Merge pull request #142 from OpenDataServices/2023-12-08

Jsonschema 4.18+ is now required

CHANGELOG.md

@@ -10,12 +10,13 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 ### Removed
 
 - Dropped support for Python 3.6 & 3.7, as these are now end of life.
-- Drop jsonschema 3 support
+- Jsonschema 4.18+ is now required. Support for 3 and older versions of 4 is removed.
 
 ### Changed
 
 - Restore jsonschema's type validator, as its performance has improved in recent Python versions https://github.com/OpenDataServices/lib-cove/pull/127
 - Allow `SchemaJsonMixin` classes to define a `validator` method, that accepts lib-cove's JSON Schema draft 4 validator class and its format checker, and returns a validator instance. https://github.com/OpenDataServices/lib-cove/pull/128
+- If the above `validator` option is not being used, allow `SchemaJsonMixin` classes to define a `registry` value. Should be a referencing.Registry instance. https://github.com/OpenDataServices/lib-cove/pull/123
 
 ### Fixed
 

libcove/lib/common.py

@@ -16,14 +16,15 @@
 import jsonref
 import jsonschema.validators
 import requests
+from referencing import Registry, Resource
 
 try:
     from functools import cached_property
 except ImportError:
     from cached_property import cached_property
 
 from flattentool import unflatten
-from jsonschema import FormatChecker, RefResolver
+from jsonschema import FormatChecker
 from jsonschema._utils import extras_msg, find_additional_properties, uniq
 from jsonschema.exceptions import UndefinedTypeCheck, ValidationError
 
@@ -822,25 +823,30 @@ def get_schema_validation_errors(
     if hasattr(schema_obj, "validator"):
         our_validator = schema_obj.validator(validator, format_checker)
     else:
-        if getattr(schema_obj, "extended", None):
-            resolver = CustomRefResolver(
-                "",
-                pkg_schema_obj,
-                config=getattr(schema_obj, "config", None),
-                schema_url=schema_obj.schema_host,
-                schema_file=schema_obj.extended_schema_file,
-                file_schema_name=schema_obj.schema_name,
-            )
+        if hasattr(schema_obj, "registry"):
+            registry = schema_obj.registry
         else:
-            resolver = CustomRefResolver(
-                "",
-                pkg_schema_obj,
-                config=getattr(schema_obj, "config", None),
-                schema_url=schema_obj.schema_host,
-            )
+            if getattr(schema_obj, "extended", None):
+                resolver = CustomRefResolver(
+                    "",
+                    pkg_schema_obj,
+                    config=getattr(schema_obj, "config", None),
+                    schema_url=schema_obj.schema_host,
+                    schema_file=schema_obj.extended_schema_file,
+                    file_schema_name=schema_obj.schema_name,
+                )
+            else:
+                resolver = CustomRefResolver(
+                    "",
+                    pkg_schema_obj,
+                    config=getattr(schema_obj, "config", None),
+                    schema_url=schema_obj.schema_host,
+                )
+
+            registry = Registry(retrieve=resolver.retrieve)
 
         our_validator = validator(
-            pkg_schema_obj, format_checker=format_checker, resolver=resolver
+            pkg_schema_obj, format_checker=format_checker, registry=registry
         )
 
     for e in our_validator.iter_errors(json_data):
@@ -1165,7 +1171,7 @@ def get_fields_present(*args, **kwargs):
     }
 
 
-class CustomRefResolver(RefResolver):
+class CustomRefResolver:
     """This RefResolver is only for use with the jsonschema library"""
 
     def __init__(self, *args, **kw):
@@ -1178,26 +1184,19 @@ def __init__(self, *args, **kw):
         # this is ignored when you supply a file
         self.schema_url = kw.pop("schema_url", "")
         self.config = kw.pop("config", "")
-        super().__init__(*args, **kw)
 
-    def resolve_remote(self, uri):
+    def retrieve(self, uri):
         schema_name = uri.split("/")[-1]
         if self.schema_file and self.file_schema_name == schema_name:
             uri = self.schema_file
         else:
             uri = urljoin(self.schema_url, schema_name)
 
-        document = self.store.get(uri)
-
-        if document:
-            return document
         if uri.startswith("http"):
             # This branch of the if-statement in-lines `RefResolver.resolve_remote()`, but using `get_request()`.
+            # https://github.com/python-jsonschema/jsonschema/blob/dbc398245a583cb2366795dc529ae042d10c1577/jsonschema/validators.py#L1008-L1023
             scheme = urlsplit(uri).scheme
-
-            if scheme in self.handlers:
-                result = self.handlers[scheme](uri)
-            elif scheme in ["http", "https"]:
+            if scheme in ("http", "https"):
                 # Requests has support for detecting the correct encoding of
                 # json over http
                 result = get_request(uri, config=self.config).json()
@@ -1206,16 +1205,12 @@ def resolve_remote(self, uri):
                 with urlopen(uri) as url:
                     result = json.loads(url.read().decode("utf-8"))
 
-            if self.cache_remote:
-                self.store[uri] = result
-            return result
         else:
             with open(uri) as schema_file:
                 result = json.load(schema_file)
 
         add_is_codelist(result)
-        self.store[uri] = result
-        return result
+        return Resource.from_contents(result)
 
 
 def _get_schema_deprecated_paths(

setup.py

@@ -11,7 +11,8 @@
     long_description="A data review library",
     install_requires=[
         "jsonref",
-        "jsonschema>=4",
+        "jsonschema>=4.18",
+        "referencing",
         "requests",
         "cached-property;python_version<'3.8'",
         "flattentool>=0.11.0",

tests/lib/test_common.py

@@ -7,6 +7,7 @@
 import jsonschema
 import pytest
 from freezegun import freeze_time
+from referencing.exceptions import CannotDetermineSpecification
 
 from libcove.lib.common import (
     SchemaJsonMixin,
@@ -766,7 +767,7 @@ def get_pkg_schema_obj(self):
     assert "[Decimal('3.1')] is too short" in validation_error_json
 
 
-def test_property_that_is_not_json_schema_doesnt_raise_exception(caplog, tmpdir):
+def test_property_that_is_not_json_schema_does_raise_exception(tmpdir):
     tmpdir.join("test.json").write(
         json.dumps({"properties": {"bad_property": "not_a_json_schema"}})
     )
@@ -778,12 +779,8 @@ class DummySchemaObj:
         def get_pkg_schema_obj(self):
             return {"$ref": "test.json"}
 
-    validation_errors = get_schema_validation_errors({}, DummySchemaObj(), "", {}, {})
-    assert validation_errors == {}
-    assert (
-        "A 'properties' object contains a 'bad_property' value that is not a JSON Schema: 'not_a_json_schema'"
-        in caplog.text
-    )
+    with pytest.raises(CannotDetermineSpecification):
+        get_schema_validation_errors({}, DummySchemaObj(), "", {}, {})
 
 
 @pytest.mark.parametrize(