stjames 0.0.42__tar.gz → 0.0.44__tar.gz

{stjames-0.0.42/stjames.egg-info → stjames-0.0.44}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: stjames
-Version: 0.0.42
+Version: 0.0.44
 Summary: standardized JSON atom/molecule encoding scheme
 Author-email: Corin Wagen <corin@rowansci.com>
 Project-URL: Homepage, https://github.com/rowansci/stjames

{stjames-0.0.42 → stjames-0.0.44}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "stjames"
-version = "0.0.42"
+version = "0.0.44"
 description = "standardized JSON atom/molecule encoding scheme"
 readme = "README.md"
 requires-python = ">=3.11"

{stjames-0.0.42 → stjames-0.0.44}/stjames/atom.py

@@ -60,7 +60,7 @@ class Atom(Base):
         Atom(1, [0.00000, 0.00000, 0.00000])
         """
         name, *xyz = xyz_line.split()
-        symbol = int(name) if name.isdigit() else SYMBOL_ELEMENT[name]
+        symbol = int(name) if name.isdigit() else SYMBOL_ELEMENT[name.title()]
         if not len(xyz) == 3:
             raise ValueError("XYZ file should have 3 coordinates per atom")
         return cls(atomic_number=symbol, position=xyz)

{stjames-0.0.42 → stjames-0.0.44}/stjames/molecule.py

@@ -1,8 +1,9 @@
+import re
 from pathlib import Path
 from typing import Iterable, Optional, Self
 
 import pydantic
-from pydantic import NonNegativeInt, PositiveInt
+from pydantic import NonNegativeInt, PositiveInt, ValidationError
 
 from .atom import Atom
 from .base import Base
@@ -54,6 +55,8 @@ class Molecule(Base):
     thermal_enthalpy_corr: Optional[float] = None
     thermal_free_energy_corr: Optional[float] = None
 
+    smiles: Optional[str] = None
+
     def __len__(self) -> int:
         return len(self.atoms)
 
@@ -135,6 +138,8 @@ class Molecule(Base):
             match format:
                 case "xyz":
                     return cls.from_xyz_lines(f.readlines(), charge=charge, multiplicity=multiplicity)
+                case "extxyz":
+                    return cls.from_extxyz_lines(f.readlines(), charge=charge, multiplicity=multiplicity)
                 case _:
                     raise ValueError(f"Unsupported {format=}")
 
@@ -161,7 +166,7 @@ class Molecule(Base):
 
         try:
             return cls(atoms=[Atom.from_xyz(line) for line in lines], charge=charge, multiplicity=multiplicity)
-        except Exception as e:
+        except (ValueError, ValidationError) as e:
             raise MoleculeReadError("Error reading molecule from xyz") from e
 
     def to_xyz(self, comment: str = "", out_file: Path | str | None = None) -> str:
@@ -190,3 +195,85 @@ class Molecule(Base):
                 f.write(out)
 
         return out
+
+    @classmethod
+    def from_extxyz(cls: type[Self], extxyz: str, charge: int = 0, multiplicity: PositiveInt = 1) -> Self:
+        r"""
+        Generate a Molecule from a EXTXYZ string. Currently only supporting Lattice and Properties fields.
+
+        >>> Molecule.from_extxyz('''
+        ... 2
+        ... Lattice="6.0 0.0 0.0 6.0 0.0 0.0 6.0 0.0 0.0"Properties=species:S:1:pos:R:3
+        ... H 0 0 0
+        ... H 0 0 1
+        ... ''').cell.lattice_vectors
+        ((6.0, 0.0, 0.0), (6.0, 0.0, 0.0), (6.0, 0.0, 0.0))
+        """
+
+        return cls.from_extxyz_lines(extxyz.strip().splitlines(), charge=charge, multiplicity=multiplicity)
+
+    @classmethod
+    def from_extxyz_lines(cls: type[Self], lines: Iterable[str], charge: int = 0, multiplicity: PositiveInt = 1) -> Self:
+        # ensure first line is number of atoms
+        lines = list(lines)
+        if len(lines[0].split()) == 1:
+            natoms = lines[0].strip()
+            if not natoms.isdigit() or (int(lines[0]) != len(lines) - 2):
+                raise MoleculeReadError(f"First line of EXTXYZ file should be the number of atoms, got: {lines[0]} != {len(lines) - 2}")
+            lines = lines[1:]
+        else:
+            raise MoleculeReadError(f"First line of EXTXYZ should be only an int denoting number of atoms. Got {lines[0].split()}")
+
+        # ensure second line contains key-value pairs
+        if "=" not in lines[0]:
+            raise MoleculeReadError(f"Invalid property line, got {lines[0]}")
+
+        cell = parse_comment_line(lines[0])
+        lines = lines[1:]
+
+        try:
+            return cls(atoms=[Atom.from_xyz(line) for line in lines], cell=cell, charge=charge, multiplicity=multiplicity)
+        except (ValueError, ValidationError) as e:
+            raise MoleculeReadError("Error reading molecule from extxyz") from e
+
+
+def parse_comment_line(line: str) -> PeriodicCell:
+    """
+    currently only supporting lattice and porperites fields from comment line
+    modify in future to support other fields from comment from_xyz_lines
+    ex: name, mulitplicity, charge, etc.
+    """
+    cell = None
+    # Regular expression to match key="value", key='value', or key=value
+    pattern = r"(\S+?=(?:\".*?\"|\'.*?\'|\S+))"
+    pairs = re.findall(pattern, line)
+
+    prop_dict = {}
+    for pair in pairs:
+        key, value = pair.split("=", 1)
+        if key.lower() == "lattice":
+            value = value.strip("'\"").split()
+            if len(value) != 9:
+                raise MoleculeReadError(f"Lattice should have 9 entries got {len(value)}")
+
+            # Convert the value to a 3x3 tuple of tuples of floats
+            try:
+                cell = tuple(tuple(map(float, value[i : i + 3])) for i in range(0, 9, 3))
+            except ValueError:
+                raise MoleculeReadError(f"Lattice should be floats, got {value}")
+
+            prop_dict[key] = value
+
+        elif key.lower() == "properties":
+            if value.lower() != "species:s:1:pos:r:3":
+                raise MoleculeReadError(f"Only accepting properties of form species:S:1:pos:R:3, got {value}")
+            prop_dict[key] = value
+        else:
+            raise MoleculeReadError(f"Currently only accepting lattice and propery keys. Got {key}")
+
+    if cell is None:
+        raise MoleculeReadError("Lattice field is required but missing.")
+
+    if "properties" not in [key.lower() for key in prop_dict.keys()]:
+        raise MoleculeReadError(f"Property field is required, got keys {prop_dict.keys()}")
+    return PeriodicCell(lattice_vectors=cell)

