Merge pull request #4604 from akolson/embed-content-logic

Extract content urls

contentcuration/automation/migrations/0001_initial.py

@@ -0,0 +1,44 @@
+# Generated by Django 3.2.24 on 2024-08-05 21:23
+import uuid
+
+import django.db.models.deletion
+from django.db import migrations
+from django.db import models
+
+
+class Migration(migrations.Migration):
+    initial = True
+
+    dependencies = [
+        ('kolibri_public', '0003_alter_file_preset'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='RecommendationsCache',
+            fields=[
+                ('id', models.UUIDField(default=uuid.uuid4, editable=False, primary_key=True,
+                                        serialize=False)),
+                ('request_hash', models.CharField(max_length=32, null=True)),
+                ('rank', models.FloatField(default=0.0, null=True)),
+                ('override_threshold', models.BooleanField(default=False)),
+                ('timestamp', models.DateTimeField(auto_now_add=True)),
+                ('contentnode', models.ForeignKey(blank=True, null=True,
+                                                  on_delete=django.db.models.deletion.CASCADE,
+                                                  related_name='recommendations',
+                                                  to='kolibri_public.contentnode')),
+            ],
+        ),
+        migrations.AddIndex(
+            model_name='recommendationscache',
+            index=models.Index(fields=['request_hash'], name='request_hash_idx'),
+        ),
+        migrations.AddIndex(
+            model_name='recommendationscache',
+            index=models.Index(fields=['contentnode'], name='contentnode_idx'),
+        ),
+        migrations.AlterUniqueTogether(
+            name='recommendationscache',
+            unique_together={('request_hash', 'contentnode')},
+        ),
+    ]

contentcuration/automation/models.py

@@ -1,3 +1,30 @@
-# from django.db import models
+import uuid
 
-# Create your models here.
+from django.db import models
+from kolibri_public.models import ContentNode
+
+
+REQUEST_HASH_INDEX_NAME = "request_hash_idx"
+CONTENTNODE_INDEX_NAME = "contentnode_idx"
+
+
+class RecommendationsCache(models.Model):
+    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
+    request_hash = models.CharField(max_length=32, null=True)
+    contentnode = models.ForeignKey(
+        ContentNode,
+        null=True,
+        blank=True,
+        related_name='recommendations',
+        on_delete=models.CASCADE,
+    )
+    rank = models.FloatField(default=0.0, null=True)
+    override_threshold = models.BooleanField(default=False)
+    timestamp = models.DateTimeField(auto_now_add=True)
+
+    class Meta:
+        unique_together = ('request_hash', 'contentnode')
+        indexes = [
+            models.Index(fields=['request_hash'], name=REQUEST_HASH_INDEX_NAME),
+            models.Index(fields=['contentnode'], name=CONTENTNODE_INDEX_NAME),
+        ]

contentcuration/automation/tests/appnexus/test_base.py

@@ -1,69 +1,26 @@
 import time
-import pytest
-import requests
 from unittest.mock import patch
 
-from automation.utils.appnexus.base import Adapter
+import mock
+import pytest
+import requests
 from automation.utils.appnexus.base import Backend
 from automation.utils.appnexus.base import BackendRequest
+from automation.utils.appnexus.base import BackendResponse
 from automation.utils.appnexus.base import SessionWithMaxConnectionAge
 from automation.utils.appnexus.errors import ConnectionError
+from automation.utils.appnexus.errors import InvalidResponse
 
 
-class MockBackend(Backend):
-    base_url = 'https://kolibri-dev.learningequality.org'
-    connect_endpoint = '/status'
-    def connect(self) -> None:
-        return super().connect()
-
-    def make_request(self, request):
-        return super().make_request(request)
-
-class ErrorBackend(Backend):
-    base_url = 'https://bad-url.com'
-    connect_endpoint = '/status'
-    def connect(self) -> None:
-        return super().connect()
-
-    def make_request(self, request):
-        return super().make_request(request)
-
-
-class MockAdapter(Adapter):
-    def mockoperation(self):
-        pass
-
-
-def test_backend_singleton():
-    b1, b2 = MockBackend(), MockBackend()
-    assert id(b1) == id(b2)
-
-
-def test_adapter_creation():
-    a = MockAdapter(backend=MockBackend)
-    assert isinstance(a, Adapter)
-
-
-def test_adapter_backend_default():
-    b = MockBackend()
-    adapter = Adapter(backend=b)
-    assert isinstance(adapter.backend, Backend)
-
-
-def test_adapter_backend_custom():
-    b = MockBackend()
-    a = Adapter(backend=b)
-    assert a.backend is b
-
 def test_session_with_max_connection_age_request():
     with patch.object(requests.Session, 'request') as mock_request:
         session = SessionWithMaxConnectionAge()
         session.request('GET', 'https://example.com')
         assert mock_request.call_count == 1
 
