jebus/enviPy-bayer
1
0
Fork 0
forked from enviPath/enviPy
Code Issues Pull Requests 1 Actions Releases Activity

[Feature] Generate Compounds by Molfile (#423)

Browse Source 
Co-authored-by: Tim Lorsbach <tim@lorsba.ch>
Reviewed-on: enviPath/enviPy#423
This commit is contained in:
jebus
2026-07-15 23:56:24 +12:00
parent 9bc9f86ff1
commit 2c2437e3f5
8 changed files with 439 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

148
epdb/models.py
View File

@ -31,7 +31,10 @@ from sklearn.model_selection import ShuffleSplit
from bridge.contracts import Property
from bridge.dto import RunResult, PropertyPrediction
from epdb.exceptions import InvalidSMILESException
from epdb.exceptions import (
InvalidMolfileException,
InvalidSMILESException,
)
from utilities.chem import FormatConverter, IndigoUtils, PredictionResult, ProductSet
from utilities.ml import (
ApplicabilityDomainPCA,
@ -789,6 +792,13 @@ class Compound(
external_identifiers = GenericRelation("ExternalIdentifier")
def get_structure_by_smiles(self, smiles: str) -> "CompoundStructure":
for struct in self.structures.all():
if struct.smiles == smiles:
return struct
raise ValueError(f"No structure with SMILES {smiles} found for {self.get_name()}")
@property
def structures(self) -> QuerySet:
return CompoundStructure.objects.filter(compound=self)
@ -857,8 +867,24 @@ class Compound(
@staticmethod
@transaction.atomic
def create(
package: "Package", smiles: str, name: str = None, description: str = None, *args, **kwargs
package: "Package",
smiles: str,
molfile: str | None = None,
name: str | None = None,
description: str | None = None,
*args,
**kwargs,
) -> "Compound":
# Molfile has precendence over SMILES
if molfile is not None and molfile.strip() != "":
mol = FormatConverter.from_molfile(molfile)
if mol is None:
raise InvalidMolfileException("Given molfile is invalid")
else:
# Overwrite SMILES from molfile
smiles = FormatConverter.to_smiles(mol)
if smiles is None or smiles.strip() == "":
raise InvalidSMILESException("SMILES is required")
@ -878,11 +904,16 @@ class Compound(
if CompoundStructure.objects.filter(
smiles=standardized_smiles, compound__package=package
).exists():
# TODO should we add a structure?
return CompoundStructure.objects.get(
found_c = CompoundStructure.objects.get(
smiles=standardized_smiles, compound__package=package
).compound
# As a direct match wasn't found add a new structure
# TODO return newly created structure
_ = found_c.add_structure(smiles, molfile=molfile, name=name, description=description)
return found_c
# Generate Compound
c = Compound()
c.package = package
@ -914,7 +945,12 @@ class Compound(
)
cs = CompoundStructure.create(
c, smiles, name=name, description=description, normalized_structure=is_standardized
c,
smiles,
molfile=molfile,
name=name,
description=description,
normalized_structure=is_standardized,
)
c.default_structure = cs
@ -927,11 +963,22 @@ class Compound(
self,
smiles: str,
name: str = None,
molfile: str = None,
description: str = None,
default_structure: bool = False,
*args,
**kwargs,
) -> "CompoundStructure":
# Molfile has precendence over SMILES
if molfile is not None and molfile.strip() != "":
mol = FormatConverter.from_molfile(molfile)
if mol is None:
raise InvalidMolfileException("Given molfile is invalid")
else:
# Overwrite SMILES from molfile
smiles = FormatConverter.to_smiles(mol)
if smiles is None or smiles == "":
raise ValueError("SMILES is required")
@ -951,16 +998,28 @@ class Compound(
)
if is_standardized:
CompoundStructure.objects.get(smiles__in=smiles, compound__package=self.package)
CompoundStructure.objects.get(smiles=smiles, compound__package=self.package)
# Check if we find a direct match for a given SMILES and/or its standardized SMILES
if CompoundStructure.objects.filter(
smiles__in=smiles, compound__package=self.package
).exists():
return CompoundStructure.objects.get(smiles__in=smiles, compound__package=self.package)