stjames-0.0.44/stjames/settings.py

@@ -0,0 +1,216 @@
+from typing import Any, Optional, Self, TypeVar
+
+from pydantic import computed_field, field_validator, model_validator
+
+from .base import Base, UniqueList
+from .basis_set import BasisSet
+from .correction import Correction
+from .method import METHODS_WITH_CORRECTION, PREPACKAGED_METHODS, Method
+from .mode import Mode
+from .opt_settings import OptimizationSettings
+from .scf_settings import SCFSettings
+from .solvent import SolventSettings
+from .task import Task
+from .thermochem_settings import ThermochemistrySettings
+
+_T = TypeVar("_T")
+
+
+class Settings(Base):
+    mode: Mode = Mode.AUTO
+
+    method: Method = Method.HARTREE_FOCK
+    basis_set: Optional[BasisSet] = None
+    tasks: UniqueList[Task] = [Task.ENERGY, Task.CHARGE, Task.DIPOLE]
+    corrections: UniqueList[Correction] = []
+
+    solvent_settings: Optional[SolventSettings] = None
+
+    # scf/opt settings will be set automatically based on mode, but can be overridden manually
+    scf_settings: SCFSettings = SCFSettings()
+    opt_settings: OptimizationSettings = OptimizationSettings()
+    thermochem_settings: ThermochemistrySettings = ThermochemistrySettings()
+
+    # mypy has this dead wrong (https://docs.pydantic.dev/2.0/usage/computed_fields/)
+    # Python 3.12 narrows the reason for the ignore to prop-decorator
+    @computed_field  # type: ignore[misc, prop-decorator, unused-ignore]
+    @property
+    def level_of_theory(self) -> str:
+        corrections = list(filter(lambda x: x not in (None, ""), self.corrections))
+
+        if self.method in PREPACKAGED_METHODS or self.basis_set is None:
+            method = self.method.value
+        elif self.method in METHODS_WITH_CORRECTION or len(corrections) == 0:
+            method = f"{self.method.value}/{self.basis_set.name.lower()}"
+        else:
+            method = f"{self.method.value}-{'-'.join([c.value for c in corrections])}/{self.basis_set.name.lower()}"
+
+        if self.solvent_settings is not None:
+            method += f"/{self.solvent_settings.model.value}({self.solvent_settings.solvent.value})"
+
+        return method
+
+    @field_validator("mode")
+    @classmethod
+    def set_mode_auto(cls, mode: Mode) -> Mode:
+        """Set the mode to RAPID if AUTO is selected."""
+        if mode == Mode.AUTO:
+            return Mode.RAPID
+
+        return mode
+
+    @model_validator(mode="after")
+    def validate_and_build(self) -> Self:
+        if self.mode == Mode.AUTO:
+            self.mode = Mode.RAPID
+
+        self.scf_settings = _assign_scf_settings_by_mode(self.mode, self.scf_settings)
+        self.opt_settings = _assign_opt_settings_by_mode(self.mode, self.opt_settings)
+
+        return self
+
+    def model_post_init(self, __context: Any) -> None:
+        # figure out `optimize_ts`
+        if Task.OPTIMIZE_TS in self.tasks:
+            self.tasks.pop(self.tasks.index(Task.OPTIMIZE_TS))
+            self.tasks.append(Task.OPTIMIZE)
+            self.opt_settings.transition_state = True
+
+        # composite methods have their own basis sets, so overwrite user stuff
+        if self.method == Method.HF3C:
+            self.basis_set = BasisSet(name="minix")
+        elif self.method == Method.B973C:
+            self.basis_set = BasisSet(name="def2-mTZVP")
+        elif self.method == Method.R2SCAN3C:
+            self.basis_set = BasisSet(name="def2-mTZVPP")
+        elif self.method == Method.WB97X3C:
+            self.basis_set = BasisSet(name="vDZP")
+
+    @field_validator("basis_set", mode="before")
+    @classmethod
+    def parse_basis_set(cls, v: Any) -> BasisSet | dict[str, Any] | None:
+        """Turn a string into a ``BasisSet`` object. (This is a little crude.)"""
+        if isinstance(v, BasisSet):
+            return None if v.name is None else v
+        elif isinstance(v, dict):
+            return None if v.get("name") is None else v
+        elif isinstance(v, str):
+            if len(v):
+                return BasisSet(name=v)
+            # "" is basically None, let's be real here...
+            return None
+        elif v is None:
+            return None
+        else:
+            raise ValueError(f"invalid value ``{v}`` for ``basis_set``")
+
+    @field_validator("corrections", mode="before")
+    @classmethod
+    def remove_empty_string(cls, v: list[_T]) -> list[_T]:
+        """Remove empty string values."""