+
 def test_session_with_max_connection_age_not_closing_connections():
-    with patch.object(requests.Session, 'close') as mock_close,\
-        patch.object(requests.Session, 'request') as mock_request:
+    with patch.object(requests.Session, 'close') as mock_close, patch.object(requests.Session, 'request') as mock_request:
         session = SessionWithMaxConnectionAge(60)
         session.request('GET', 'https://example.com')
         time.sleep(0.1)
@@ -72,9 +29,9 @@ def test_session_with_max_connection_age_not_closing_connections():
         assert mock_close.call_count == 0
         assert mock_request.call_count == 2
 
+
 def test_session_with_max_connection_age_closing_connections():
-    with patch.object(requests.Session, 'close') as mock_close,\
-        patch.object(requests.Session, 'request') as mock_request:
+    with patch.object(requests.Session, 'close') as mock_close, patch.object(requests.Session, 'request') as mock_request:
         session = SessionWithMaxConnectionAge(1)
         session.request('GET', 'https://example.com')
         time.sleep(2)
@@ -83,33 +40,55 @@ def test_session_with_max_connection_age_closing_connections():
         assert mock_close.call_count == 1
         assert mock_request.call_count == 2
 
-def test_backend_connect():
-    backend = MockBackend()
-    connected = backend.connect()
 
-    assert connected is True
+@mock.patch("automation.utils.appnexus.base.Backend.connect")
+def test_backend_connect(mock_connect):
+    mock_connect.return_value = True
+
+    backend = Backend()
+    result = backend.connect()
+
+    mock_connect.assert_called_once()
+    assert result is True
 
-def test_backend_connect_error():
-    backend = ErrorBackend()
-    connected = backend.connect()
 
-    assert connected is False
+@mock.patch("automation.utils.appnexus.base.Backend.connect")
+def test_backend_connect_error(mock_connect):
+    mock_connect.side_effect = [ConnectionError("Failed to connect"), False]
 
-def test_backend_request():
-    request = BackendRequest('GET', '/api/public/info')
+    backend = Backend()
 
-    backend = MockBackend()
+    with pytest.raises(ConnectionError) as exc_info:
+        backend.connect()
+    assert str(exc_info.value) == "Failed to connect"
+
+    result = backend.connect()
+    assert result is False
+
+    assert mock_connect.call_count == 2
+
+
+@mock.patch("automation.utils.appnexus.base.Backend.make_request")
+def test_backend_request(mock_make_request):
+    mock_response = BackendResponse(data=[{"key": "value"}])
+    mock_make_request.return_value = mock_response
+
+    backend = Backend()
+    request = BackendRequest(method="GET", path="/api/test")
     response = backend.make_request(request)
 
-    assert response.status_code == 200
-    assert len(response.__dict__) > 0
+    assert response == mock_response
+    mock_make_request.assert_called_once_with(request)
+
 
-def test_backend_request_error():
-    request = BackendRequest('GET', '/api/public/info')
+@mock.patch("automation.utils.appnexus.base.Backend.make_request")
+def test_backend_request_error(mock_make_request):
+    mock_make_request.side_effect = InvalidResponse("Request failed")
 
-    backend = ErrorBackend()
+    backend = Backend()
+    request = BackendRequest(method="GET", path="/api/test")
 
-    with pytest.raises(ConnectionError) as error:
+    with pytest.raises(InvalidResponse) as exc_info:
         backend.make_request(request)
-
-    assert "Unable to connect to" in str(error.value)
+    assert str(exc_info.value) == "Request failed"
+    mock_make_request.assert_called_once_with(request)

