stjames 0.0.92__tar.gz → 0.0.94__tar.gz

{stjames-0.0.92/stjames.egg-info → stjames-0.0.94}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: stjames
-Version: 0.0.92
+Version: 0.0.94
 Summary: standardized JSON atom/molecule encoding scheme
 Author-email: Corin Wagen <corin@rowansci.com>
 Project-URL: Homepage, https://github.com/rowansci/stjames

{stjames-0.0.92 → stjames-0.0.94}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "stjames"
-version = "0.0.92"
+version = "0.0.94"
 description = "standardized JSON atom/molecule encoding scheme"
 readme = "README.md"
 requires-python = ">=3.11"

{stjames-0.0.92 → stjames-0.0.94}/stjames/__init__.py

@@ -7,12 +7,9 @@ from .molecule import *
 from .workflows import *
 
 from .scf_settings import *
-from .int_settings import *
 from .opt_settings import *
 from .compute_settings import *
-from .diis_settings import *
 from .thermochem_settings import *
-from .grid_settings import *
 from .settings import *
 
 from .method import *

{stjames-0.0.92 → stjames-0.0.94}/stjames/pdb.py

@@ -448,6 +448,7 @@ def _format_atom_line(
         chg = "  "
 
     atom_name = atom.name if atom.name else atom.element
+
     occupancy = atom.occupancy if atom.occupancy else 1.0
 
     # Construct the line.
@@ -455,7 +456,7 @@ def _format_atom_line(
     line = (
         f"{record_type}"
         f"{serial:5d} "  # atom serial number (columns 7-11)
-        f"{atom_name:<4}"  # atom name (columns 13-16, left-justified in this snippet)
+        f"{(' ' + atom_name if len(atom_name) < 4 else atom_name):<4}"  # atom name (columns 13-16, left-justified in this snippet)
         f"{alt_loc_char}"  # altLoc (column 17)
         f"{residue_name:>3}"  # residue name (columns 18-20)
         f" {chain_char}"  # chain ID (column 22)

stjames-0.0.94/stjames/scf_settings.py

@@ -0,0 +1,7 @@
+from pydantic import BaseModel
+
+
+class SCFSettings(BaseModel):
+    max_iters: int = 250
+
+    soscf: bool = False

{stjames-0.0.92 → stjames-0.0.94}/stjames/settings.py

@@ -68,7 +68,6 @@ class Settings(Base):
         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
@@ -115,74 +114,6 @@ class Settings(Base):
         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.

{stjames-0.0.92 → stjames-0.0.94}/stjames/status.py

@@ -19,3 +19,6 @@ class Status(int, Enum):
 
     # Status if a user has exceeded their allowed max_concurrency
     AWAITING_QUEUE = 5
+
+    # Job is not yet submitted
+    DRAFT = 6

{stjames-0.0.92 → stjames-0.0.94}/stjames/workflows/docking.py

@@ -7,6 +7,7 @@ from pydantic import AfterValidator, ConfigDict, field_validator, model_validato
 from ..base import Base, round_float
 from ..pdb import PDB
 from ..types import UUID, Vector3D
+from .conformer_search import ConformerGenSettings, ETKDGSettings
 from .workflow import MoleculeWorkflow
 
 
@@ -20,6 +21,28 @@ class Score(Base):
 
     pose: UUID | None  # for calculation
     score: Annotated[float, AfterValidator(round_float(3))]
+    posebusters_valid: bool
+
+
+class DockingSettings(Base):
+    """
+    Base class for controlling how docked poses are generated.
+
+    :param max_poses: the maximum number of poses generated per input molecule
+    """
+
+    max_poses: int = 4
+
+
+class VinaSettings(DockingSettings):
+    """
+    Controls how AutoDock Vina is run.
+
+    :param exhaustiveness: how many times Vina attempts to find a pose.
+        8 is typical, 32 is considered relatively careful.
+    """
+
+    exhaustiveness: int = 8
 
 
 class DockingWorkflow(MoleculeWorkflow):
@@ -35,31 +58,38 @@ class DockingWorkflow(MoleculeWorkflow):
     :param mode: Mode for workflow (currently unused)
 
     New:
-    :param molecules: Molecules to dock (optional)
-    :param smiles: SMILES strings of the ligands (optional)
+    :param docking_engine: which docking method to use
     :param do_csearch: whether to csearch starting structures
+    :param csearch_settings: settings for initial conformer search.
     :param do_optimization: whether to optimize starting structures
+    :param opt_settings: settings for conformer optimization.
     :param do_pose_refinement: whether to optimize non-rotatable bonds in output poses
-    :param conformers: UUIDs of optimized conformers
     :param target: PDB of the protein.
     :param target_uuid: UUID of the protein.
     :param pocket: center (x, y, z) and size (x, y, z) of the pocket
 
     Results:
+    :param conformers: UUIDs of optimized conformers
     :param scores: docked poses sorted by score
     """
 
     model_config = ConfigDict(arbitrary_types_allowed=True)
 