+        return [c for c in v if c] if v is not None else v
+
+
+def _assign_scf_settings_by_mode(mode: Mode, scf_settings: SCFSettings) -> SCFSettings:
+    """
+    Assign SCF settings based on the mode.
+
+    Values based off of the following sources:
+    QChem:
+        - https://manual.q-chem.com/5.2/Ch4.S3.SS2.html
+        - https://manual.q-chem.com/5.2/Ch4.S5.SS2.html
+
+    Gaussian:
+        - https://gaussian.com/integral/
+        - https://gaussian.com/overlay5/
+
+    Orca:
+        - manual 4.2.1, §9.6.1 and §9.7.3
+
+    Psi4:
+        - https://psicode.org/psi4manual/master/autodir_options_c/module__scf.html
+        - https://psicode.org/psi4manual/master/autodoc_glossary_options_c.html
+
+    TeraChem:
+        - Manual, it's easy to locate everything.
+
+    The below values are my best attempt at homogenizing various sources.
+    In general, eri_threshold should be 3 OOM lower than SCF convergence.
+    """
+    if mode == Mode.MANUAL:
+        return scf_settings
+
+    match mode:
+        case Mode.RECKLESS:
+            scf_settings.energy_threshold = 1e-5
+            scf_settings.rms_error_threshold = 1e-7
+            scf_settings.max_error_threshold = 1e-5
+            scf_settings.rebuild_frequency = 100
+            scf_settings.int_settings.eri_threshold = 1e-8
+            scf_settings.int_settings.csam_multiplier = 3.0
+            scf_settings.int_settings.pair_overlap_threshold = 1e-8
+        case Mode.RAPID | Mode.CAREFUL:
+            scf_settings.energy_threshold = 1e-6
+            scf_settings.rms_error_threshold = 1e-9
+            scf_settings.max_error_threshold = 1e-7
+            scf_settings.rebuild_frequency = 10
+            scf_settings.int_settings.eri_threshold = 1e-10
+            scf_settings.int_settings.csam_multiplier = 1.0
+            scf_settings.int_settings.pair_overlap_threshold = 1e-10
+        case Mode.METICULOUS:
+            scf_settings.energy_threshold = 1e-8
+            scf_settings.rms_error_threshold = 1e-9
+            scf_settings.max_error_threshold = 1e-7
+            scf_settings.rebuild_frequency = 5
+            scf_settings.int_settings.eri_threshold = 1e-12
+            scf_settings.int_settings.csam_multiplier = 1.0
+            scf_settings.int_settings.pair_overlap_threshold = 1e-12
+        case Mode.DEBUG:
+            scf_settings.energy_threshold = 1e-9
+            scf_settings.rms_error_threshold = 1e-10
+            scf_settings.max_error_threshold = 1e-9
+            scf_settings.rebuild_frequency = 1
+            scf_settings.int_settings.eri_threshold = 1e-14
+            scf_settings.int_settings.csam_multiplier = 1e10  # in other words, disable CSAM
+            scf_settings.int_settings.pair_overlap_threshold = 1e-14
+        case _:
+            raise ValueError(f"Unknown mode ``{mode.value}``!")
+
+    return scf_settings
+
+
+def _assign_opt_settings_by_mode(mode: Mode, opt_settings: OptimizationSettings) -> OptimizationSettings:
+    """
+    Assign optimization settings based on the mode.
+
+    Constraints lead to a lot of noise, so we need to loosen the thresholds.
+
+    cf. DLFIND manual, and https://www.cup.uni-muenchen.de/ch/compchem/geom/basic.html
+    and the discussion at https://geometric.readthedocs.io/en/latest/how-it-works.html
+    in periodic systems, "normal" is 0.05 eV/Å ~= 2e-3 Hartree/Å, and "careful" is 0.01 ~= 4e-4
+
+    Note: thresholds here are in units of Hartree/Å, not Hartree/Bohr as listed in many places.
+    """
+    opt_settings.energy_threshold = 1e-6
+    match mode:
+        case Mode.RECKLESS:
+            opt_settings.energy_threshold = 2e-5
+            opt_settings.max_gradient_threshold = 7e-3
+            opt_settings.rms_gradient_threshold = 6e-3
+        case Mode.RAPID:
+            opt_settings.energy_threshold = 5e-5
+            opt_settings.max_gradient_threshold = 5e-3
+            opt_settings.rms_gradient_threshold = 3.5e-3
+        case Mode.CAREFUL:
+            opt_settings.max_gradient_threshold = 9e-4
+            opt_settings.rms_gradient_threshold = 6e-4
+        case Mode.METICULOUS:
+            opt_settings.max_gradient_threshold = 3e-5
+            opt_settings.rms_gradient_threshold = 2e-5
+        case Mode.DEBUG:
+            opt_settings.max_gradient_threshold = 4e-6
+            opt_settings.rms_gradient_threshold = 2e-6
+        case _:
+            raise ValueError(f"Unknown mode ``{mode.value}``!")
+
+    return opt_settings