contentcuration/automation/tests/test_recommendations_cache_model.py

@@ -0,0 +1,64 @@
+import uuid
+
+from automation.models import RecommendationsCache
+from django.db import IntegrityError
+from kolibri_public.models import ContentNode
+
+from contentcuration.tests.base import StudioTestCase
+
+
+class TestRecommendationsCache(StudioTestCase):
+
+    def setUp(self):
+        self.content_node = ContentNode.objects.create(
+            id=uuid.uuid4(),
+            title='Test Content Node',
+            content_id=uuid.uuid4(),
+            channel_id=uuid.uuid4(),
+        )
+        self.cache = RecommendationsCache.objects.create(
+            request_hash='test_hash',
+            contentnode=self.content_node,
+            rank=1.0,
+            override_threshold=False
+        )
+
+    def test_cache_creation(self):
+        self.assertIsInstance(self.cache, RecommendationsCache)
+        self.assertEqual(self.cache.request_hash, 'test_hash')
+        self.assertEqual(self.cache.contentnode, self.content_node)
+        self.assertEqual(self.cache.rank, 1.0)
+        self.assertFalse(self.cache.override_threshold)
+
+    def test_cache_retrieval(self):
+        retrieved_cache = RecommendationsCache.objects.get(request_hash='test_hash')
+        self.assertEqual(retrieved_cache, self.cache)
+
+    def test_cache_uniqueness(self):
+        with self.assertRaises(IntegrityError):
+            RecommendationsCache.objects.create(
+                request_hash='test_hash',
+                contentnode=self.content_node,
+                rank=2.0,
+                override_threshold=True
+            )
+
+    def test_bulk_create_ignore_conflicts_true(self):
+        initial_count = RecommendationsCache.objects.count()
+        try:
+            RecommendationsCache.objects.bulk_create(
+                [self.cache, self.cache],
+                ignore_conflicts=True
+            )
+        except IntegrityError:
+            self.fail("bulk_create raised IntegrityError unexpectedly!")
+
+        final_count = RecommendationsCache.objects.count()
+        self.assertEqual(initial_count, final_count)
+
+    def test_bulk_create_ignore_conflicts_false(self):
+        with self.assertRaises(IntegrityError):
+            RecommendationsCache.objects.bulk_create(
+                [self.cache, self.cache],
+                ignore_conflicts=False
+            )

contentcuration/automation/utils/appnexus/base.py

@@ -56,8 +56,7 @@ def __init__(
 
 class BackendResponse(object):
     """ Class that should be inherited by specific backend for its responses"""
-    def __init__(self, error=None, **kwargs):
-        self.error = error
+    def __init__(self, **kwargs):
         for key, value in kwargs.items():
             setattr(self, key, value)
 
@@ -169,11 +168,10 @@ def connect(self, **kwargs):
 
     def make_request(self, request):
         """ Make a request to the backend service. """
-        response = self._make_request(request)
         try:
-            info = response.json()
-            info.update({"status_code": response.status_code})
-            return BackendResponse(**info)
+            response = self._make_request(request)
+            response_body = dict(data=response.json())
+            return BackendResponse(**response_body)
         except ValueError as e:
             logging.exception(e)
             raise errors.InvalidResponse("Invalid response from backend")

contentcuration/contentcuration/settings.py

@@ -89,6 +89,7 @@
     'django.contrib.postgres',
     'django_celery_results',
     'kolibri_public',
+    'automation',
 )
 
 SESSION_ENGINE = "django.contrib.sessions.backends.cached_db"

contentcuration/contentcuration/tests/utils/test_automation_manager.py

@@ -1,5 +1,4 @@
 import unittest
-from unittest.mock import MagicMock
 
 from contentcuration.utils.automation_manager import AutomationManager
 
@@ -11,30 +10,3 @@ def setUp(self):
     def test_creation(self):
         # Check if an instance of AutomationManager is created successfully
         self.assertIsInstance(self.automation_manager, AutomationManager)
-
-    def test_generate_embedding(self):
-        text = "Some text that needs to be embedded"
-        # Mock the generate_embedding method of RecommendationsAdapter
-        # as the implementation is yet to be done
-        self.automation_manager.recommendations_backend_adapter.generate_embedding = MagicMock(return_value=[0.1, 0.2, 0.3])
-        embedding_vector = self.automation_manager.generate_embedding(text)
-        self.assertIsNotNone(embedding_vector)
-
-    def test_embedding_exists(self):
-        embedding_vector = [0.1, 0.2, 0.3]
-        # Currently no solid implementation exists for this
-        # So the embadding_exists function returns true anyways
-        exists = self.automation_manager.embedding_exists(embedding_vector)
-        self.assertTrue(exists)
-
-    def test_load_recommendations(self):
-        embedding_vector = [0.1, 0.2, 0.3]
-        self.automation_manager.recommendations_backend_adapter.get_recommendations = MagicMock(return_value=["item1", "item2"])
-        recommendations = self.automation_manager.load_recommendations(embedding_vector)
-        self.assertIsInstance(recommendations, list)
-
-    def test_cache_embeddings(self):
-        embeddings_list = [[0.1, 0.2, 0.3]]
-        # Currently the function returns true anyways
-        success = self.automation_manager.cache_embeddings(embeddings_list)
-        self.assertTrue(success)