fix!: community models should not inherit from Statement / EvidenceLine (#22)

close #21

* Use model validator instead of inheriting from `Statement` or
`EvidenceLine`
* `VariantOncogenicityFunctionalImpactEvidenceLine` and
`VariantPathogenicityFunctionalImpactEvidenceLine` should NOT inherit
from `EvidenceLine`
* `VariantOncogenicityStudyStatement`, `VariantPathogenicityStatement`,
`VariantDiagnosticStudyStatement`, `VariantPrognosticStudyStatement`,
and `VariantTherapeuticResponseStudyStatement` should NOT inherit from
`Statement`

Author: korikuzma

pyproject.toml

@@ -32,7 +32,7 @@ requires-python = ">=3.10"
 dynamic = ["version"]
 dependencies = [
     "ga4gh.vrs==2.*",
-    "ga4gh.cat_vrs~=0.4.0",
+    "ga4gh.cat_vrs~=0.5.0",
     "pydantic==2.*"
 ]
 

src/ga4gh/va_spec/aac_2017/models.py

@@ -9,14 +9,15 @@
 from ga4gh.core.models import MappableConcept, iriReference
 from ga4gh.va_spec.base.core import (
     Method,
-    Statement,
+    StatementValidatorMixin,
     VariantDiagnosticProposition,
     VariantPrognosticProposition,
     VariantTherapeuticResponseProposition,
 )
 from ga4gh.va_spec.base.enums import System
 from ga4gh.va_spec.base.validators import validate_mappable_concept
 from pydantic import (
+    BaseModel,
     Field,
     field_validator,
 )
@@ -46,8 +47,11 @@ class Classification(str, Enum):
 AMP_ASCO_CAP_TIERS = [v.value for v in Classification.__members__.values()]
 
 
-class _ValidatorMixin:
-    """Mixin class for reusable AMP/ASCO/CAP field validators"""
+class AmpAscoCapValidatorMixin(StatementValidatorMixin):
+    """Mixin class for reusable AMP/ASCO/CAP field validators
+
+    Should be used with classes that inherit from Pydantic BaseModel
+    """
 
     @field_validator("strength")
     @classmethod
@@ -74,7 +78,7 @@ def validate_classification(cls, v: MappableConcept) -> MappableConcept:
         return validate_mappable_concept(v, System.AMP_ASCO_CAP, AMP_ASCO_CAP_TIERS)
 
 
-class VariantDiagnosticStudyStatement(Statement, _ValidatorMixin):
+class VariantDiagnosticStudyStatement(BaseModel, AmpAscoCapValidatorMixin):
     """A statement reporting a conclusion from a single study about whether a variant is
     associated with a disease (a diagnostic inclusion criterion), or absence of a
     disease (diagnostic exclusion criterion) - based on interpretation of the study's
@@ -99,7 +103,7 @@ class VariantDiagnosticStudyStatement(Statement, _ValidatorMixin):
     )
 
 
-class VariantPrognosticStudyStatement(Statement, _ValidatorMixin):
+class VariantPrognosticStudyStatement(BaseModel, AmpAscoCapValidatorMixin):
     """A statement reporting a conclusion from a single study about whether a variant is
     associated with a disease prognosis - based on interpretation of the study's
     results.
@@ -123,7 +127,7 @@ class VariantPrognosticStudyStatement(Statement, _ValidatorMixin):
     )
 
 
-class VariantTherapeuticResponseStudyStatement(Statement, _ValidatorMixin):
+class VariantTherapeuticResponseStudyStatement(BaseModel, AmpAscoCapValidatorMixin):
     """A statement reporting a conclusion from a single study about whether a variant is
     associated with a therapeutic response (positive or negative) - based on
     interpretation of the study's results.

