Fix schema, make it stricter, add tests (#705)

* Fix schema, make it stricter, add tests * format
GoogleChrome · Dec 13, 2024 · 8296ccf · 8296ccf
1 parent 1ffd306
commit 8296ccf
Show file tree

Hide file tree

Showing 3 changed files with 173 additions and 54 deletions.
diff --git a/RWS-Submission_Guidelines.md b/RWS-Submission_Guidelines.md
@@ -44,57 +44,7 @@ ccTLD (country code top-level domain) variants for the subsets above are also su
 New submissions to the canonical RWS list must be filed as <a href="https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/creating-a-pull-request">pull requests (PRs)</a> on GitHub. Submitters should ensure that submissions follow the schema template provided below. Anyone with a <a href="https://docs.github.com/en/get-started/learning-about-github/types-of-github-accounts">GitHub account</a> may make a submission.
 
 Modifications to existing sets, including deletions, must also be submitted as new PRs against the canonical RWS list.
-The canonical RWS list will be validated against this schema whenever a user files their PR:
-
-```json
-{
-  "type": "object",
-  "properties": {
-    "sets": {
-      "type": "array",
-      "items": {
-        "type": "object",
-        "properties": {
-	  "contact" : {"type": "string"},
-          "ccTLDs": {
-            "type": "object",
-            "additionalProperties": {
-              "type": "array",
-              "items": {
-                "type": "string"
-              }
-            }
-          },
-          "primary": { "type": "string" },
-          "associatedSites": {
-            "type": "array",
-            "items": {
-              "type": "string"
-            }
-          },
-          "serviceSites": {
-            "type": "array",
-            "items": {
-              "type": "string"
-            }
-          },
-          "rationaleBySite": {
-            "type": "object",
-            "additionalProperties": {
-              "type": "string"
-            }
-          }
-        },
-        "required": ["primary"],
-        "dependentRequired": {
-          "associatedSites": ["An explanation of how you clearly present the affiliation across domains to users and why users would expect your domains to be affiliated"],
-          "serviceSites": ["An explanation of how each domain in this subset supports functionality or security needs."]
-        }
-      }
-    }
-  }
-}
-```
+The canonical RWS list will be validated against [this schema](./SCHEMA.json) whenever a user files their PR.
 
 A hypothetical example of the RWS canonical list is provided below for reference. A submission should follow the structure below, with new submissions being added as items to the "sets" list.
 

diff --git a/SCHEMA.json b/SCHEMA.json
@@ -1,11 +1,13 @@
 {
     "type": "object",
+    "additionalProperties": false,
     "properties": {
-        "contact" : {"type": "string"},
         "sets":  {
             "type": "array",
+            "uniqueItems": true,
             "items": {
                 "type": "object",
+                "additionalProperties": false,
                 "properties": {
                     "ccTLDs": {
                         "type": "object",
@@ -16,15 +18,18 @@
                             }
                         }
                     },
+                    "contact": { "type": "string" },
                     "primary": {"type": "string"},
                     "associatedSites": {
                         "type": "array",
+                        "uniqueItems": true,
                         "items": {
                             "type": "string"
                         }
                     },
                     "serviceSites": {
                         "type": "array",
+                        "uniqueItems": true,
                         "items": {
                             "type": "string"
                         }
@@ -40,8 +45,13 @@
                 "dependentRequired": {
                     "associatedSites": ["rationaleBySite"],
                     "serviceSites": ["rationaleBySite"]
-                }
+                },
+                "anyOf": [
+                    { "required": ["associatedSites"] },
+                    { "required": ["serviceSites"] },
+                    { "required": ["ccTLDs"] }
+                ]
             }
         }
     }
-}
+}
diff --git a/tests/rws_tests.py b/tests/rws_tests.py
@@ -82,6 +82,87 @@ def test_parse_and_check_format(self):
 class TestValidateSchema(unittest.TestCase):
     """A test suite for the validate_schema function of RwsCheck"""
 
+    def test_valid_associated(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "associatedSites": ["https://associated1.com"],
+                    "rationaleBySite": {
+                        "https://associated1.com": "example rationale",
+                    },
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        rws_check.validate_schema("SCHEMA.json")
+
+    def test_valid_service(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "serviceSites": ["https://service1.com"],
+                    "rationaleBySite": {
+                        "https://service1.com": "example rationale",
+                    },
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        rws_check.validate_schema("SCHEMA.json")
+
+    def test_valid_ccTLDs(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "ccTLDs": {"https://primary.com": ["https://primary.ca"]},
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        rws_check.validate_schema("SCHEMA.json")
+
+    def test_valid_full(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "associatedSites": ["https://associated1.com"],
+                    "serviceSites": ["https://service1.com"],
+                    "rationaleBySite": {
+                        "https://associated1.com": "example rationale",
+                        "https://service1.com": "example rationale",
+                    },
+                    "ccTLDs": {"https://associated1.com": ["https://associated1.ca"]},
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        rws_check.validate_schema("SCHEMA.json")
+
+    def test_duplicate_set(self):
+        entry = {
+            "contact": "[email protected]",
+            "primary": "https://primary.com",
+            "associatedSites": ["https://associated1.com"],
+            "serviceSites": ["https://service1.com"],
+            "rationaleBySite": {
+                "https://associated1.com": "example rationale",
+                "https://service1.com": "example rationale",
+            },
+            "ccTLDs": {"https://associated1.com": ["https://associated1.ca"]},
+        }
+        json_dict = {"sets": [entry, entry]}
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
     def test_no_primary(self):
         json_dict = {
             "sets": [
@@ -101,6 +182,19 @@ def test_no_primary(self):
         with self.assertRaises(ValidationError):
             rws_check.validate_schema("SCHEMA.json")
 
+    def test_primary_only(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
     def test_no_rationaleBySite(self):
         json_dict = {
             "sets": [
@@ -150,6 +244,71 @@ def test_no_contact(self):
         with self.assertRaises(ValidationError):
             rws_check.validate_schema("SCHEMA.json")
 
+    def test_nonunique_associated_sites(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "associatedSites": [
+                        "https://associated1.com",
+                        "https://associated1.com",
+                    ],
+                    "rationaleBySite": {
+                        "https://associated1.com": "example rationale",
+                    },
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
+    def test_nonunique_service_sites(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "serviceSites": [
+                        "https://service1.com",
+                        "https://service1.com",
+                    ],
+                    "rationaleBySite": {
+                        "https://service1.com": "example rationale",
+                    },
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
+    def test_unexpected_top_level_property(self):
+        json_dict = {"sets": [], "foo": True}
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
+    def test_unexpected_set_level_property(self):
+        json_dict = {
+            "sets": [
+                {
+                    "contact": "[email protected]",
+                    "primary": "https://primary.com",
+                    "associatedSites": ["https://associated1.com"],
+                    "rationaleBySite": {
+                        "https://associated1.com": "example rationale",
+                    },
+                    "ccTLDs": {"https://associated1.com": ["https://associated1.ca"]},
+                    "foo": True,
+                }
+            ]
+        }
+        rws_check = RwsCheck(rws_sites=json_dict, etlds=None, icanns=set(["ca"]))
+        with self.assertRaises(ValidationError):
+            rws_check.validate_schema("SCHEMA.json")
+
 
 class TestRwsSetEqual(unittest.TestCase):
     def test_equal_case(self):