Fix tests for minimum properties and items (scrapinghub#432)

* Fix tests for minimum properties and items

In JSON Schema validation for minimum properties and items the data should be non-empty.

* Add tests for invalid empty properties and items

Add validation messages for empty properties and items that should be
non-empty.

* Fix warning from pytest trylast

Use decorator @pytest.hookimpl(trylast=True) instead @pytest.mark.trylast

* Fix warnings about literal comparations with 'is'

Change literal comparations from 'is' to '=='.

* Apply pre-commit

* Use jsonschema >= 3.2.0 for all python versions

* Define minimum version for JSON Schema

From version 4.21.0 JSON Schema to validation for minimum properties and items the data should be non-empty.

Author: rochamatcomp

setup.py

@@ -14,7 +14,7 @@
     zip_safe=False,
     include_package_data=True,
     install_requires=[
-        "jsonschema[format]>=3.2.0",
+        "jsonschema[format]>=4.21.0",
         "python-slugify",
     ],
     extras_require={

spidermon/contrib/pytest/plugins/filter_monitors.py

@@ -7,7 +7,7 @@ def pytest_report_header(config):
     return "Spidermon monitor filtering"
 
 
-@pytest.mark.trylast
+@pytest.hookimpl(trylast=True)
 def pytest_collection_modifyitems(session, config, items):
     items[:] = [
         item

spidermon/contrib/validation/jsonschema/translator.py

@@ -38,6 +38,7 @@ class JSONSchemaMessageTranslator(MessageTranslator):
         r"^.* is not allowed for .*$": messages.NOT_ALLOWED_VALUE,
         r"^.+ is too short$": messages.FIELD_TOO_SHORT,
         r"^.+ is too long$": messages.FIELD_TOO_LONG,
+        r"^.+ should be non-empty$": messages.SHOULD_BE_NON_EMPTY,
         r"^.+ does not match .*$": messages.REGEX_NOT_MATCHED,
         r"^.+ has non-unique elements$": messages.NOT_UNIQUE,
     }

spidermon/contrib/validation/messages.py

@@ -44,3 +44,4 @@
 NOT_MULTIPLE_OF = "Not multiple of"
 NOT_ALLOWED_VALUE = "Not allowed value"
 NOT_UNIQUE = "Not unique"
+SHOULD_BE_NON_EMPTY = "should be non-empty"

tests/contrib/scrapy/test_pipelines.py

@@ -88,7 +88,7 @@ class PipelineJSONSchemaValidator(PipelineTest):
             settings={SETTING_SCHEMAS: [test_schema]},
             cases=[
                 f"'{STATS_ITEM_ERRORS}' not in {{stats}}",
-                f"{{stats}}['{STATS_AMOUNTS}'] is 1",
+                f"{{stats}}['{STATS_AMOUNTS}'] == 1",
                 assert_type_in_stats(Item),
             ],
         ),
@@ -183,8 +183,8 @@ class PipelineJSONSchemaValidator(PipelineTest):
             item=TestItem(),
             settings={SETTING_SCHEMAS: {TestItem: [test_schema, tree_schema]}},
             cases=[
-                f"{{stats}}['{STATS_AMOUNTS}'] is 2",
-                f"{{stats}}['{STATS_ITEM_ERRORS}'] is 2",
+                f"{{stats}}['{STATS_AMOUNTS}'] == 2",
+                f"{{stats}}['{STATS_ITEM_ERRORS}'] == 2",
             ],
         ),
         DataTest(
@@ -198,7 +198,7 @@ class PipelineJSONSchemaValidator(PipelineTest):
             item=TreeItem(),
             settings={SETTING_SCHEMAS: {TestItem: test_schema, TreeItem: tree_schema}},
             cases=[
-                f"{{stats}}['{STATS_MISSINGS}'] is 1",
+                f"{{stats}}['{STATS_MISSINGS}'] == 1",
                 assert_type_in_stats(TestItem),
                 assert_type_in_stats(TreeItem),
             ],

tests/test_validators_jsonschema.py

@@ -907,35 +907,61 @@ class Maximum(SchemaTest):
 
 
 class MinItems(SchemaTest):
-    schema = {"minItems": 1}
+    schema = {"minItems": 2}
     data_tests = [
-        DataTest(name="longer is valid", data=[1, 2], valid=True),
-        DataTest(name="exact length is valid", data=[1], valid=True),
+        DataTest(name="longer is valid", data=[1, 2, 3], valid=True),
+        DataTest(name="exact length is valid", data=[1, 2], valid=True),
         DataTest(
             name="too short is invalid",
-            data=[],
+            data=[1],
             valid=False,
             expected_errors={"": [messages.FIELD_TOO_SHORT]},
         ),
         DataTest(name="ignores non-arrays", data="", valid=True),
     ]
 
 
+class EmptyItems(SchemaTest):
+    schema = {"minItems": 1}
+    data_tests = [
+        DataTest(
+            name="empty is invalid",
+            data=list(),
+            valid=False,
+            expected_errors={"": [messages.SHOULD_BE_NON_EMPTY]},
+        )
+    ]
+
+
 class MinProperties(SchemaTest):
-    schema = {"minProperties": 1}
+    schema = {"minProperties": 2}
     data_tests = [
-        DataTest(name="longer is valid", data={"foo": 1, "bar": 2}, valid=True),
-        DataTest(name="exact length is valid", data={"foo": 1}, valid=True),
+        DataTest(
+            name="longer is valid", data={"foo": 1, "bar": 2, "foobar": 3}, valid=True
+        ),
+        DataTest(name="exact length is valid", data={"foo": 1, "bar": 2}, valid=True),
         DataTest(
             name="too short is invalid",
-            data={},
+            data={"foo": 1},
             valid=False,
             expected_errors={"": [messages.NOT_ENOUGH_PROPERTIES]},
         ),
         DataTest(name="ignores non-objects", data="", valid=True),
     ]
 
 
+class EmptyProperties(SchemaTest):
+    schema = {"minProperties": 1}
+    data_tests = [
+        DataTest(
+            name="empty is invalid",
+            data=dict(),
+            valid=False,
+            expected_errors={"": [messages.SHOULD_BE_NON_EMPTY]},
+        )
+    ]
+
+
 class Minimum(SchemaTest):
     # exclusiveMinimum behaviour changed from draft-04 to draft-06
     # http://json-schema.org/draft-06/json-schema-release-notes.html#backwards-incompatible-changes

tox.ini

@@ -22,10 +22,10 @@ commands = pytest -s --ignore=./tests/contrib --ignore=./tests/utils/test_zyte.p
 basepython = python3.8
 deps =
     {[testenv]deps}
-    jsonschema[format]==3.2.0
+    jsonschema[format]==4.21.0
 
 [testenv:docs]
-deps = 
+deps =
     -r {toxinidir}/docs/requirements-docs.txt
 changedir = docs
 commands =