src/ga4gh/va_spec/acmg_2015/models.py

@@ -7,9 +7,9 @@
 
 from ga4gh.core.models import MappableConcept, iriReference
 from ga4gh.va_spec.base.core import (
-    EvidenceLine,
+    EvidenceLineValidatorMixin,
     Method,
-    Statement,
+    StatementValidatorMixin,
     VariantPathogenicityProposition,
 )
 from ga4gh.va_spec.base.enums import (
@@ -18,8 +18,10 @@
     STRENGTHS,
     System,
 )
-from ga4gh.va_spec.base.validators import validate_mappable_concept
-from pydantic import Field, field_validator
+from ga4gh.va_spec.base.validators import (
+    validate_mappable_concept,
+)
+from pydantic import BaseModel, Field, field_validator, model_validator
 
 
 class EvidenceOutcome(str, Enum):
@@ -51,7 +53,9 @@ class AcmgClassification(str, Enum):
 ACMG_CLASSIFICATIONS = [v.value for v in AcmgClassification.__members__.values()]
 
 
-class VariantPathogenicityFunctionalImpactEvidenceLine(EvidenceLine):
+class VariantPathogenicityFunctionalImpactEvidenceLine(
+    BaseModel, EvidenceLineValidatorMixin
+):
     """An Evidence Line that describes how information about the functional impact of a
     variant on a gene or gene product was interpreted as evidence for or against the
     variant's pathogenicity.
@@ -100,23 +104,21 @@ def validate_specified_by(cls, v: Method | iriReference) -> Method | iriReferenc
 
         return v
 
-    @field_validator("evidenceOutcome")
-    @classmethod
-    def validate_evidence_outcome(
-        cls, v: MappableConcept | None
-    ) -> MappableConcept | None:
-        """Validate evidenceOutcome
+    @model_validator(mode="before")
+    def validate_evidence_outcome(cls, values: dict) -> dict:  # noqa: N805
+        """Validate ``evidenceOutcome`` property if it exists
 
-        :param v: evidenceOutcome
-        :raises ValueError: If invalid evidenceOutcome values are provided
-        :return: Validated evidenceOutcome value
+        :param values: Input values
+        :raises ValueError: If ``evidenceOutcome`` exists and is invalid
+        :return: Validated input values. If ``evidenceOutcome`` exists, then it will be
+            validated and converted to a ``MappableConcept``
         """
-        return validate_mappable_concept(
-            v, System.ACMG, EVIDENCE_OUTCOME_VALUES, mc_is_required=False
+        return cls._validate_evidence_outcome(
+            values, System.ACMG, EVIDENCE_OUTCOME_VALUES
         )
 
 
-class VariantPathogenicityStatement(Statement):
+class VariantPathogenicityStatement(BaseModel, StatementValidatorMixin):
     """A Statement describing the role of a variant in causing an inherited condition."""
 
     proposition: VariantPathogenicityProposition | None = Field(
@@ -162,9 +164,10 @@ def validate_classification(cls, v: MappableConcept) -> MappableConcept:
             err_msg = "`primaryCoding` is required."
             raise ValueError(err_msg)
 
-        supported_systems = [System.ACMG.value, System.ACMG.value]
+        supported_systems = [System.ACMG.value, System.CLIN_GEN.value]
         if v.primaryCoding.system not in supported_systems:
             err_msg = f"`primaryCoding.system` must be one of: {supported_systems}."
+            raise ValueError(err_msg)
 
         if v.primaryCoding.system == System.ACMG:
             if v.primaryCoding.code.root not in ACMG_CLASSIFICATIONS:

src/ga4gh/va_spec/base/__init__.py

@@ -27,9 +27,11 @@
 )
 from .domain_entities import Condition, ConditionSet, Therapeutic, TherapyGroup
 from .enums import (
+    CCV_CLASSIFICATIONS,
     CLIN_GEN_CLASSIFICATIONS,
     STRENGTH_OF_EVIDENCE_PROVIDED_VALUES,
     STRENGTHS,
+    CcvClassification,
     ClinGenClassification,
     DiagnosticPredicate,
     MembershipOperator,
@@ -42,7 +44,9 @@
 