-    do_csearch: bool = True
-    do_optimization: bool = True
-    conformers: list[UUID] = []
-
     target: PDB | None = None
     target_uuid: UUID | None = None
     pocket: tuple[Vector3D, Vector3D]
 
+    do_csearch: bool = True
+    conformer_gen_settings: ConformerGenSettings = ETKDGSettings(mode="reckless")
+
+    do_optimization: bool = True
+    # optimization_settings - here in future once we have a cleaner mode sol'n, ccw 7.9.25
+
+    docking_settings: DockingSettings = VinaSettings()
+
     do_pose_refinement: bool = True
+
+    conformers: list[UUID] = []
     scores: list[Score] = []
 
     def __str__(self) -> str:
@@ -80,7 +110,7 @@ class DockingWorkflow(MoleculeWorkflow):
     def check_protein(self) -> Self:
         """Check if protein is provided."""
         if not self.target and not self.target_uuid:
-            raise ValueError("Must provide either molecules or smiles")
+            raise ValueError("Must provide either target or target_uuid")
         return self
 
     @field_validator("pocket", mode="after")

{stjames-0.0.92 → stjames-0.0.94}/stjames/workflows/protein_cofolding.py

@@ -1,6 +1,6 @@
 """Protein Cofolding Workflow."""
 
-from typing import Annotated
+from typing import Annotated, Literal
 
 from pydantic import AfterValidator, BaseModel, ConfigDict
 
@@ -17,6 +17,33 @@ class CofoldingModel(LowercaseStrEnum):
     BOLTZ_2 = "boltz_2"
 
 
+class Token(BaseModel):
+    """Either a atom in a ligand or a residue in a protein."""
+
+    input_type: Literal["ligand", "protein"]
+    input_index: int
+    token_index: int
+
+
+class ContactConstraint(BaseModel):
+    """Contact constraint to be used for prediction."""
+
+    token_1: Token
+    token_2: Token
+    max_distance: float  # Angstroms
+    force: bool = False  # Whether to use potentials to enforce the constraint
+
+
+class PocketConstraint(BaseModel):
+    """Pocket constraint to be used for prediction."""
+
+    input_type: Literal["ligand", "protein"]
+    input_index: int
+    contacts: list[Token]
+    max_distance: float  # Angstroms
+    force: bool = False  # Whether to use potentials to enforce the constraint
+
+
 class CofoldingScores(BaseModel):
     confidence_score: Annotated[float, AfterValidator(round_float(3))]
     ptm: Annotated[float, AfterValidator(round_float(3))]  # predicted template modeling score
@@ -52,6 +79,9 @@ class ProteinCofoldingWorkflow(FASTAWorkflow):
     use_msa_server: bool = False
     use_templates_server: bool = False
     use_potentials: bool = False
+    contact_constraints: list[ContactConstraint] = []
+    pocket_constraints: list[PocketConstraint] = []
+
     predicted_structure_uuid: UUID | None = None
     scores: CofoldingScores | None = None
     model: CofoldingModel = CofoldingModel.BOLTZ_2

{stjames-0.0.92 → stjames-0.0.94/stjames.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: stjames
-Version: 0.0.92
+Version: 0.0.94
 Summary: standardized JSON atom/molecule encoding scheme
 Author-email: Corin Wagen <corin@rowansci.com>
 Project-URL: Homepage, https://github.com/rowansci/stjames

{stjames-0.0.92 → stjames-0.0.94}/stjames.egg-info/SOURCES.txt

@@ -10,10 +10,7 @@ stjames/calculation.py
 stjames/compute_settings.py
 stjames/constraint.py
 stjames/correction.py
-stjames/diis_settings.py
 stjames/engine.py
-stjames/grid_settings.py
-stjames/int_settings.py
 stjames/message.py
 stjames/method.py
 stjames/mode.py

stjames-0.0.92/stjames/diis_settings.py

@@ -1,23 +0,0 @@
-import pydantic
-
-from .base import Base, LowercaseStrEnum
-
-
-class DIISStrategy(LowercaseStrEnum):
-    # regular Pulay DIIS
-    DIIS = "diis"
-
-    # Hu/Yang JCP 2010 - ADIIS
-    ADIIS = "adiis"
-
-    # first ADIIS, then DIIS
-    ADIIS_DIIS = "adiis_diis"
-
-
-class DIISSettings(Base):
-    strategy: DIISStrategy = DIISStrategy.ADIIS_DIIS
-    subspace_size: pydantic.PositiveInt = 12
-
-    # if it's a hybrid strategy, where do we transition?
-    adiis_diis_blend_start: pydantic.PositiveFloat = 1e-1
-    adiis_diis_blend_stop: pydantic.PositiveFloat = 1e-4