if CompoundStructure.objects.filter(smiles=smiles, compound__package=self.package).exists():
found_cs = CompoundStructure.objects.get(smiles=smiles, compound__package=self.package)
if found_cs.molfile is None and (molfile is not None and molfile.strip() != ""):
logger.info(
f"Setting molfile for found CompoundStructure(uuid={found_cs.uuid}) as it was empty"
)
found_cs.molfile = molfile
found_cs.save()
return found_cs
cs = CompoundStructure.create(
self, smiles, name=name, description=description, normalized_structure=is_standardized
self,
smiles,
name=name,
molfile=molfile,
description=description,
normalized_structure=is_standardized,
)
if default_structure:
@ -1141,10 +1200,35 @@ class CompoundStructure(
@staticmethod
@transaction.atomic
def create(
compound: Compound, smiles: str, name: str = None, description: str = None, *args, **kwargs
compound: Compound,
smiles: str,
molfile: str = None,
name: str = None,
description: str = None,
*args,
**kwargs,
):
# Molfile has precendence over SMILES
if molfile is not None and molfile.strip() != "":
mol = FormatConverter.from_molfile(molfile)
if mol is None:
raise InvalidMolfileException("Given molfile is invalid")
else:
# Overwrite SMILES from molfile
smiles = FormatConverter.to_smiles(mol)
if CompoundStructure.objects.filter(compound=compound, smiles=smiles).exists():
return CompoundStructure.objects.get(compound=compound, smiles=smiles)
found_cs = CompoundStructure.objects.get(smiles=smiles, compound=compound)
if found_cs.molfile is None and (molfile is not None and molfile.strip() != ""):
logger.info(
f"Setting molfile for found CompoundStructure(uuid={found_cs.uuid}) as it was empty"
)
found_cs.molfile = molfile
found_cs.save()
return found_cs
if compound.pk is None:
raise ValueError("Unpersisted Compound! Persist compound first!")
@ -2174,11 +2258,12 @@ class Pathway(EnviPathModel, AliasMixin, ScenarioMixin, AdditionalInformationMix
def add_node(
self,
smiles: str,
name: Optional[str] = None,
description: Optional[str] = None,
depth: Optional[int] = -1,
molfile: str | None = None,
name: str | None = None,
description: str | None = None,
depth: int = -1,
):
return Node.create(self, smiles, depth, name=name, description=description)
return Node.create(self, smiles, depth, molfile=molfile, name=name, description=description)
@transaction.atomic
def add_edge(
@ -2336,28 +2421,43 @@ class Node(EnviPathModel, AliasMixin, ScenarioMixin, AdditionalInformationMixin)
pathway: "Pathway",
smiles: str,
depth: int,
name: Optional[str] = None,
description: Optional[str] = None,
molfile: str | None = None,
name: str | None = None,
description: str | None = None,
):
# Molfile has precendence over SMILES
if molfile is not None and molfile.strip() != "":
mol = FormatConverter.from_molfile(molfile)
if mol is None:
raise InvalidMolfileException("Given molfile is invalid")
else:
# Overwrite SMILES from molfile
smiles = FormatConverter.to_smiles(mol)
stereo_removed = False
if pathway.predicted and FormatConverter.has_stereo(smiles):
smiles = FormatConverter.standardize(smiles, remove_stereo=True)
stereo_removed = True
c = Compound.create(pathway.package, smiles, name=name, description=description)
c = Compound.create(
pathway.package, smiles, molfile=molfile, name=name, description=description
)
if Node.objects.filter(pathway=pathway, default_node_label=c.default_structure).exists():
return Node.objects.get(pathway=pathway, default_node_label=c.default_structure)
structure = c.get_structure_by_smiles(smiles)
if Node.objects.filter(pathway=pathway, default_node_label=structure).exists():
return Node.objects.get(pathway=pathway, default_node_label=structure)
n = Node()
n.stereo_removed = stereo_removed
n.pathway = pathway
n.depth = depth
n.default_node_label = c.default_structure
n.default_node_label = structure
n.save()
n.node_labels.add(c.default_structure)
n.node_labels.add(structure)
n.save()
return n