-all', raise_atomtype_exception=True):
+    def from_inchi(self, inchistr, backend='openbabel-first', raise_atomtype_exception=True):
         """
         Convert an InChI string `inchistr` to a molecular structure.
+
+        RDKit and Open Babel are the two backends used in RMG. It is possible to use a
+        single backend or try different backends in sequence. The available options for the ``backend``
+        argument: 'openbabel-first'(default), 'rdkit-first', 'rdkit', or 'openbabel'.
         """
         translator.from_inchi(self, inchistr, backend, raise_atomtype_exception=raise_atomtype_exception)
         return self
@@ -1782,9 +1786,13 @@ def from_augmented_inchi(self, aug_inchi, raise_atomtype_exception=True):
         translator.from_augmented_inchi(self, aug_inchi, raise_atomtype_exception=raise_atomtype_exception)
         return self
 
-    def from_smiles(self, smilesstr, backend='try-all', raise_atomtype_exception=True):
+    def from_smiles(self, smilesstr, backend='openbabel-first', raise_atomtype_exception=True):
         """
         Convert a SMILES string `smilesstr` to a molecular structure.
+
+        RDKit and Open Babel are the two backends used in RMG. It is possible to use a
+        single backend or try different backends in sequence. The available options for the ``backend``
+        argument: 'openbabel-first'(default), 'rdkit-first', 'rdkit', or 'openbabel'.
         """
         translator.from_smiles(self, smilesstr, backend, raise_atomtype_exception=raise_atomtype_exception)
         return self
@@ -1863,62 +1871,78 @@ def to_single_bonds(self, raise_atomtype_exception=True):
         new_mol.update_atomtypes(raise_exception=raise_atomtype_exception)
         return new_mol
 
-    def to_inchi(self):
+    def to_inchi(self, backend='rdkit-first'):
         """
         Convert a molecular structure to an InChI string. Uses
         `RDKit <http://rdkit.org/>`_ to perform the conversion.
         Perceives aromaticity.
-        
+
         or
-        
+
         Convert a molecular structure to an InChI string. Uses
         `OpenBabel <http://openbabel.org/>`_ to perform the conversion.
+
+        It is possible to use a single backend or try different backends in sequence.
+        The available options for the ``backend`` argument: 'rdkit-first'(default),
+        'openbabel-first', 'rdkit', or 'openbabel'.
         """
         try:
-            return translator.to_inchi(self)
+            return translator.to_inchi(self, backend=backend)
         except:
             logging.exception(f"Error for molecule \n{self.to_adjacency_list()}")
             raise
 
-    def to_augmented_inchi(self):
+    def to_augmented_inchi(self, backend='rdkit-first'):
         """
         Adds an extra layer to the InChI denoting the multiplicity
         of the molecule.
-        
+
         Separate layer with a forward slash character.
+
+        RDKit and Open Babel are the two backends used in RMG. It is possible to use a
+        single backend or try different backends in sequence. The available options for the ``backend``
+        argument: 'rdkit-first'(default), 'openbabel-first', 'rdkit', or 'openbabel'.
         """
         try:
-            return translator.to_inchi(self, aug_level=2)
+            return translator.to_inchi(self, backend=backend, aug_level=2)
         except:
             logging.exception(f"Error for molecule \n{self.to_adjacency_list()}")
             raise
 
-    def to_inchi_key(self):
+    def to_inchi_key(self, backend='rdkit-first'):
         """
         Convert a molecular structure to an InChI Key string. Uses
         `OpenBabel <http://openbabel.org/>`_ to perform the conversion.
-        
-        or 
-        
+
+        or
+
         Convert a molecular structure to an InChI Key string. Uses
         `RDKit <http://rdkit.org/>`_ to perform the conversion.
+
+        It is possible to use a single backend or try different backends in sequence.
+        The available options for the ``backend`` argument: 'rdkit-first'(default),
+        'openbabel-first', 'rdkit', or 'openbabel'.
         """
         try:
-            return translator.to_inchi_key(self)
+            return translator.to_inchi_key(self, backend=backend)
         except:
             logging.exception(f"Error for molecule \n{self.to_adjacency_list()}")
             raise
 
-    def to_augmented_inchi_key(self):
+    def to_augmented_inchi_key(self, backend='rdkit-first'):
         """
         Adds an extra layer to the InChIKey denoting the multiplicity
         of the molecule.
 
         Simply append the multiplicity string, do not separate by a
         character like forward slash.
+
+        RDKit and Open Babel are the two backends used in RMG. It is possible to use a
+        single backend or try different backends in sequence. The available options for the ``backend``
+        argument: 'rdkit-first'(default), 'openbabel-first', 'rdkit', or 'openbabel'.
         """
         try:
-            return translator.to_inchi_key(self, aug_level=2)
+            return translator.to_inchi_key(self, backend=backend, aug_level=2)
         except:
             logging.exception(f"Error for molecule \n{self.to_adjacency_list()}")
             raise

rmgpy/molecule/translator.py

@@ -169,7 +169,7 @@ def to_inchi(mol, backend='rdkit-first', aug_level=0):
     Uses RDKit or OpenBabel for conversion.
 
     Args:
-        backend     choice of backend, 'try-all', 'rdkit', or 'openbabel'
+        backend     choice of backend, 'rdkit-first' (default), 'openbabel-first', 'rdkit', or 'openbabel'
         aug_level   level of augmentation, 0, 1, or 2
     """
     cython.declare(inchi=str, ulayer=str, player=str, mlayer=str)
@@ -205,7 +205,7 @@ def to_inchi_key(mol, backend='rdkit-first', aug_level=0):
     Uses RDKit or OpenBabel for conversion.
 