stjames-0.0.92/stjames/grid_settings.py

@@ -1,27 +0,0 @@
-import pydantic
-
-from .base import Base, LowercaseStrEnum
-
-
-class RadialGridType(LowercaseStrEnum):
-    """What sort of radial grid (only one option for now)"""
-
-    KRACK_KOSTER = "krack_koster"
-    LMG = "lmg"
-
-
-class GridSettings(Base):
-    radial_grid_type: RadialGridType = RadialGridType.LMG
-    angular_num_points: pydantic.PositiveInt = 434
-
-    weight_cutoff: pydantic.PositiveFloat = 1e-11
-
-    # for LMG
-    radial_precision: pydantic.PositiveFloat = 1e-11
-
-    # for other schemes, like KK
-    radial_num_points: pydantic.PositiveInt = 75
-
-    # pruning?
-    prune: bool = True
-    min_angular_points: pydantic.PositiveInt = 50

stjames-0.0.92/stjames/int_settings.py

@@ -1,31 +0,0 @@
-import pydantic
-
-from .base import Base, LowercaseStrEnum
-
-
-class ERIStrategy(LowercaseStrEnum):
-    # direct SCF, as per Almlof/Faegri/Korsell
-    DIRECT = "direct"
-
-    # Raffinetti-style supermatrix
-    SUPERMATRIX = "supermatrix"
-
-    # let the software choose between direct and supermatrix based on size
-    AUTO = "auto"
-
-    # resolution of the identity for J and K
-    RIJK = "rijk"
-
-
-class IntSettings(Base):
-    strategy: ERIStrategy = ERIStrategy.AUTO
-
-    # these will get overwritten by ``mode`` anyway, for the most part
-    eri_threshold: pydantic.PositiveFloat = 1e-9
-    csam_multiplier: pydantic.PositiveFloat = pydantic.Field(default=1, ge=1)
-    pair_overlap_threshold: pydantic.PositiveFloat = 1e-10
-
-    @property
-    def resolution_of_the_identity(self) -> bool:
-        """Abstracting in case we add RIJCOSX, etc later"""
-        return self.strategy == ERIStrategy.RIJK

stjames-0.0.92/stjames/scf_settings.py

@@ -1,77 +0,0 @@
-from typing import Any, Optional
-
-import pydantic
-from pydantic import PositiveFloat, PositiveInt
-
-from .base import Base, LowercaseStrEnum
-from .diis_settings import DIISSettings
-from .grid_settings import GridSettings
-from .int_settings import IntSettings
-
-
-class SCFInitMethod(LowercaseStrEnum):
-    # (See https://manual.q-chem.com/5.2/Ch4.S4.SS2.html for a nice overview here.)
-    SAD = "sad"
-    CORE = "core"
-    #    GWH = "gwh"
-    READ = "read"
-
-
-class OrthonormalizationMethod(LowercaseStrEnum):
-    SYMMETRIC = "symmetric"
-    CANONICAL = "canonical"
-    # Cholesky, in future?
-
-
-class SCFSettings(Base):
-    max_iters: int = 100
-    init_method: SCFInitMethod = SCFInitMethod.SAD
-
-    int_settings: IntSettings = IntSettings()
-    grid_settings: GridSettings = GridSettings()
-    diis_settings: DIISSettings = DIISSettings()
-
-    #### orthonormalization
-    orthonormalization: OrthonormalizationMethod = OrthonormalizationMethod.CANONICAL
-
-    #### damping
-    do_damping: bool = True
-    # when should we stop damping?
-    end_damping_error: PositiveFloat = 0.1
-    # what damping factor should we use?
-    damping_factor: float = pydantic.Field(ge=0, le=1, default=0.7)
-
-    #### level shifting
-    do_level_shift: bool = True
-    # how much? (Eh)
-    level_shift_magnitude: PositiveFloat = 0.25
-    # when should we stop?
-    end_level_shift_error: PositiveFloat = 0.1
-
-    #### incremental
-    # do incremental fock build?
-    do_incremental: bool = True
-    # reset incremental fock build
-    rebuild_frequency: PositiveInt = 20
-
-    #### when are we converged?
-    energy_threshold: PositiveFloat = 1e-6
-    rms_error_threshold: PositiveFloat = 1e-8
-    max_error_threshold: PositiveFloat = 1e-5
-
-    #### DIIS
-    do_diis: bool = True
-    # error below which we'll start DIIS
-    start_diis_max_error: pydantic.PositiveFloat = 0.2
-    # first iteration we'll consider starting DIIS
-    start_diis_iter: PositiveInt = 3
-    # iteration past which we'll start DIIS even if error is high
-    start_diis_anyway: PositiveInt = 7
-
-    # if ``read`` initialization is selected
-    initial_density_matrix_guess: Optional[list[list[float]]] = None
-
-    def model_post_init(self, __context: Any) -> None:
-        # disable incremental Fock for RI
-        if self.int_settings.resolution_of_the_identity:
-            self.do_incremental = False