Encode spaces in urls for SMK (#3439)

* Encode spaces in urls for smk

* Undo change to default for missing urls

* Make a separate test for spaces

* Undo change to legacy test

Author: stacimc

catalog/dags/providers/provider_api_scripts/smk.py

@@ -126,6 +126,8 @@ def get_record_data(self, data: dict) -> dict | None:
             return None
 
         thumbnail_url = data.get("image_thumbnail")
+        if thumbnail_url:
+            thumbnail_url = thumbnail_url.replace(" ", "%20")
         height = data.get("image_height")
         width = data.get("image_width")
         filesize = data.get("image_size") or data.get("size")
@@ -135,7 +137,7 @@ def get_record_data(self, data: dict) -> dict | None:
             "foreign_landing_url": foreign_landing_url,
             "license_info": license_info,
             "title": self._get_title(data),
-            "url": url,
+            "url": url.replace(" ", "%20"),
             "thumbnail_url": thumbnail_url,
             "height": height,
             "width": width,

catalog/tests/dags/providers/provider_api_scripts/resources/smk/expected_image_data_spaces.json

@@ -0,0 +1,12 @@
+{
+  "filesize": 11784886,
+  "height": 1059,
+  "url": "https://api.smk.dk/api/v1/thumbnail/KKSgb5743%20verso.jpg",
+  "thumbnail_url": "https://api.smk.dk/api/v1/thumbnail/KKSgb5743%20verso.jpg",
+  "width": 3887,
+  "creator": null,
+  "foreign_identifier": "1170012466_object",
+  "foreign_landing_url": "https://open.smk.dk/en/artwork/image/KKSgb5743%20verso",
+  "meta_data": {},
+  "title": null
+}

catalog/tests/dags/providers/provider_api_scripts/resources/smk/image_data_spaces.json

@@ -0,0 +1,23 @@
+{
+  "alternative_images": [
+    {
+      "height": 1576,
+      "iiif_id": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2",
+      "iiif_info": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/info.json",
+      "mime_type": "image/tiff",
+      "native": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/full/full/0/native.jpg",
+      "orientation": "landscape",
+      "size": 19269857,
+      "thumbnail": "https://iip-thumb.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/full/!1024,/0/default.jpg",
+      "width": 4073
+    }
+  ],
+  "id": "1170012466_object",
+  "object_number": "KKSgb5743 verso",
+  "image_height": 1059,
+  "image_native": "https://api.smk.dk/api/v1/thumbnail/KKSgb5743 verso.jpg",
+  "image_size": 11784886,
+  "image_thumbnail": "https://api.smk.dk/api/v1/thumbnail/KKSgb5743 verso.jpg",
+  "image_width": 3887,
+  "rights": "https://creativecommons.org/share-your-work/public-domain/cc0/"
+}

catalog/tests/dags/providers/provider_api_scripts/test_smk.py

@@ -177,3 +177,12 @@ def test_get_record_data_returns_none_with_falsy_required_property(property_name
         item[prop] = ""
 
     assert smk.get_record_data(item) is None
+
+
+def test_get_record_data_with_spaces_in_urls():
+    item = _get_resource_json("image_data_spaces.json")
+    expected_image_data = _get_resource_json("expected_image_data_spaces.json")
+    expected_record_data = {**expected_image_data, "license_info": CC0_SMK}
+    actual_record_data = smk.get_record_data(item)
+
+    assert actual_record_data == expected_record_data