     Args:
-        backend     choice of backend, 'try-all', 'rdkit', or 'openbabel'
+        backend     choice of backend, 'rdkit-first' (default), 'openbabel-first', 'rdkit', or 'openbabel'
         aug_level   level of augmentation, 0, 1, or 2
     """
     cython.declare(key=str, ulayer=str, player=str, mlayer=str)
@@ -274,11 +274,11 @@ def to_smiles(mol, backend='default'):
         return output
 
 
-def from_inchi(mol, inchistr, backend='try-all', raise_atomtype_exception=True):
+def from_inchi(mol, inchistr, backend='openbabel-first', raise_atomtype_exception=True):
     """
     Convert an InChI string `inchistr` to a molecular structure. Uses
-    a user-specified backend for conversion, currently supporting
-    rdkit (default) and openbabel.
+    a user-specified backend for conversion, currently supporting 'openbabel-first' (default), rdkit-first,
+    rdkit, and openbabel.
     """
     if inchiutil.INCHI_PREFIX in inchistr:
         return _read(mol, inchistr, 'inchi', backend, raise_atomtype_exception=raise_atomtype_exception)
@@ -325,11 +325,11 @@ def from_smarts(mol, smartsstr, backend='rdkit', raise_atomtype_exception=True):
     return _read(mol, smartsstr, 'sma', backend, raise_atomtype_exception=raise_atomtype_exception)
 
 
-def from_smiles(mol, smilesstr, backend='try-all', raise_atomtype_exception=True):
+def from_smiles(mol, smilesstr, backend='openbabel-first', raise_atomtype_exception=True):
     """
     Convert a SMILES string `smilesstr` to a molecular structure. Uses
-    a user-specified backend for conversion, currently supporting
-    rdkit (default) and openbabel.
+    a user-specified backend for conversion, currently supporting openbabel-first (default), rdkit-first,
+    rdkit and openbabel.
     """
     return _read(mol, smilesstr, 'smi', backend, raise_atomtype_exception=raise_atomtype_exception)
 
@@ -569,14 +569,14 @@ def _get_backend_list(backend):
     """
     if not isinstance(backend, str):
         raise ValueError("The backend argument should be a string. "
-                         "Accepted values are 'try-all', 'rdkit-first', 'rdkit', and 'openbabel'")
+                         "Accepted values are 'openbabel-first', 'rdkit-first', 'rdkit', and 'openbabel'")
     backend = backend.strip().lower()
-    if backend == 'try-all':
+    if backend == 'openbabel-first':
         return BACKENDS
     elif backend == 'rdkit-first':
         return reversed(BACKENDS)
     elif backend in ['rdkit', 'openbabel']:
         return [backend]
     else:
         raise ValueError("Unrecognized value for backend argument. "
-                         "Accepted values are 'try-all', 'rdkit-first', 'rdkit', and 'openbabel'")
+                         "Accepted values are 'openbabel-first', 'rdkit-first', 'rdkit', and 'openbabel'")

rmgpy/qm/molecule.py

@@ -520,11 +520,11 @@ def load_thermo_data(self):
         self.qm_data = local_context["qmData"]
         return thermo
 
-    def get_augmented_inchi_key(self):
+    def get_augmented_inchi_key(self, backend='rdkit-first'):
         """
         Returns the augmented InChI from self.molecule
         """
-        return self.molecule.to_augmented_inchi_key()
+        return self.molecule.to_augmented_inchi_key(backend=backend)
 
     def get_mol_file_path_for_calculation(self, attempt):
         """

rmgpy/species.py

@@ -740,17 +740,17 @@ def copy(self, deep=False):
 
         return other
 
-    def get_augmented_inchi(self):
+    def get_augmented_inchi(self, backend='rdkit-first'):
         if self.aug_inchi is None:
-            self.aug_inchi = self.generate_aug_inchi()
+            self.aug_inchi = self.generate_aug_inchi(backend=backend)
         return self.aug_inchi
 
-    def generate_aug_inchi(self):
+    def generate_aug_inchi(self, backend='rdkit-first'):
         candidates = []
         self.generate_resonance_structures()
         for mol in self.molecule:
             try:
-                cand = [mol.to_augmented_inchi(), mol]
+                cand = [mol.to_augmented_inchi(backend=backend), mol]
             except ValueError:
                 pass  # not all resonance structures can be parsed into InChI (e.g. if containing a hypervalance atom)
             else:

scripts/checkModels.py

@@ -284,6 +284,18 @@ def initialize_log(verbose, log_file_name='checkModels.log'):
     `verbose` parameter is an integer specifying the amount of log text seen
     at the console; the levels correspond to those of the :data:`logging` module.
     """
+    # since RDKit 2022.03.1, logging is done using the Python logger instead of the
+    # Cout streams. This does not affect running RMG normally, but this testing file
+    # only works properly if it is the only logger
+    # see https://github.com/rdkit/rdkit/pull/4846 for the changes in RDKit
+
+    # clear all other existing loggers
+    # https://stackoverflow.com/a/12158233
+    for handler in logging.root.handlers[:]:
+        logging.root.removeHandler(handler)
+
+    # once moved to a more recent python (at least 3.8), just add force=true to this statement
+    # and remove the above
     logging.basicConfig(
         filename=log_file_name,
         filemode='w',