{stjames-0.0.42 → stjames-0.0.44}/stjames/workflows/bde.py

@@ -56,6 +56,7 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
 
     Inherited:
     :param initial_molecule: Molecule of interest
+    :param mode: Mode for workflow
     :param multistage_opt_settings: set by mode unless mode=MANUAL (ignores additional settings if set)
     :param solvent: solvent to use
     :param xtb_preopt: pre-optimize with xtb (sets based on mode when None)
@@ -69,7 +70,6 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
     :param transition_state: whether this is a transition state (not supported)
 
     New:
-    :param mode: Mode for workflow
     :param optimize_fragments: whether to optimize the fragments, or just the starting molecule (default depends on mode)
     :param atoms: atoms to dissociate (1-indexed)
     :param fragment_indices: fragments to dissociate (all fields feed into this, 1-indexed)
@@ -80,7 +80,6 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
     :param bdes: BDE results
     """
 
-    mode: Mode
     mso_mode: Mode = _sentinel_mso_mode  # type: ignore [assignment]
     frequencies: bool = False
     optimize_fragments: bool = None  # type: ignore [assignment]
@@ -107,15 +106,6 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
         """
         return f"{type(self).__name__} {self.mode.name}\n" + "\n".join(map(str, self.fragment_indices))
 
-    def __repr__(self) -> str:
-        """
-        Return a string representation of the BDE workflow.
-
-        >>> BDEWorkflow(initial_molecule=Molecule.from_xyz("He 0 0 0"), mode=Mode.METICULOUS, atoms=[])
-        <BDEWorkflow METICULOUS>
-        """
-        return f"<{type(self).__name__} {self.mode.name}>"
-
     @property
     def energies(self) -> tuple[float | None, ...]:
         return tuple(bde.energy for bde in self.bdes)
@@ -128,22 +118,21 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
 
         return value
 
-    @field_validator("mode")
-    @classmethod
-    def set_mode_auto(cls, mode: Mode) -> Mode:
-        if mode == Mode.AUTO:
-            return Mode.RAPID
-
-        return mode
-
     @field_validator("initial_molecule", mode="before")
     @classmethod
-    def no_charge_or_spin(cls, mol: Molecule) -> Molecule:
+    def no_charge_or_spin(cls, val: Molecule | dict[str, Any]) -> Molecule | dict[str, Any]:
         """Ensure the molecule has no charge or spin."""
+        if isinstance(val, dict):
+            mol = Molecule(**val)
+        elif isinstance(val, Molecule):
+            mol = val
+        else:
+            raise ValueError(f"{val=} is not a Molecule.")
+
         if mol.charge != 0 or mol.multiplicity != 1:
             raise ValueError("Charge and spin partitioning undefined for BDE, only neutral singlet molecules supported.")
 
-        return mol
+        return val
 
     @model_validator(mode="before")
     @classmethod
@@ -159,10 +148,10 @@ class BDEWorkflow(Workflow, MultiStageOptMixin):
         self.fragment_indices = tuple(map(tuple, self.fragment_indices))
 
         match self.mode:
-            case Mode.RECKLESS | Mode.RAPID:
-                # Default off
-                self.optimize_fragments = self.optimize_fragments or False
-            case Mode.CAREFUL | Mode.METICULOUS:
+            case Mode.RECKLESS:
+                # GFN-FF doesn't support open-shell species
+                self.optimize_fragments = False
+            case Mode.RAPID | Mode.CAREFUL | Mode.METICULOUS:
                 # Default on
                 self.optimize_fragments = self.optimize_fragments or self.optimize_fragments is None
             case _:

{stjames-0.0.42 → stjames-0.0.44}/stjames/workflows/molecular_dynamics.py

@@ -56,5 +56,7 @@ class MolecularDynamicsWorkflow(Workflow):
     calc_settings: Settings
     calc_engine: str | None = None
 
+    save_interval: PositiveInt = 10
+
     # uuids of scan points
     frames: list[Frame] = []

{stjames-0.0.42 → stjames-0.0.44}/stjames/workflows/multistage_opt.py

@@ -22,9 +22,9 @@ class MultiStageOptSettings(BaseModel):
     RAPID *default
         r²SCAN-3c//GFN2-xTB with GFN0-xTB pre-opt (off by default)
     CAREFUL
-        wB97X-3c//B97-3c with GFN2-xTB pre-opt
+        wB97X-3c//r²SCAN-3c with GFN2-xTB pre-opt
     METICULOUS
-        wB97M-D3BJ/def2-TZVPPD//wB97X-3c//B97-3c with GFN2-xTB pre-opt
+        wB97M-D3BJ/def2-TZVPPD//wB97X-3c//r²SCAN-3c with GFN2-xTB pre-opt
 
     Notes:
     - No solvent in pre-opt
@@ -163,7 +163,7 @@ class MultiStageOptSettings(BaseModel):
                 self.xtb_preopt = (self.xtb_preopt is None) or self.xtb_preopt
                 self.optimization_settings = [
                     *gfn2_pre_opt * self.xtb_preopt,
-                    opt(Method.B973C, solvent=self.solvent, freq=self.frequencies),
+                    opt(Method.R2SCAN3C, solvent=self.solvent, freq=self.frequencies),
                 ]
                 self.singlepoint_settings = sp(Method.WB97X3C, solvent=self.solvent)
 
