Add a structure hash for rubrics, so we can find rubrics that have the same…

Add a structure hash for rubrics, so we can find rubrics that have the same structure, even if the prompts are different.

apps/openassessment/assessment/admin.py

@@ -18,7 +18,8 @@ class RubricAdmin(admin.ModelAdmin):
     list_display_links = ('id', 'content_hash')
     search_fields = ('id', 'content_hash')
     readonly_fields = (
-        'id', 'content_hash', 'points_possible', 'criteria_summary', 'data'
+        'id', 'content_hash', 'structure_hash',
+        'points_possible', 'criteria_summary', 'data'
     )
 
     def criteria_summary(self, rubric_obj):

apps/openassessment/assessment/models/base.py

@@ -57,9 +57,12 @@ class Rubric(models.Model):
     creating a new Rubric instead. This makes it easy to cache and do hash-based
     lookups.
     """
-    # SHA1 hash
+    # SHA1 hash, including prompts and explanations
     content_hash = models.CharField(max_length=40, unique=True, db_index=True)
 
+    # SHA1 hash of just the rubric structure (criteria / options / points)
+    structure_hash = models.CharField(max_length=40, db_index=True)
+
     class Meta:
         app_label = "assessment"
 
@@ -91,6 +94,38 @@ class Rubric(models.Model):
         canonical_form = json.dumps(rubric_dict, sort_keys=True)
         return sha1(canonical_form).hexdigest()
 
+    @staticmethod
+    def structure_hash_from_dict(rubric_dict):
+        """
+        Generate a hash of the rubric that includes only structural information:
+            * Criteria names and order
+            * Option names / points / order number
+
+        We do NOT include prompt text or option explanations.
+
+        NOTE: currently, we use the criterion and option names as unique identifiers,
+        so we include them in the structure.  In the future, we plan to assign
+        criteria/options unique IDs -- when we do that, we will need to update
+        this method and create a data migration for existing rubrics.
+        """
+        structure = [
+            {
+                "criterion_name": criterion.get('name'),
+                "criterion_order": criterion.get('order_num'),
+                "options": [
+                    {
+                        "option_name": option.get('name'),
+                        "option_points": option.get('points'),
+                        "option_order": option.get('order_num')
+                    }
+                    for option in criterion.get('options', [])
+                ]
+            }
+            for criterion in rubric_dict.get('criteria', [])
+        ]
+        canonical_form = json.dumps(structure, sort_keys=True)
+        return sha1(canonical_form).hexdigest()
+
     def options_ids(self, options_selected):
         """Given a mapping of selected options, return the option IDs.
 

apps/openassessment/assessment/serializers/base.py

@@ -92,7 +92,7 @@ class RubricSerializer(NestedModelSerializer):
 
     class Meta:
         model = Rubric
-        fields = ('id', 'content_hash', 'criteria', 'points_possible')
+        fields = ('id', 'content_hash', 'structure_hash', 'criteria', 'points_possible')
 
     def validate_criteria(self, attrs, source):
         """Make sure we have at least one Criterion in the Rubric."""
@@ -283,6 +283,7 @@ def rubric_from_dict(rubric_dict):
         rubric = Rubric.objects.get(content_hash=content_hash)
     except Rubric.DoesNotExist:
         rubric_dict["content_hash"] = content_hash
+        rubric_dict["structure_hash"] = Rubric.structure_hash_from_dict(rubric_dict)
         for crit_idx, criterion in enumerate(rubric_dict.get("criteria", {})):
             if "order_num" not in criterion:
                 criterion["order_num"] = crit_idx

apps/openassessment/assessment/test/test_rubric.py

@@ -3,10 +3,12 @@
 Tests for assessment models.
 """
 
+import copy
 from openassessment.test_utils import CacheResetTest
 from openassessment.assessment.models import (
     Rubric, Criterion, CriterionOption, InvalidOptionSelection
 )
+from openassessment.assessment.test.constants import RUBRIC
 
 
 class TestRubricOptionIds(CacheResetTest):
@@ -122,7 +124,7 @@ class TestRubricOptionIds(CacheResetTest):
     def test_options_ids_points_caching(self):
         # First call: the dict is not cached
         with self.assertNumQueries(1):
-            options_ids = self.rubric.options_ids_for_points({
+            self.rubric.options_ids_for_points({
                 'test criterion 0': 0,
                 'test criterion 1': 1,
                 'test criterion 2': 2,
@@ -131,7 +133,7 @@ class TestRubricOptionIds(CacheResetTest):
 
         # Second call: the dict is not cached
         with self.assertNumQueries(0):
-            options_ids = self.rubric.options_ids_for_points({
+            self.rubric.options_ids_for_points({
                 'test criterion 0': 1,
                 'test criterion 1': 2,
                 'test criterion 2': 1,
@@ -163,3 +165,66 @@ class TestRubricOptionIds(CacheResetTest):
                 'test criterion 2': 1,
                 'test criterion 3': 0
             })
+
+    def test_structure_hash_identical(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+
+        # Same structure, but different text should have the same structure hash
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['prompt'] = 'altered!'
+        for criterion in altered_rubric['criteria']:
+            criterion['prompt'] = 'altered!'
+            for option in criterion['options']:
+                option['explanation'] = 'altered!'
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+
+        # Expect that the two hashes are the same
+        self.assertEqual(first_hash, second_hash)
+
+    def test_structure_hash_extra_keys(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+
+        # Same structure, add some extra keys
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['extra'] = 'extra!'
+        altered_rubric['criteria'][0]['extra'] = 'extra!'
+        altered_rubric['criteria'][0]['options'][0]['extra'] = 'extra!'
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+
+        # Expect that the two hashes are the same
+        self.assertEqual(first_hash, second_hash)
+
+    def test_structure_hash_criterion_order_changed(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['criteria'][0]['order_num'] = 5
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+        self.assertNotEqual(first_hash, second_hash)
+
+    def test_structure_hash_criterion_name_changed(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['criteria'][0]['name'] = 'altered!'
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+        self.assertNotEqual(first_hash, second_hash)
+
+    def test_structure_hash_option_order_changed(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['criteria'][0]['options'][0]['order_num'] = 5
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+        self.assertNotEqual(first_hash, second_hash)
+
+    def test_structure_hash_option_name_changed(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['criteria'][0]['options'][0]['name'] = 'altered!'
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+        self.assertNotEqual(first_hash, second_hash)
+
+    def test_structure_hash_option_points_changed(self):
+        first_hash = Rubric.structure_hash_from_dict(RUBRIC)
+        altered_rubric = copy.deepcopy(RUBRIC)
+        altered_rubric['criteria'][0]['options'][0]['points'] = 'altered!'
+        second_hash = Rubric.structure_hash_from_dict(altered_rubric)
+        self.assertNotEqual(first_hash, second_hash)

apps/openassessment/xblock/test/data/student_training_mixin.json

@@ -9,6 +9,7 @@
             "training_rubric": {
                 "id": 2,
                 "content_hash": "de2bb2b7e2c6e3df014e53b8c65f37d511cc4344",
+                "structure_hash": "a513b20d93487d6d80e31e1d974bf22519332567",
                 "criteria": [
                     {
                         "order_num": 0,
@@ -67,4 +68,4 @@
             }
         }
     }
-}
\ No newline at end of file
+}