Merge pull request #996 from LibraryOfCongress/more-api-test-coverage

Add more test coverage for API endpoints

concordia/tests/test_api_views.py

@@ -1,8 +1,17 @@
+from urllib.parse import urlparse
+
 from django.test import TestCase, override_settings
 from django.urls import reverse
 from django.utils.timezone import now
 
-from concordia.models import Asset, Item, Transcription, TranscriptionStatus, User
+from concordia.models import (
+    Asset,
+    Campaign,
+    Item,
+    Transcription,
+    TranscriptionStatus,
+    User,
+)
 from concordia.utils import get_anonymous_user
 
 from .utils import JSONAssertMixin, create_asset, create_item, create_project
@@ -18,11 +27,14 @@ def setUpTestData(cls):
             username="reviewer", email="tester@example.com"
         )
 
-        project = create_project()
+        cls.test_project = create_project()
 
         cls.items = [
             create_item(
-                item_id=f"item_{i}", title=f"Item {i}", project=project, do_save=False
+                item_id=f"item_{i}",
+                title=f"Item {i}",
+                project=cls.test_project,
+                do_save=False,
             )
             for i in range(0, 3)
         ]
@@ -57,20 +69,73 @@ def setUpTestData(cls):
         submitted_t.full_clean()
         submitted_t.save()
 
-    def get_asset_list(self, url, page_size=23):
-        resp = self.client.get(url, {"per_page": page_size})
+    def get_api_response(self, url, **request_args):
+        """
+        This issues a call to one of our API views and confirms that the
+        response follows our basic conventions of returning a valid JSON
+        response
+        """
+
+        qs = {"format": "json"}
+        if request_args is not None:
+            qs.update(request_args)
+
+        resp = self.client.get(url, qs)
         data = self.assertValidJSON(resp)
+        return resp, data
 
-        self.assertIn("objects", data)
-        object_count = len(data["objects"])
-        self.assertLessEqual(object_count, 23)
+    def get_api_list_response(self, url, page_size=23, **request_args):
+        """
+        This issues a call to one of our API views and confirms that the
+        response follows our basic conventions of returning a top level object
+        with members“objects” (list) and “pagination” (object).
+        """
+
+        qs = {"per_page": page_size}
+        if request_args is not None:
+            qs.update(request_args)
+
+        resp, data = self.get_api_response(url, **qs)
 
+        self.assertIn("objects", data)
         self.assertIn("pagination", data)
 
-        self.assertAssetsHaveLatestTranscriptions(data["objects"])
+        object_count = len(data["objects"])
+        self.assertLessEqual(object_count, page_size)
+
+        self.assertAbsoluteURLs(data["objects"])
+        self.assertAbsoluteURLs(data["pagination"])
 
         return resp, data
 
+    def assertAbsoluteUrl(self, url, allow_none=True):
+        """Require a URL to either be None or an absolute URL"""
+
+        if url is None and allow_none:
+            return
+
+        parsed = urlparse(url)
+        self.assertIn(
+            parsed.scheme, ["http", "https"], msg=f"Expected {url} to have HTTP scheme"
+        )
+
+        self.assertTrue(parsed.netloc)
+
+    def assertAbsoluteURLs(self, data):
+        if isinstance(data, dict):
+            for k, v in data.items():
+                if k.endswith("url"):
+                    self.assertAbsoluteUrl(v)
+                elif isinstance(v, (dict, list)):
+                    self.assertAbsoluteURLs(v)
+        elif isinstance(data, list):
+            for i in data:
+                self.assertAbsoluteURLs(i)
+        else:
+            raise TypeError(
+                "assertAbsoluteURLs must be called with a dictionary or list"
+            )
+
     def assertAssetStatuses(self, asset_list, expected_statuses):
         asset_pks = [i["id"] for i in asset_list]
 