@@ -171,7 +171,7 @@ class MultiStageOptSettings(BaseModel):
                 self.xtb_preopt = (self.xtb_preopt is None) or self.xtb_preopt
                 self.optimization_settings = [
                     *gfn2_pre_opt * self.xtb_preopt,
-                    opt(Method.B973C, solvent=self.solvent),
+                    opt(Method.R2SCAN3C, solvent=self.solvent),
                     opt(Method.WB97X3C, solvent=self.solvent, freq=self.frequencies),
                 ]
                 self.singlepoint_settings = sp(Method.WB97MD3BJ, "def2-TZVPPD", solvent=self.solvent)
@@ -212,6 +212,12 @@ class MultiStageOptWorkflow(Workflow, MultiStageOptSettings):
     # Populated while running the workflow
     calculations: list[UUID | None] = Field(default_factory=list)
 
+    def __repr__(self) -> str:
+        if self.mode != Mode.MANUAL:
+            return f"<{type(self).__name__} {self.mode.name}>"
+
+        return f"<{type(self).__name__} {self.level_of_theory}>"
+
 
 # the id of a mutable object may change, thus using object()
 _sentinel_msos = object()

{stjames-0.0.42 → stjames-0.0.44}/stjames/workflows/spin_states.py

@@ -49,6 +49,7 @@ class SpinStatesWorkflow(Workflow, MultiStageOptMixin):
 
     Inherited
     :param initial_molecule: Molecule of interest
+    :param mode: Mode for workflow
     :param multistage_opt_settings: set by mode unless mode=MANUAL (ignores additional settings if set)
     :param solvent: solvent to use
     :param xtb_preopt: pre-optimize with xtb (sets based on mode when None)
@@ -60,7 +61,6 @@ class SpinStatesWorkflow(Workflow, MultiStageOptMixin):
     :param mso_mode: Mode for MultiStageOptSettings
 
     New:
-    :param mode: Mode for workflow
     :param states: multiplicities of the spin state targetted
     :param spin_states: resulting spin states data
 
@@ -72,16 +72,12 @@ class SpinStatesWorkflow(Workflow, MultiStageOptMixin):
     '<SpinStatesWorkflow [1, 3, 5] RAPID>'
     """
 
-    mode: Mode
     mso_mode: Mode = _sentinel_mso_mode  # type: ignore [assignment]
     states: list[PositiveInt]
 
     # Results
     spin_states: list[SpinState] = Field(default_factory=list)
 
-    def __str__(self) -> str:
-        return repr(self)
-
     def __repr__(self) -> str:
         if self.mode != Mode.MANUAL:
             return f"<{type(self).__name__} {self.states} {self.mode.name}>"
@@ -121,14 +117,6 @@ class SpinStatesWorkflow(Workflow, MultiStageOptMixin):
         values["mso_mode"] = values["mode"]
         return values
 
-    @field_validator("mode")
-    @classmethod
-    def set_mode_auto(cls, mode: Mode) -> Mode:
-        if mode == Mode.AUTO:
-            return Mode.RAPID
-
-        return mode
-
     @field_validator("spin_states")
     @classmethod
     def validate_spin_states(cls, spin_states: list[SpinState]) -> list[SpinState]:

{stjames-0.0.42 → stjames-0.0.44}/stjames/workflows/workflow.py

@@ -1,4 +1,4 @@
-from pydantic import BaseModel, ConfigDict
+from pydantic import field_validator
 
 from ..base import Base
 from ..message import Message
@@ -8,30 +8,17 @@ from ..types import UUID
 
 
 class Workflow(Base):
-    """All workflows should have these properties."""
-
-    initial_molecule: Molecule
-    messages: list[Message] = []
-
-
-class DBCalculation(Base):
-    """Encodes a calculation that's in the database. This isn't terribly useful by itself."""
-
-    uuid: UUID
-
-
-class WorkflowInput(BaseModel):
     """
-    Input for a workflow.
+    Base class for Workflows.
 
     :param initial_molecule: Molecule of interest
-    :param mode: Mode for workflow
+    :param mode: Mode to use
+    :param messages: messages to display
     """
 
-    model_config = ConfigDict(extra="forbid")
-
     initial_molecule: Molecule
-    mode: Mode
+    mode: Mode = Mode.AUTO
+    messages: list[Message] = []
 
     def __str__(self) -> str:
         return repr(self)
@@ -39,8 +26,17 @@ class WorkflowInput(BaseModel):
     def __repr__(self) -> str:
         return f"<{type(self).__name__} {self.mode.name}>"
 
+    @field_validator("mode")
+    @classmethod
+    def set_mode_auto(cls, mode: Mode) -> Mode:
+        """Set the mode to RAPID if AUTO is selected."""
+        if mode == Mode.AUTO:
+            return Mode.RAPID
+
+        return mode
 
-class WorkflowResults(BaseModel):
-    """Results of a workflow."""
 
