model reorganization (#366)

bare rearrangement

bare rearrangement 2

fix up imports

new molecule fns

mol v2

convert foundational models

clear away version_stamp and placeholder Mol v3

basisset move and OptRes.trajectory rename

post-Christmas, mostly opt/td edits

.github/workflows/CI.yaml

@@ -5,12 +5,12 @@ on:
     branches:
       - master
       - next2024
-      - next2024_atop_v29
+      - next2025
   pull_request:
     branches:
       - master
       - next2024
-      - next2024_atop_v29
+      - next2025
   schedule:
     - cron: "9 16 * * 1"
 

docs/changelog.rst

@@ -37,8 +37,33 @@ New Features
 
 Enhancements
 ++++++++++++
-- (:pr:`363`) 
-- (:pr:`363`) 
+- (:pr:`364`) 
+- (:pr:`364`) 
+- (:pr:`364`) 
+- (:pr:`364`) main storage for ``v2.TorsionDriveResult`` moved from ``optimization_history`` to ``scan_results``.
+- (:pr:`364`) ``v2.TorsionDriveInput.initial_molecule`` restored from ``initial_molecules``.
+- (:pr:`364`) default of OptimizationProtocol.trajectory_results changed to "none" from "all" in v1. much info can now come from properties.
+- (:pr:`364`) v2.OptimizationProtocol renamed trajectory_results from trajectory in accordance with the protocol naming the controlled field. no default change yet.
+- (:pr:`364`) v1/v2: import ElectronShell, BasisCenter, ECPPotential from top level models
+- (:pr:`364`) molparse learns to pass through schema v3, though no new field for Mol yet.
+- (:pr:`364`) ``v2.FailedOperation`` gained schema_name and schema_version=2. unversioned in v1
+- (:pr:`364`) ``v2.BasisSet.schema_version`` is now 2, with no layout change.
+- (:pr:`364`) ``v2.Molecule.schema_version`` is now 3. convert_v of all the models learned to handle the new schema_version.
+- (:pr:`364`) v2: standardizing on In/Res get versions, Ptcl/Kw/Spec get only schema_name. At, Opt, TD
+- (:pr:`364`) v1/v2: removing the version_stamps from the models: At, Opt, TD, Fail, BAsis, Mol. so it will error rather than clobber if constructed with wrong version. convert_v now handles.
+- (:pr:`364`) convert_v functions learned to handle model.basis=BasisSet, not just str.
+- (:pr:`364`) ``Molecule`` and ``BasisSet``  and ``WavefunctionProperties`` learned to ``convert_v`` to interconvert between v1 and v2. No layout changes. 
+  ``BasisSet.schema_name`` standardized to ``qcschema_basis_set``.
+  Both classes get their ``schema_name`` as Literal now
+- (:pr:`360`) ``Molecule`` learned new functions ``element_composition`` and ``molecular_weight``.
+  The first gives a dictionary of element symbols and counts, while the second gives the weight in amu.
+  Both can access the whole molecule or per-fragment like the existing ``nelectrons`` and
+  ``nuclear_repulsion_energy``. All four can now select all atoms or exclude ghosts (default).
+- (:pr:`364`) separated procedures.py and renamed results.py so models are separated into atomic.py, optimization.py, torsion_drive.py, failure models moved to failed_operation.py. basis.py to basis_set.py
+- (:pr:`364`) ``schema_name`` output chanded to result ``qcschema_output`` to ``qcschema_atomic_result``. also opt
+- (:pr:`364`) ``TDKeywords`` renamed to ``TorsionDriveKeywords``
+- (:pr:`364`) ``AtomicResultProtocols`` renamed to ``AtomicProtocols`` and ``AtomicResultProperties`` to ``AtomicProperties``
+- (:pr:`364`) new ``v2.TorsionDriveProtocols`` model with field ``scan_results`` to control all/none/lowest saving of optimizationresults at each grid point. Use "all" for proper conversion to v1.
 - (:pr:`363`) ``v2.TorsionDriveResult`` no longer inherits from Input and now has indep id and extras and new native_files.
 - (:pr:`363`) ``v2.TorsionDriveInput.initial_molecule`` now ``initial_molecules`` as it's a list of >=1 molecules. keep change?
 - (:pr:`363`) ``v2. TorsionDriveSpecification`` is a new model. instead of ``v2.TorsionDriveInput`` having a ``input_specification`` and an ``optimization_spec`` fields, it has a ``specification`` field that is a ``TorsionDriveSpecification`` which in turn hold opt info and in turn gradient/atomic info. 

qcelemental/models/v1/__init__.py

@@ -2,7 +2,7 @@
 from .align import AlignmentMill
 from .basemodels import AutodocBaseSettings  # remove when QCFractal merges `next`
 from .basemodels import ProtoModel
-from .basis import BasisSet
+from .basis import BasisCenter, BasisSet, ECPPotential, ElectronShell
 from .common_models import ComputeError, DriverEnum, FailedOperation, Model, Provenance
 from .molecule import Molecule
 from .procedures import Optimization  # scheduled for removal

qcelemental/models/v1/basemodels.py

@@ -200,11 +200,22 @@ def compare(self, other: Union["ProtoModel", BaseModel], **kwargs) -> bool:
         bool
             True if the objects match.
         """
-        from ..testing import compare_recursive
+        from ...testing import compare_recursive
 
         return compare_recursive(self, other, **kwargs)
 
