DEV-2783: add unique link test (#1069)

Add test for detecting duplicate link definition via name or backref.

setup.cfg

@@ -8,15 +8,16 @@ author_email = [email protected]
 maintainer = GDC Feature Team
 maintainer_email = [email protected]
 license = Apache Software License 2.0
-license_file = LICENSE
+license_files = LICENSE
 url = https://github.com/NCI-GDC/gdcdictionary
 keywords = Genomic Data Commons, JSON schema, Data Dictionary, GDC, Cancer Research
 classifiers =
-    Programming Language :: PythonLicense :: OSI Approved :: Apache Software License
     Operating System :: Unix
     Operating System :: POSIX :: Linux
+    License :: OSI Approved :: Apache Software License
     Topic :: Software Development :: Libraries :: Python Modules
     Topic :: Internet
+    Programming Language :: Python
     Programming Language :: Python :: 3.7
     Programming Language :: Python :: 3.8
     Programming Language :: Python :: 3.9
@@ -25,7 +26,7 @@ classifiers =
 
 [options]
 zip_safe = True
-packages = find:
+packages = find_namespace:
 package_dir =
     =src
 python_requires = >=3.7

src/gdcdictionary/validators.py

@@ -81,9 +81,9 @@ def from_values(cls, schema: str, message: str, keys: List[str]) -> SchemaValida
         if "Additional properties are not allowed" in message:
             message = f"Key(s) {keys} not a valid property for type '{schema}'"
         logger.debug(
-            "json schema violation for '%s'",
+            "gdcdictionary schema violation for '%s'",
             schema,
-            extra={"keys": keys, "message": message},
+            extra={"keys": keys, "violation": message},
         )
         return SchemaValidationError(schema, message, keys)
 
@@ -120,7 +120,7 @@ def iter_errors(
                     extra={
                         "partial": partial,
                         "keys": violation.keys,
-                        "message": violation.message,
+                        "violation": violation.message,
                     },
                 )
             violation = SchemaValidationError.from_values(self.name, message, keys)
@@ -130,7 +130,7 @@ def iter_errors(
                     extra={
                         "partial": partial,
                         "keys": violation.keys,
-                        "message": violation.message,
+                        "violation": violation.message,
                     },
                 )
                 continue

tests/test_gdcdictionary.py

@@ -1,3 +1,6 @@
+import dataclasses
+from typing import Any, Mapping, Sequence
+
 try:
     from importlib.resources import files
 except ImportError:
@@ -8,6 +11,97 @@
 import gdcdictionary
 
 
+@dataclasses.dataclass(frozen=True)
+class Association:
+    """Represents a directed link between source and target.
+
+    A single link definition in a node will expand into two
+    associations, one for each direction.
+
+    For example, the definition below on the node `aliquot`
+    ```
+        id: aliquot
+        links:
+          - exclusive: true
+            required: true
+            subgroup:
+              - name: analytes
+                backref: aliquots
+                label: derived_from
+                target_type: analyte
+                multiplicity: many_to_one
+                required: false
+              - name: samples
+                backref: aliquots
+                label: derived_from
+                target_type: sample
+                multiplicity: many_to_many
+                required: false
+          - name: centers
+            backref: aliquots
+            label: shipped_to
+            target_type: center
+            multiplicity: many_to_one
+            required: false
+    ```
+    This results in 6 possible associations
+        centers: aliquot --> center
+        aliquots: center --> aliquot
+        analytes: aliquot --> analyte
+        aliquots: analyte --> aliquot
+        samples: aliquot --> sample
+        aliquots: sample --> aliquot
+
+    Within a source node, the association name should be unique.
+    For example, on aliquot there should be only one association named `centers`
+    """
+
+    name: str
+    source: str
+    target: str = dataclasses.field(compare=False, hash=False)
+
+
+class Associations(set):
+    """An unordered collection of unique association."""
+
+    def add(self, association: Association) -> None:
+        """Raise key error if entry already exists."""
+        if association in self:
+            raise KeyError(f"{association} already exists - Duplicate links not allowed.")
+        super().add(association)
+
+    def update(self, associations: "Associations") -> None:
+        """Raise key error if entry already exists."""
+        for association in associations:
+            if association in self:
+                raise ValueError(f"{association} already exists - Duplicate links not allowed.")
+        super().update(associations)
+
+
+def extract_links(source: str, links: Sequence[Mapping[str, Any]]) -> Associations:
+    """Inspect links collected from the schema of a node and generate Associations."""
+    associations = Associations()
+    for link in links:
+        if "subgroup" in link:
+            associations.update(extract_links(source, link["subgroup"]))
+            continue
+        forward = Association(name=link["name"], source=source, target=link["target_type"])
+        reverse = Association(name=link["backref"], target=source, source=link["target_type"])
+
+        associations.add(forward)
+        associations.add(reverse)
+    return associations
+
+
+def read_associations(dictionary: gdcdictionary.GDCDictionary) -> Associations:
+    """Read all association in the dictionary."""
+    associations = Associations()
+    for schema in dictionary.schema.values():
+        name = schema["id"]
+        associations.update(extract_links(name, schema["links"]))
+    return associations
+
+
 def test_load_dictionary__invalid_location() -> None:
     with pytest.raises(IOError):
         gdcdictionary.GDCDictionary(root_dir="invalid/dir/path")
@@ -19,3 +113,8 @@ def test_load_dictionary() -> None:
             root_dir=str(path.parent / "src/gdcdictionary/schemas"), lazy=False
         )
         assert dictionary.loaded is True
+
+
+def test_unique_associations():
+    """passes if no exception is raised"""
+    read_associations(gdcdictionary.gdcdictionary)

tests/test_validators.py

@@ -94,7 +94,7 @@ def test_validate_instances__invalid_types(partial: bool) -> None:
     )
 
 
-def test_partials_validation():
+def test_partials_validation() -> None:
     # example missing required fields
     instances = [
         {"type": "case", "days_to_consent": 123, "submitter_id": "UNSC-2"},