-    model_config = ConfigDict(extra="forbid", frozen=True)
+class DBCalculation(Base):
+    """Encodes a calculation that's in the database. This isn't terribly useful by itself."""
+
+    uuid: UUID

{stjames-0.0.42 → stjames-0.0.44/stjames.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: stjames
-Version: 0.0.42
+Version: 0.0.44
 Summary: standardized JSON atom/molecule encoding scheme
 Author-email: Corin Wagen <corin@rowansci.com>
 Project-URL: Homepage, https://github.com/rowansci/stjames

{stjames-0.0.42 → stjames-0.0.44}/stjames.egg-info/SOURCES.txt

@@ -53,4 +53,6 @@ stjames/workflows/scan.py
 stjames/workflows/spin_states.py
 stjames/workflows/tautomer.py
 stjames/workflows/workflow.py
-tests/test_molecule.py
+tests/test_from_extxyz.py
+tests/test_molecule.py
+tests/test_settings.py

stjames-0.0.44/tests/test_from_extxyz.py

@@ -0,0 +1,231 @@
+import pytest
+
+from stjames import Atom, Molecule, MoleculeReadError, PeriodicCell
+
+valid_extxyz = """
+5
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_num_atoms = """
+6
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+not_digit_num_atoms = """
+v
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+many_num_atoms = """
+6 9
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+no_num_atoms = """
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+xyz_style = """
+5
+Comment
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+missing_lattice = """
+5
+Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+missing_properties = """
+5
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0"
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_properites = """
+5
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3foo:1
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_extra = """
+5
+Lattice="6.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 6.0 3.14" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_equals = """
+5
+Lattice="6.0 0.0 =0.0 0.0 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_str = """
+5
+Lattice="6.0 0.0 0.0 hi 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_extra_string = """
+5
+Lattice="6.0 0.0 0.0 0.0 sup 6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+
+incorrect_lattice_single_quote = """
+5
+Lattice="6.0 0.0 0.0 0.0 6.0 '0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_double_quote = """
+5
+Lattice="6.0 0.0 0.0 0.0 "6.0 0.0 0.0 0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_double_single_quote = """
+5
+Lattice="6.0 0.0 0.0 0.0 '6.0 0.0 0.0 '0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+incorrect_lattice_double_double_quote = """
+5
+Lattice="6.0 0.0 "0.0 0.0 6.0 0.0 0.0 "0.0 6.0" Properties=species:S:1:pos:R:3
+C        0.0        0.0        0.0
+H        0.0        0.0        1.0
+H        1.0        0.0        0.0
+H        0.0        1.0        0.0
+H        1.0        1.0        1.0
+"""
+
+
+expected_cell = (
+    (6.0, 0.0, 0.0),
+    (0.0, 6.0, 0.0),
+    (0.0, 0.0, 6.0),
+)
+
+expected_atoms = [
+    Atom(atomic_number=6, position=(0.0, 0.0, 0.0)),  # C
+    Atom(atomic_number=1, position=(0.0, 0.0, 1.0)),  # H
+    Atom(atomic_number=1, position=(1.0, 0.0, 0.0)),  # H
+    Atom(atomic_number=1, position=(0.0, 1.0, 0.0)),  # H
+    Atom(atomic_number=1, position=(1.0, 1.0, 1.0)),  # H
+]
+
+expected_molecule = Molecule(
+    charge=0,
+    multiplicity=1,
+    atoms=expected_atoms,
+    cell=PeriodicCell(lattice_vectors=expected_cell),
+)
+
+
+def test_molecule_from_extxyz_valid() -> None:
+    """
+    Test case for valid extxyz string.
+    """
+    molecule = Molecule.from_extxyz(valid_extxyz)
+    assert molecule == expected_molecule, f"Valid case failed: got {molecule}, expected {expected_molecule}"
+
+
+@pytest.mark.parametrize(
+    "invalid_extxyz",
+    [
+        incorrect_num_atoms,
+        no_num_atoms,
+        not_digit_num_atoms,
+        many_num_atoms,
+        xyz_style,
+        missing_lattice,
+        missing_properties,
+        incorrect_properites,
+        incorrect_lattice_extra,
+        incorrect_lattice_equals,
+        incorrect_lattice_str,
+        incorrect_lattice_extra_string,
+        incorrect_lattice_single_quote,
+        incorrect_lattice_double_quote,
+        incorrect_lattice_double_single_quote,
+        incorrect_lattice_double_double_quote,
+    ],
+)
+def test_molecule_from_extxyz_invalid(invalid_extxyz: str) -> None:
+    """
+    Test case for invalid extxyz strings, ensuring they raise MoleculeReadError.
+    """
+    with pytest.raises(MoleculeReadError):
+        Molecule.from_extxyz(invalid_extxyz)

stjames-0.0.44/tests/test_settings.py

@@ -0,0 +1,34 @@
+from stjames import Constraint, Mode, OptimizationSettings, Settings
+
+
+def test_set_mode_auto() -> None:
+    Settings()
+    assert Settings().mode == Mode.RAPID
+
+
+def test_opt_settings() -> None:
+    settings_rapid = Settings(mode=Mode.RAPID)
+    settings_meticulous = Settings(mode=Mode.METICULOUS)
+
+    cons = [Constraint(atoms=[1, 2], constraint_type="bond")]
+    settings_careful = Settings(mode=Mode.CAREFUL, opt_settings=OptimizationSettings(constraints=cons))
+
+    rap_opt_set = settings_rapid.opt_settings
+    car_opt_set = settings_careful.opt_settings
+    met_opt_set = settings_meticulous.opt_settings
+
+    assert not rap_opt_set.constraints
+    assert not met_opt_set.constraints
+    assert car_opt_set.constraints == cons
+
+    assert rap_opt_set.energy_threshold == 5e-5
+    assert rap_opt_set.max_gradient_threshold == 5e-3
+    assert rap_opt_set.rms_gradient_threshold == 3.5e-3
+
+    assert car_opt_set.energy_threshold == 1e-6
+    assert car_opt_set.max_gradient_threshold == 9e-4
+    assert car_opt_set.rms_gradient_threshold == 6e-4
+
+    assert met_opt_set.energy_threshold == 1e-6
+    assert met_opt_set.max_gradient_threshold == 3e-5
+    assert met_opt_set.rms_gradient_threshold == 2e-5