 
+def check_convertible_version(ver: int, error: str):
+    """Standardize version/error handling for v1 QCSchema."""
+
+    if ver == 1:
+        return "self"
+    elif ver == 2:
+        return True
+    else:
+        raise ValueError(f"QCSchema {error} version={ver} does not exist for conversion.")
+
+
 # remove when QCFractal merges `next`
 class AutodocBaseSettings(BaseSettings):
     """Old class for pydantic docstring before autodoc-pydantic came about."""

qcelemental/models/v1/basis.py

@@ -1,10 +1,13 @@
 from enum import Enum
-from typing import Dict, List, Literal, Optional
+from typing import TYPE_CHECKING, Dict, List, Literal, Optional, Union
 
 from pydantic.v1 import ConstrainedInt, Field, constr, validator
 
 from ...exceptions import ValidationError
-from .basemodels import ProtoModel, qcschema_draft
+from .basemodels import ProtoModel, check_convertible_version, qcschema_draft
+
+if TYPE_CHECKING:
+    import qcelemental
 
 
 class NonnegativeInt(ConstrainedInt):
@@ -175,10 +178,6 @@ class Config(ProtoModel.Config):
         def schema_extra(schema, model):
             schema["$schema"] = qcschema_draft
 
-    @validator("schema_version", pre=True)
-    def _version_stamp(cls, v):
-        return 1
-
     @validator("atom_map")
     def _check_atom_map(cls, v, values):
         sv = set(v)
@@ -230,3 +229,23 @@ def _calculate_nbf(cls, atom_map, center_data) -> int:
             ret += center_count[center]
 
         return ret
+
+    def convert_v(
+        self, target_version: int, /
+    ) -> Union["qcelemental.models.v1.BasisSet", "qcelemental.models.v2.BasisSet"]:
+        """Convert to instance of particular QCSchema version."""
+        import qcelemental as qcel
+
+        if check_convertible_version(target_version, error="BasisSet") == "self":
+            return self
+
+        dself = self.model_dump()
+        if target_version == 2:
+            dself.pop("schema_name")  # changes in v2
+            dself.pop("schema_version")  # changes in v2
+
+            self_vN = qcel.models.v2.BasisSet(**dself)
+        else:
+            assert False, target_version
+
+        return self_vN

qcelemental/models/v1/common_models.py

@@ -4,12 +4,14 @@
 import numpy as np
 from pydantic.v1 import Field
 
-from .basemodels import ProtoModel, qcschema_draft
+from .basemodels import ProtoModel, check_convertible_version, qcschema_draft
 from .basis import BasisSet
 
 if TYPE_CHECKING:
     from pydantic.v1.typing import ReprArgs
 
+    import qcelemental
+
 
 # Encoders, to be deprecated
 ndarray_encoder = {np.ndarray: lambda v: v.flatten().tolist()}
@@ -147,15 +149,6 @@ def convert_v(
         return self_vN
 
 
-def check_convertible_version(ver: int, error: str):
-    if ver == 1:
-        return "self"
-    elif ver == 2:
-        return True
-    else:
-        raise ValueError(f"QCSchema {error} version={version} does not exist for conversion.")
-
-
 qcschema_input_default = "qcschema_input"
 qcschema_output_default = "qcschema_output"
 qcschema_optimization_input_default = "qcschema_optimization_input"

qcelemental/models/v1/molecule.py

@@ -28,13 +28,15 @@
 from ...physical_constants import constants
 from ...testing import compare, compare_values
 from ...util import deserialize, measure_coordinates, msgpackext_loads, provenance_stamp, which_import
-from .basemodels import ProtoModel, qcschema_draft
+from .basemodels import ProtoModel, check_convertible_version, qcschema_draft
 from .common_models import Provenance, qcschema_molecule_default
 from .types import Array
 
 if TYPE_CHECKING:
     from pydantic.v1.typing import ReprArgs
 
+    import qcelemental
+
 # Rounding quantities for hashing
 GEOMETRY_NOISE = 8
 MASS_NOISE = 6
@@ -376,12 +378,6 @@ def __init__(self, orient: bool = False, validate: Optional[bool] = None, **kwar
         elif validate or geometry_prep:
             values["geometry"] = float_prep(values["geometry"], geometry_noise)
 
-    @validator("schema_version", pre=True)
-    def _version_stamp(cls, v):
-        # seemingly unneeded, this lets conver_v re-label the model w/o discarding model and
-        #   submodel version fields first.
-        return 2
-
     @validator("geometry")
     def _must_be_3n(cls, v, values, **kwargs):
         n = len(values["symbols"])
@@ -1562,6 +1558,28 @@ def scramble(
 
         return cmol, {"rmsd": rmsd, "mill": perturbation}
 
+    def convert_v(
+        self, target_version: int, /
+    ) -> Union["qcelemental.models.v1.Molecule", "qcelemental.models.v2.Molecule"]:
+        """Convert to instance of particular QCSchema version."""
+        import qcelemental as qcel
+
+        if check_convertible_version(target_version, error="Molecule") == "self":
+            return self
+
+        loss_store = {}
+        dself = self.model_dump()
+        if target_version == 2:
+            # below is assignment rather than popping so Mol() records as set and future Mol.model_dump() includes the field.
+            #   QCEngine _build_model convert_v(2) can lose it otherwise, and molparse machinery wants to see the field.
+            dself["schema_version"] = 3
+
+            self_vN = qcel.models.v2.Molecule(**dself)
+        else:
+            assert False, target_version
+
+        return self_vN
+
 
 def _filter_defaults(dicary):
     nat = len(dicary["symbols"])