 __all__ = [
     "Agent",
+    "CCV_CLASSIFICATIONS",
     "CLIN_GEN_CLASSIFICATIONS",
+    "CcvClassification",
     "ClinGenClassification",
     "ClinGenClassification",
     "ClinicalVariantProposition",

src/ga4gh/va_spec/base/core.py

@@ -20,19 +20,24 @@
 from ga4gh.va_spec.base.enums import (
     DiagnosticPredicate,
     PrognosticPredicate,
+    System,
     TherapeuticResponsePredicate,
 )
+from ga4gh.va_spec.base.validators import validate_mappable_concept
 from ga4gh.vrs.models import Allele, MolecularVariation
 from pydantic import (
+    BaseModel,
     ConfigDict,
     Field,
     RootModel,
     StringConstraints,
     ValidationError,
     field_validator,
+    model_validator,
 )
 
-StatementType = TypeVar("StatementType", bound="Statement")
+StatementType = TypeVar("StatementType")
+EvidenceLineType = TypeVar("EvidenceLineType")
 
 #########################################
 # Abstract Core Classes
@@ -484,7 +489,7 @@ class EvidenceLine(InformationEntity, BaseModelForbidExtra):
         description="The possible fact against which evidence items contained in an Evidence Line were collectively evaluated, in determining the overall strength and direction of support they provide. For example, in an ACMG Guideline-based assessment of variant pathogenicity, the support provided by distinct lines of evidence are assessed against a target proposition that the variant is pathogenic for a specific disease.",
     )
     hasEvidenceItems: (
-        list[StudyResult | StatementType | EvidenceLine | iriReference] | None
+        list[StudyResult | StatementType | EvidenceLineType | iriReference] | None
     ) = Field(
         None,
         description="An individual piece of information that was evaluated as evidence in building the argument represented by an Evidence Line.",
@@ -509,7 +514,7 @@ class EvidenceLine(InformationEntity, BaseModelForbidExtra):
     @field_validator("hasEvidenceItems", mode="before")
     def validate_has_evidence_items(
         cls,  # noqa: N805
-        v: list[StudyResult, StatementType, EvidenceLine, iriReference] | None,
+        v: list | None,
     ) -> list | None:
         """Ensure hasEvidenceItems is correct type
 
@@ -539,15 +544,12 @@ def validate_has_evidence_items(
                     obj_
                     for _, obj_ in vars(imported_module).items()
                     if inspect.isclass(obj_)
-                    and issubclass(obj_, Statement)
-                    and obj_ is not Statement
+                    and issubclass(obj_, BaseModel)
+                    and obj_.__name__.endswith(("Statement", "EvidenceLine"))
                 ]
             )
 
-        has_evidence_items_models.extend(
-            [Statement, StudyResult, EvidenceLine, iriReference]
-        )
-
+        has_evidence_items_models.extend([Statement, StudyResult, EvidenceLine])
         for evidence_item in v:
             if isinstance(evidence_item, dict):
                 found_model = False
@@ -561,10 +563,13 @@ def validate_has_evidence_items(
                         found_model = True
                         break
                 if not found_model:
-                    err_msg = "Unable to find valid model"
+                    err_msg = "Unable to find valid model for `hasEvidenceItems`"
                     raise ValueError(err_msg)
+            elif isinstance(evidence_item, str):
+                evidence_items.append(iriReference(root=evidence_item))
             else:
-                evidence_items.append(evidence_item)
+                err_msg = "Unable to find valid model for `hasEvidenceItems`"
+                raise ValueError(err_msg)
         return evidence_items
 