stjames-0.0.42/stjames/settings.py

@@ -1,209 +0,0 @@
-from typing import Any, Optional, TypeVar
-
-import pydantic
-
-from .base import Base, UniqueList
-from .basis_set import BasisSet
-from .correction import Correction
-from .method import METHODS_WITH_CORRECTION, PREPACKAGED_METHODS, Method
-from .mode import Mode
-from .opt_settings import OptimizationSettings
-from .scf_settings import SCFSettings
-from .solvent import SolventSettings
-from .task import Task
-from .thermochem_settings import ThermochemistrySettings
-
-_T = TypeVar("_T")
-
-
-class Settings(Base):
-    method: Method = Method.HARTREE_FOCK
-    basis_set: Optional[BasisSet] = None
-    tasks: UniqueList[Task] = [Task.ENERGY, Task.CHARGE, Task.DIPOLE]
-    corrections: UniqueList[Correction] = []
-
-    mode: Mode = Mode.AUTO
-
-    solvent_settings: Optional[SolventSettings] = None
-
-    # scf/opt settings will be set automatically based on mode, but can be overridden manually
-    scf_settings: SCFSettings = SCFSettings()
-    opt_settings: OptimizationSettings = OptimizationSettings()
-    thermochem_settings: ThermochemistrySettings = ThermochemistrySettings()
-
-    # mypy has this dead wrong (https://docs.pydantic.dev/2.0/usage/computed_fields/)
-    # Python 3.12 narrows the reason for the ignore to prop-decorator
-    @pydantic.computed_field  # type: ignore[misc, prop-decorator, unused-ignore]
-    @property
-    def level_of_theory(self) -> str:
-        corrections = list(filter(lambda x: x not in (None, ""), self.corrections))
-
-        if self.method in PREPACKAGED_METHODS or self.basis_set is None:
-            method = self.method.value
-        elif self.method in METHODS_WITH_CORRECTION or len(corrections) == 0:
-            method = f"{self.method.value}/{self.basis_set.name.lower()}"
-        else:
-            method = f"{self.method.value}-{'-'.join([c.value for c in corrections])}/{self.basis_set.name.lower()}"
-
-        if self.solvent_settings is not None:
-            method += f"/{self.solvent_settings.model.value}({self.solvent_settings.solvent.value})"
-
-        return method
-
-    def model_post_init(self, __context: Any) -> None:
-        _assign_settings_by_mode(self)
-
-        # figure out `optimize_ts`
-        if Task.OPTIMIZE_TS in self.tasks:
-            self.tasks.pop(self.tasks.index(Task.OPTIMIZE_TS))
-            self.tasks.append(Task.OPTIMIZE)
-            self.opt_settings.transition_state = True
-
-        # composite methods have their own basis sets, so overwrite user stuff
-        if self.method == Method.HF3C:
-            self.basis_set = BasisSet(name="minix")
-        elif self.method == Method.B973C:
-            self.basis_set = BasisSet(name="def2-mTZVP")
-        elif self.method == Method.R2SCAN3C:
-            self.basis_set = BasisSet(name="def2-mTZVPP")
-        elif self.method == Method.WB97X3C:
-            self.basis_set = BasisSet(name="vDZP")
-
-    @pydantic.field_validator("basis_set", mode="before")
-    @classmethod
-    def parse_basis_set(cls, v: Any) -> BasisSet | dict[str, Any] | None:
-        """Turn a string into a ``BasisSet`` object. (This is a little crude.)"""
-        if isinstance(v, BasisSet):
-            return None if v.name is None else v
-        elif isinstance(v, dict):
-            return None if v.get("name") is None else v
-        elif isinstance(v, str):
-            if len(v):
-                return BasisSet(name=v)
-            # "" is basically None, let's be real here...
-            return None
-        elif v is None:
-            return None
-        else:
-            raise ValueError(f"invalid value ``{v}`` for ``basis_set``")
-
-    @pydantic.field_validator("corrections", mode="before")
-    @classmethod
-    def remove_empty_string(cls, v: list[_T]) -> list[_T]:
-        """Remove empty string values."""
-        return [c for c in v if c] if v is not None else v
-
-
-def _assign_settings_by_mode(settings: Settings) -> None:
-    """Modifies ``scf_settings`` and ``opt_settings`` based on preset ``mode``."""
-    mode = settings.mode
-
-    if mode == Mode.AUTO:
-        if (Task.OPTIMIZE in settings.tasks) or (Task.GRADIENT in settings.tasks) or (Task.FREQUENCIES in settings.tasks) or (Task.HESSIAN in settings.tasks):
-            # noisy gradient! struggles to converge
-            if settings.method == Method.AIMNET2_WB97MD3:
-                mode = Mode.RAPID
-            else:
-                mode = Mode.CAREFUL
-        else:
-            mode = Mode.RAPID
