Make a JSON schema for CC Certificate and Dataset objects.

3 files changed, 63 insertions, 0 deletions

diff --git a/tests/cc/test_cc_analysis.py b/tests/cc/test_cc_analysis.py
index 4606383c..42c97bac 100644
--- a/tests/cc/test_cc_analysis.py
+++ b/tests/cc/test_cc_analysis.py
@@ -1,14 +1,19 @@
 from __future__ import annotations
 
+import json
 import shutil
 from collections.abc import Generator
 from importlib import resources
 from pathlib import Path
+from tempfile import TemporaryDirectory
 
 import pytest
 import tests.data.cc.analysis
 import tests.data.common
+from jsonschema import Draft7Validator, validate
+from referencing import Registry, Resource
 
+import sec_certs.serialization.schemas
 from sec_certs.cert_rules import SARS_IMPLIED_FROM_EAL
 from sec_certs.dataset.auxiliary_dataset_handling import (
     CPEDatasetHandler,
@@ -258,3 +263,35 @@ def test_segment_extractor(reference_dataset: CCDataset):
     df = ReferenceSegmentExtractor()(reference_dataset.certs.values())
     assert df is not None
     assert not df.empty
+
+
+def test_schema_validate_single(reference_dataset: CCDataset):
+    with (
+        resources.open_text(sec_certs.serialization.schemas, "cc_certificate.json") as schema_file,
+        TemporaryDirectory() as tmp_dir,
+    ):
+        schema = json.load(schema_file)
+        for cert in reference_dataset:
+            fname = Path(tmp_dir) / (cert.dgst + ".json")
+            cert.to_json(fname)
+            with fname.open("r") as handle:
+                validate(instance=json.load(handle), schema=schema)
+
+
+def test_schema_validate_full(reference_dataset: CCDataset):
+    registry = Registry()
+    with (
+        resources.open_text(sec_certs.serialization.schemas, "cc_certificate.json") as cert_schema_file,
+        resources.open_text(sec_certs.serialization.schemas, "cc_dataset.json") as dataset_schema_file,
+    ):
+        cert_schema = json.load(cert_schema_file)
+        registry = Resource.from_contents(cert_schema) @ registry
+        dataset_schema = json.load(dataset_schema_file)
+        registry = Resource.from_contents(dataset_schema) @ registry
+    validator = Draft7Validator(schema={"$ref": "http://sec-certs.org/schemas/cc_dataset.json"}, registry=registry)
+
+    with TemporaryDirectory() as tmp_dir:
+        fname = Path(tmp_dir) / "dataset.json"
+        reference_dataset.to_json(fname)
+        with fname.open("r") as handle:
+            validator.validate(instance=json.load(handle))
diff --git a/tests/cc/test_cc_certificate.py b/tests/cc/test_cc_certificate.py
index d24f4c30..d53663dd 100644
--- a/tests/cc/test_cc_certificate.py
+++ b/tests/cc/test_cc_certificate.py
@@ -7,7 +7,9 @@ from pathlib import Path
 import pytest
 import tests.data.cc.analysis
 import tests.data.cc.certificate
+from jsonschema import validate
 
+import sec_certs.serialization.schemas
 from sec_certs.dataset import CCDataset
 from sec_certs.sample import CCCertificate
 
@@ -100,3 +102,11 @@ def test_cert_older_dgst(cert_one: CCCertificate):
     cert_one.report_link = None
     with pytest.raises(RuntimeError):
         cert_one.older_dgst
+
+
+def test_schema_validation(data_dir: Path):
+    with (
+        (data_dir / "fictional_cert.json").open("r") as cert,
+        resources.open_text(sec_certs.serialization.schemas, "cc_certificate.json") as schema,
+    ):
+        validate(json.load(cert), json.load(schema))
diff --git a/tests/cc/test_cc_dataset.py b/tests/cc/test_cc_dataset.py
index 3d13edeb..071f5330 100644
--- a/tests/cc/test_cc_dataset.py
+++ b/tests/cc/test_cc_dataset.py
@@ -1,10 +1,13 @@
 import json
 import shutil
+from importlib import resources
 from pathlib import Path
 from tempfile import TemporaryDirectory
 
 import pytest
+from jsonschema import validate
 
+import sec_certs.serialization.schemas
 from sec_certs import constants
 from sec_certs.configuration import config
 from sec_certs.dataset.cc import CCDataset
@@ -166,3 +169,16 @@ def test_to_pandas(toy_dataset: CCDataset):
     assert df.shape == (len(toy_dataset), len(CCCertificate.pandas_columns))
     assert df.index.name == "dgst"
     assert set(df.columns) == (set(CCCertificate.pandas_columns).union({"year_from"})) - {"dgst"}
+
+
+def test_schema_validate(toy_dataset: CCDataset):
+    with (
+        resources.open_text(sec_certs.serialization.schemas, "cc_certificate.json") as schema_file,
+        TemporaryDirectory() as tmp_dir,
+    ):
+        schema = json.load(schema_file)
+        for cert in toy_dataset:
+            fname = Path(tmp_dir) / (cert.dgst + ".json")
+            cert.to_json(fname)
+            with fname.open("r") as handle:
+                validate(instance=json.load(handle), schema=schema)