 
@@ -580,9 +585,17 @@ class Statement(InformationEntity, BaseModelForbidExtra):
     type: Literal["Statement"] = Field(
         CoreType.STATEMENT.value, description=f"MUST be '{CoreType.STATEMENT.value}'."
     )
-    proposition: Proposition = Field(
+    proposition: (
+        ExperimentalVariantFunctionalImpactProposition
+        | VariantDiagnosticProposition
+        | VariantOncogenicityProposition
+        | VariantPathogenicityProposition
+        | VariantPrognosticProposition
+        | VariantTherapeuticResponseProposition
+    ) = Field(
         ...,
         description="A possible fact, the validity of which is assessed and reported by the Statement. A Statement can put forth the proposition as being true, false, or uncertain, and may provide an assessment of the level of confidence/evidence supporting this claim.",
+        discriminator="type",
     )
     direction: Direction = Field(
         ...,
@@ -624,3 +637,70 @@ class StudyGroup(Entity, BaseModelForbidExtra):
         None,
         description="A feature or role shared by all members of the StudyGroup, representing a criterion for membership in the group.",
     )
+
+
+class StatementValidatorMixin:
+    """Mixin class for reusable Statement model validators
+
+    Should be used with classes that inherit from Pydantic BaseModel
+    """
+
+    model_config = ConfigDict(extra="allow")
+
+    @model_validator(mode="after")
+    def statement_validator(cls, model: BaseModel) -> BaseModel:  # noqa: N805
+        """Validate that the model is a ``Statement``.
+
+        :param model: Pydantic BaseModel to validate
+        :raises ValueError: If ``model`` does not validate against a ``Statement``
+        :return: Validated model
+        """
+        try:
+            Statement(**model.model_dump())
+        except ValidationError as e:
+            err_msg = f"Must be a `Statement`: {e}"
+            raise ValueError(err_msg) from e
+        return model
+
+
+class EvidenceLineValidatorMixin:
+    """Mixin class for reusable EvidenceLine model validators
+
+    Should be used with classes that inherit from Pydantic BaseModel
+    """
+
+    model_config = ConfigDict(extra="allow")
+
+    @staticmethod
+    def _validate_evidence_outcome(
+        values: dict, system: System, codes: list[str]
+    ) -> dict:
+        """Validate ``evidenceOutcome`` property if it exists
+
+        :param values: Input values
+        :param system: System that should be used in ``MappableConcept``
+        :param codes: Codes that should be used in ``MappableConcept``
+        :raises ValueError: If ``evidenceOutcome`` exists and is invalid
+        :return: Validated input values. If ``evidenceOutcome`` exists, then it will be
+            validated and converted to a ``MappableConcept``
+        """
+        if "evidenceOutcome" in values:
+            mc = MappableConcept(**values["evidenceOutcome"])
+            values["evidenceOutcome"] = mc
+            validate_mappable_concept(mc, system, codes, mc_is_required=False)
+        return values
+
+    @model_validator(mode="after")
+    def evidence_line_validator(cls, model: BaseModel) -> BaseModel:  # noqa: N805
+        """Validate that the model is a ``EvidenceLine``.
+
+        :param model: Pydantic BaseModel to validate
+        :raises ValueError: If ``model`` does not validate against a ``EvidenceLine``
+        :return: Validated model
+        """
+        try:
+            EvidenceLine(**model.model_dump())
+        except ValidationError as e:
+            err_msg = f"Must be an `EvidenceLine`: {e}"
+            raise ValueError(err_msg) from e
+        return model

src/ga4gh/va_spec/base/enums.py

@@ -75,6 +75,19 @@ class ClinGenClassification(str, Enum):
 CLIN_GEN_CLASSIFICATIONS = [v.value for v in ClinGenClassification.__members__.values()]
 
 
+class CcvClassification(str, Enum):
+    """Define constraints for CCV classifications"""
+
+    ONCOGENIC = "oncogenic"
+    LIKELY_ONCOGENIC = "likely oncogenic"
+    UNCERTAIN_SIGNIFICANCE = "uncertain significance"
+    LIKELY_BENIGN = "likely benign"
+    BENIGN = "benign"
+
+
+CCV_CLASSIFICATIONS = [v.value for v in CcvClassification.__members__.values()]
+
+
 class System(str, Enum):
     """Define constraints for systems"""