-    elif mode == Mode.MANUAL:
-        return
-
-    # modify scf settings!
-    #
-    # values based off of the following sources:
-    # qchem:
-    #   https://manual.q-chem.com/5.2/Ch4.S3.SS2.html
-    #   https://manual.q-chem.com/5.2/Ch4.S5.SS2.html
-    #
-    # gaussian:
-    #   https://gaussian.com/integral/
-    #   https://gaussian.com/overlay5/
-    #
-    # orca:
-    #   manual 4.2.1, §9.6.1 and §9.7.3
-    #
-    # psi4:
-    #   https://psicode.org/psi4manual/master/autodir_options_c/module__scf.html
-    #   https://psicode.org/psi4manual/master/autodoc_glossary_options_c.html
-    #
-    # terachem:
-    #   manual, it's easy to locate everything.
-    #
-    # the below values are my best attempt at homogenizing various sources.
-    # in general, eri_threshold should be 3 OOM lower than scf convergence
-    scf_settings = settings.scf_settings
-    if mode == Mode.RECKLESS:
-        scf_settings.energy_threshold = 1e-5
-        scf_settings.rms_error_threshold = 1e-7
-        scf_settings.max_error_threshold = 1e-5
-        scf_settings.rebuild_frequency = 100
-        scf_settings.int_settings.eri_threshold = 1e-8
-        scf_settings.int_settings.csam_multiplier = 3.0
-        scf_settings.int_settings.pair_overlap_threshold = 1e-8
-    elif mode == Mode.RAPID:
-        scf_settings.energy_threshold = 5e-5
-        scf_settings.rms_error_threshold = 1e-8
-        scf_settings.max_error_threshold = 1e-6
-        scf_settings.rebuild_frequency = 20
-        scf_settings.int_settings.eri_threshold = 1e-9
-        scf_settings.int_settings.csam_multiplier = 1.0
-        scf_settings.int_settings.pair_overlap_threshold = 1e-9
-    elif mode == Mode.CAREFUL:
-        scf_settings.energy_threshold = 1e-6
-        scf_settings.rms_error_threshold = 1e-9
-        scf_settings.max_error_threshold = 1e-7
-        scf_settings.rebuild_frequency = 10
-        scf_settings.int_settings.eri_threshold = 1e-10
-        scf_settings.int_settings.csam_multiplier = 1.0
-        scf_settings.int_settings.pair_overlap_threshold = 1e-10
-    elif mode == Mode.METICULOUS:
-        scf_settings.energy_threshold = 1e-8
-        scf_settings.rms_error_threshold = 1e-9
-        scf_settings.max_error_threshold = 1e-7
-        scf_settings.rebuild_frequency = 5
-        scf_settings.int_settings.eri_threshold = 1e-12
-        scf_settings.int_settings.csam_multiplier = 1.0
-        scf_settings.int_settings.pair_overlap_threshold = 1e-12
-    elif mode == Mode.DEBUG:
-        scf_settings.energy_threshold = 1e-9
-        scf_settings.rms_error_threshold = 1e-10
-        scf_settings.max_error_threshold = 1e-9
-        scf_settings.rebuild_frequency = 1
-        scf_settings.int_settings.eri_threshold = 1e-14
-        scf_settings.int_settings.csam_multiplier = 1e10  # in other words, disable CSAM
-        scf_settings.int_settings.pair_overlap_threshold = 1e-14
-    else:
-        raise ValueError(f"Unknown mode ``{mode.value}``!")
-
-    opt_settings = settings.opt_settings
-
-    # constrained optimizations warrant loosening the settings a bit
-    has_constraints = len(opt_settings.constraints) > 0
-
-    # cf. DLFIND manual, and https://www.cup.uni-muenchen.de/ch/compchem/geom/basic.html
-    # and the discussion at https://geometric.readthedocs.io/en/latest/how-it-works.html
-    # in periodic systems, "normal" is 0.05 eV/Å ~= 2e-3 Hartree/Å, and "careful" is 0.01 ~= 4e-4
-    if mode == Mode.RECKLESS:
-        opt_settings.energy_threshold = 2e-5
-        opt_settings.max_gradient_threshold = 7e-3
-        opt_settings.rms_gradient_threshold = 6e-3
-    elif mode == Mode.RAPID or (mode == Mode.CAREFUL and has_constraints):
-        opt_settings.energy_threshold = 5e-5
-        opt_settings.max_gradient_threshold = 5e-3
-        opt_settings.rms_gradient_threshold = 3.5e-3
-    elif mode == Mode.CAREFUL or (mode == Mode.METICULOUS and has_constraints):
-        opt_settings.energy_threshold = 1e-6
-        opt_settings.max_gradient_threshold = 9e-4
-        opt_settings.rms_gradient_threshold = 6e-4
-    elif mode == Mode.METICULOUS:
-        opt_settings.energy_threshold = 1e-6
-        opt_settings.max_gradient_threshold = 3e-5
-        opt_settings.rms_gradient_threshold = 2e-5
-    elif mode == Mode.DEBUG:
-        opt_settings.energy_threshold = 1e-6
-        opt_settings.max_gradient_threshold = 4e-6
-        opt_settings.rms_gradient_threshold = 2e-6
-    else:
-        raise ValueError(f"Unknown mode ``{mode.value}``!")