fix: forbid extra attributes when additionalProperties is set to False (

#506)

close #505

src/ga4gh/core/models.py

@@ -18,6 +18,12 @@
 from ga4gh.core.identifiers import GA4GH_IR_REGEXP
 
 
+class BaseModelForbidExtra(BaseModel):
+    """Base Pydantic model class with extra attributes forbidden."""
+
+    model_config = ConfigDict(extra="forbid")
+
+
 class Relation(str, Enum):
     """A mapping relation between concepts as defined by the Simple Knowledge
     Organization System (SKOS).
@@ -129,7 +135,7 @@ class Element(BaseModel, ABC):
 #########################################
 
 
-class Coding(Element):
+class Coding(Element, BaseModelForbidExtra):
     """A structured representation of a code for a defined concept in a terminology or
     code system.
     """
@@ -149,7 +155,7 @@ class Coding(Element):
     code: code  # Cannot use Field due to PydanticUserError: field name and type annotation must not clash.
 
 
-class ConceptMapping(Element):
+class ConceptMapping(Element, BaseModelForbidExtra):
     """A mapping to a concept in a terminology or code system."""
 
     model_config = ConfigDict(use_enum_values=True)
@@ -164,7 +170,7 @@ class ConceptMapping(Element):
     )
 
 
-class Extension(Element):
+class Extension(Element, BaseModelForbidExtra):
     """The Extension class provides entities with a means to include additional
     attributes that are outside of the specified standard but needed by a given content
     provider or system implementer. These extensions are not expected to be natively
@@ -186,7 +192,7 @@ class Extension(Element):
     )
 
 
-class MappableConcept(Element):
+class MappableConcept(Element, BaseModelForbidExtra):
     """A concept name that may be mapped to one or more `Codings`."""
 
     conceptType: Optional[str] = Field(  # noqa: N815

src/ga4gh/vrs/models.py

@@ -37,7 +37,13 @@
     PrevVrsVersion,
     sha512t24u,
 )
-from ga4gh.core.models import Element, Entity, MappableConcept, iriReference
+from ga4gh.core.models import (
+    BaseModelForbidExtra,
+    Element,
+    Entity,
+    MappableConcept,
+    iriReference,
+)
 from ga4gh.core.pydantic import get_pydantic_root, getattr_in
 
 
@@ -369,7 +375,7 @@ class ga4gh(_ValueObject.ga4gh):  # noqa: N801
         prefix: str
 
 
-class Expression(Element):
+class Expression(Element, BaseModelForbidExtra):
     """Representation of a variation by a specified nomenclature or syntax for a
     Variation object. Common examples of expressions for the description of molecular
     variation include the HGVS and ISCN nomenclatures.
@@ -465,7 +471,7 @@ class sequenceString(RootModel):
 #########################################
 
 
-class LengthExpression(_ValueObject):
+class LengthExpression(_ValueObject, BaseModelForbidExtra):
     """A sequence expressed only by its length."""
 
     type: Literal["LengthExpression"] = Field(
@@ -479,7 +485,7 @@ class ga4gh(_ValueObject.ga4gh):
         inherent = ["length", "type"]
 
 
-class ReferenceLengthExpression(_ValueObject):
+class ReferenceLengthExpression(_ValueObject, BaseModelForbidExtra):
     """An expression of a length of a sequence from a repeating reference."""
 
     type: Literal["ReferenceLengthExpression"] = Field(
@@ -501,7 +507,7 @@ class ga4gh(_ValueObject.ga4gh):
         inherent = ["length", "repeatSubunitLength", "type"]
 
 
-class LiteralSequenceExpression(_ValueObject):
+class LiteralSequenceExpression(_ValueObject, BaseModelForbidExtra):
     """An explicit expression of a Sequence."""
 
     type: Literal["LiteralSequenceExpression"] = Field(
@@ -519,7 +525,7 @@ class ga4gh(_ValueObject.ga4gh):
 #########################################
 
 
-class SequenceReference(_ValueObject):
+class SequenceReference(_ValueObject, BaseModelForbidExtra):
     """A sequence of nucleic or amino acid character codes."""
 