@@ -100,19 +165,157 @@ def assertAssetsHaveLatestTranscriptions(self, asset_list):
                 )
 
     def test_asset_list(self):
-        resp, data = self.get_asset_list(reverse("assets-list-json"))
+        resp, data = self.get_api_list_response(reverse("assets-list-json"))
+
+        self.assertAssetsHaveLatestTranscriptions(data["objects"])
 
     def test_transcribable_asset_list(self):
-        resp, data = self.get_asset_list(reverse("transcribe-assets-json"))
+        resp, data = self.get_api_list_response(reverse("transcribe-assets-json"))
 
         self.assertAssetStatuses(
             data["objects"],
             [TranscriptionStatus.NOT_STARTED, TranscriptionStatus.IN_PROGRESS],
         )
 
+        self.assertAssetsHaveLatestTranscriptions(data["objects"])
+
     def test_reviewable_asset_list(self):
-        resp, data = self.get_asset_list(reverse("review-assets-json"))
+        resp, data = self.get_api_list_response(reverse("review-assets-json"))
 
         self.assertAssetStatuses(data["objects"], [TranscriptionStatus.SUBMITTED])
 
         self.assertGreater(len(data["objects"]), 0)
+
+        self.assertAssetsHaveLatestTranscriptions(data["objects"])
+
+    def test_campaign_list(self):
+        resp, data = self.get_api_list_response(reverse("transcriptions:campaign-list"))
+
+        self.assertGreater(len(data["objects"]), 0)
+
+        test_campaigns = {
+            i["id"]: i
+            for i in Campaign.objects.published().values(
+                "id", "title", "description", "short_description", "slug"
+            )
+        }
+
+        for obj in data["objects"]:
+            self.assertIn("id", obj)
+            self.assertIn("url", obj)
+            self.assertDictContainsSubset(test_campaigns[obj["id"]], obj)
+
+    def test_campaign_detail(self):
+        resp, data = self.get_api_response(
+            reverse(
+                "transcriptions:campaign-detail",
+                kwargs={"slug": self.test_project.campaign.slug},
+            )
+        )
+
+        self.assertIn("object", data)
+        self.assertNotIn("objects", data)
+
+        serialized_project = data["object"]
+
+        self.assertIn("id", serialized_project)
+        self.assertIn("url", serialized_project)
+        campaign = self.test_project.campaign
+        self.assertDictContainsSubset(
+            {
+                "id": campaign.id,
+                "title": campaign.title,
+                "description": campaign.description,
+                "slug": campaign.slug,
+                "metadata": campaign.metadata,
+                "thumbnail_image": campaign.thumbnail_image,
+            },
+            serialized_project,
+        )
+        self.assertURLEqual(
+            serialized_project["url"], f"http://testserver{campaign.get_absolute_url()}"
+        )
+
+    def test_project_detail(self):
+        project = self.test_project
+
+        resp, data = self.get_api_list_response(project.get_absolute_url())
+
+        # Until we clean up the project view code, projects have two key
+        # elements: objects lists the children (i.e. items) and the project
+        # itself is in a second top-level “project” object:
+        self.assertIn("objects", data)
+        self.assertIn("project", data)
+        self.assertNotIn("object", data)
+
+        serialized_project = data["project"]
+
+        self.assertIn("id", serialized_project)
+        self.assertIn("url", serialized_project)
+
+        self.assertURLEqual(
+            serialized_project["url"], f"http://testserver{project.get_absolute_url()}"
+        )
+        self.assertDictContainsSubset(
+            {
+                "description": project.description,
+                "id": project.id,
+                "metadata": project.metadata,
+                "slug": project.slug,
+                "thumbnail_image": project.thumbnail_image,
+                "title": project.title,
+            },
+            serialized_project,
+        )
+
+        for obj in data["objects"]:
+            self.assertIn("description", obj)
+            self.assertIn("item_id", obj)
+            self.assertIn("item_url", obj)
+            self.assertIn("metadata", obj)
+            self.assertIn("thumbnail_url", obj)
+            self.assertIn("title", obj)
+            self.assertIn("url", obj)
+
+    def test_item_detail(self):
+        item = self.test_project.item_set.first()
+        resp, data = self.get_api_list_response(item.get_absolute_url())
+
+        # Until we clean up the project view code, projects have two key
+        # elements: objects lists the children (i.e. items) and the project
+        # itself is in a second top-level “project” object:
+        self.assertIn("objects", data)
+        self.assertIn("item", data)
+        self.assertNotIn("object", data)
+
+        serialized_item = data["item"]
+
+        self.assertIn("id", serialized_item)
+        self.assertIn("url", serialized_item)
+        self.assertIn("thumbnail_url", serialized_item)
+
+        self.assertURLEqual(
+            serialized_item["url"], f"http://testserver{item.get_absolute_url()}"
+        )
+        self.assertDictContainsSubset(
+            {
+                "description": item.description,
+                "id": item.id,
+                "item_id": item.item_id,
+                "metadata": item.metadata,
+                "title": item.title,
+            },
+            serialized_item,
+        )
+
+        for obj in data["objects"]:
+            self.assertIn("description", obj)
+            self.assertIn("difficulty", obj)
+            self.assertIn("metadata", obj)
+            self.assertIn("image_url", obj)
+            self.assertIn("thumbnail_url", obj)
+            self.assertIn("resource_url", obj)
+            self.assertIn("title", obj)
+            self.assertIn("slug", obj)
+            self.assertIn("url", obj)
+            self.assertIn("year", obj)