     model_config = ConfigDict(use_enum_values=True)
@@ -554,7 +560,7 @@ class ga4gh(_ValueObject.ga4gh):
         inherent = ["refgetAccession", "type"]
 
 
-class SequenceLocation(Ga4ghIdentifiableObject):
+class SequenceLocation(Ga4ghIdentifiableObject, BaseModelForbidExtra):
     """A `Location` defined by an interval on a `Sequence`."""
 
     type: Literal["SequenceLocation"] = Field(
@@ -660,7 +666,7 @@ class _VariationBase(Ga4ghIdentifiableObject, ABC):
 #########################################
 
 
-class Allele(_VariationBase):
+class Allele(_VariationBase, BaseModelForbidExtra):
     """The state of a molecule at a `Location`."""
 
     type: Literal["Allele"] = Field(
@@ -704,7 +710,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):  # noqa: N801
         inherent = ["location", "state", "type"]
 
 
-class CisPhasedBlock(_VariationBase):
+class CisPhasedBlock(_VariationBase, BaseModelForbidExtra):
     """An ordered set of co-occurring `Variation` on the same molecule."""
 
     type: Literal["CisPhasedBlock"] = Field(
@@ -736,7 +742,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):
 #########################################
 
 
-class Adjacency(_VariationBase):
+class Adjacency(_VariationBase, BaseModelForbidExtra):
     """The `Adjacency` class represents the adjoining of the end of a sequence with the
     beginning of an adjacent sequence, potentially with an intervening linker sequence.
     """
@@ -779,7 +785,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):
         inherent = ["adjoinedSequences", "linker", "type"]
 
 
-class Terminus(_VariationBase):
+class Terminus(_VariationBase, BaseModelForbidExtra):
     """The `Terminus` data class provides a structure for describing the end
     (terminus) of a sequence. Structurally similar to Adjacency but the linker sequence
     is not allowed and it removes the unnecessary array structure.
@@ -797,7 +803,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):  # noqa: N815
         inherent = ["location", "type"]
 
 
-class TraversalBlock(_ValueObject):
+class TraversalBlock(_ValueObject, BaseModelForbidExtra):
     """A component used to describe the orientation of applicable molecular variation
     within a DerivativeMolecule.
     """
@@ -820,7 +826,7 @@ class ga4gh(_ValueObject.ga4gh):
         inherent = ["component", "orientation", "type"]
 
 
-class DerivativeMolecule(_VariationBase):
+class DerivativeMolecule(_VariationBase, BaseModelForbidExtra):
     """The "Derivative Molecule" data class is a structure for describing a derivate
     molecule composed from multiple sequence components.
     """
@@ -851,7 +857,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):  # noqa: N815
 #########################################
 
 
-class CopyNumberCount(_VariationBase):
+class CopyNumberCount(_VariationBase, BaseModelForbidExtra):
     """The absolute count of discrete copies of a `Location`, within a system
     (e.g. genome, cell, etc.).
     """
@@ -872,7 +878,7 @@ class ga4gh(Ga4ghIdentifiableObject.ga4gh):  # noqa: N815
         inherent = ["copies", "location", "type"]
 
 
-class CopyNumberChange(_VariationBase):
+class CopyNumberChange(_VariationBase, BaseModelForbidExtra):
     """An assessment of the copy number of a `Location` within a system
     (e.g. genome, cell, etc.) relative to a baseline ploidy.
     """

tests/validation/test_schemas.py

@@ -105,6 +105,11 @@ def test_schema_class_fields(gks_schema, pydantic_models):
 
         required_schema_fields = set(mapping.schema[schema_model]["required"])
 
+        if mapping.schema[schema_model].get("additionalProperties") is False:
+            assert pydantic_model.model_config.get("extra") == "forbid", (
+                f"{pydantic_model} should forbid extra attributes"
+            )
+
         for prop, property_def in schema_properties.items():
             pydantic_model_field_info = pydantic_model.model_fields[prop]
             pydantic_field_required = pydantic_model_field_info.is_required()