concordia/utils.py

@@ -2,6 +2,8 @@
 
 from django.contrib.auth.models import User
 
+from .templatetags.concordia_media_tags import asset_media_url
+
 
 def get_anonymous_user():
     """
@@ -26,3 +28,20 @@ def get_or_create_reservation_token(request):
     if "reservation_token" not in request.session:
         request.session["reservation_token"] = token_hex(25)
     return request.session["reservation_token"]
+
+
+def get_image_urls_from_asset(asset):
+    """
+    Given an Asset, return a tuple containing the normalized full-size and
+    thumbnail-size image URLs
+    """
+
+    image_url = asset_media_url(asset)
+    if asset.download_url and "iiif" in asset.download_url:
+        thumbnail_url = asset.download_url.replace(
+            "http://tile.loc.gov", "https://tile.loc.gov"
+        )
+    else:
+        thumbnail_url = image_url
+
+    return image_url, thumbnail_url

concordia/views.py

@@ -61,6 +61,7 @@
 from concordia.templatetags.concordia_media_tags import asset_media_url
 from concordia.utils import (
     get_anonymous_user,
+    get_image_urls_from_asset,
     get_or_create_reservation_token,
     request_accepts_json,
 )
@@ -651,6 +652,14 @@ def get_context_data(self, **kwargs):
 
     def serialize_context(self, context):
         data = super().serialize_context(context)
+
+        for i, asset in enumerate(context["object_list"]):
+            serialized_asset = data["objects"][i]
+            serialized_asset.pop("media_url")
+            image_url, thumbnail_url = get_image_urls_from_asset(asset)
+            serialized_asset["image_url"] = image_url
+            serialized_asset["thumbnail_url"] = thumbnail_url
+
         data["item"] = self.serialize_object(context["item"])
         return data
 
@@ -1474,13 +1483,7 @@ def serialize_object(self, obj):
         project = item.project
         campaign = project.campaign
 
-        image_url = asset_media_url(obj)
-        if obj.download_url and "iiif" in obj.download_url:
-            thumbnail_url = obj.download_url.replace(
-                "http://tile.loc.gov", "https://tile.loc.gov"
-            )
-        else:
-            thumbnail_url = image_url
+        image_url, thumbnail_url = get_image_urls_from_asset(obj)
 
         metadata = {
             "id": obj.pk,