Fix edge cases and add tests

Signed-off-by: Olga Bulat <obulat@gmail.com>

openverse_catalog/dags/providers/provider_api_scripts/smk.py

@@ -108,7 +108,7 @@ def _handle_items_data(
                 image_url=img.get("image_url"),
                 height=img.get("height"),
                 width=img.get("width"),
-                filesize=img.get("image_size"),
+                filesize=img.get("filesize"),
                 filetype=extract_filetype(img.get("image_url", ""), "image"),
                 license_info=license_info,
                 thumbnail_url=img.get("thumbnail"),
@@ -137,14 +137,15 @@ def _get_images(item):
 
         height = item.get("image_height")
         width = item.get("image_width")
-
+        filesize = item.get("image_size") or item.get("size")
         images.append(
             {
                 "id": id,
                 "image_url": image_url,
                 "thumbnail": thumbnail_url,
                 "height": height,
                 "width": width,
+                "filesize": filesize,
             }
         )
 
@@ -160,13 +161,16 @@ def _get_images(item):
                 image_url, thumbnail_url = _get_image_urls(iiif_id)
                 height = alt_img.get("height")
                 width = alt_img.get("width")
+                filesize = alt_img.get("image_size") or alt_img.get("size")
+
                 images.append(
                     {
                         "id": iiif_id,
                         "image_url": image_url,
                         "thumbnail": thumbnail_url,
                         "height": height,
                         "width": width,
+                        "filesize": filesize,
                     }
                 )
     return images

tests/dags/providers/provider_api_scripts/resources/smk/expected_image_data_hq.json

@@ -1,12 +1,14 @@
 [
   {
+    "filesize": 11784886,
     "height": 1059,
     "id": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.reconstructed.tif.jp2",
     "image_url": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.reconstructed.tif.jp2/full/!2048,/0/default.jpg",
     "thumbnail": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.reconstructed.tif.jp2/full/!400,/0/default.jpg",
     "width": 3887
   },
   {
+    "filesize": 19269857,
     "height": 1576,
     "id": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2",
     "image_url": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/full/!2048,/0/default.jpg",

tests/dags/providers/provider_api_scripts/resources/smk/expected_image_data_legacy.json

@@ -1,12 +1,14 @@
 [
   {
+    "filesize": 11784886,
     "height": 1059,
     "id": "1170012466_object",
     "image_url": "https://api.smk.dk/api/v1/thumbnail/52f00edc-936e-42a7-950b-d0cd0df3864b.jpg",
     "thumbnail": "https://api.smk.dk/api/v1/thumbnail/52f00edc-936e-42a7-950b-d0cd0df3864b.jpg",
     "width": 3887
   },
   {
+    "filesize": 19269857,
     "height": 1576,
     "id": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2",
     "image_url": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/full/!2048,/0/default.jpg",

tests/dags/providers/provider_api_scripts/resources/smk/expected_image_data_partial.json

@@ -1,5 +1,6 @@
 [
   {
+    "filesize": 19269857,
     "height": 1576,
     "id": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2",
     "image_url": "https://iip.smk.dk/iiif/jp2/KKSgb6458.tif.jp2/full/!2048,/0/default.jpg",

tests/dags/providers/provider_api_scripts/test_smk.py

@@ -1,21 +1,29 @@
 import json
 import logging
-import os
+from pathlib import Path
 from unittest.mock import MagicMock, patch
 
 import requests
+from common.licenses import LicenseInfo
 from providers.provider_api_scripts import smk
 
 
-RESOURCES = os.path.join(os.path.abspath(os.path.dirname(__file__)), "resources/smk")
+RESOURCES = Path(__file__).parent.resolve() / "resources/smk"
 
 logging.basicConfig(
     format="%(asctime)s - %(name)s - %(levelname)s:  %(message)s", level=logging.DEBUG
 )
 
+CC0 = LicenseInfo(
+    "cc0",
+    "1.0",
+    "https://creativecommons.org/publicdomain/zero/1.0/",
+    None,
+)
+
 
 def _get_resource_json(json_name):
-    with open(os.path.join(RESOURCES, json_name)) as f:
+    with open(RESOURCES / json_name) as f:
         resource_json = json.load(f)
     return resource_json
 
@@ -106,6 +114,32 @@ def test_handle_items_data_success():
     assert actual_image_count == 1
 
 
+def test_handle_items_data_success_data():
+    items = _get_resource_json("items_batch.json")
+    with patch.object(smk.image_store, "add_item") as mock_add_item:
+        smk._handle_items_data(items)
+    _, actual_image = mock_add_item.call_args
+    expected_image = {
+        "foreign_identifier": "https://iip.smk.dk/iiif/jp2/kks1615.tif.jp2",
+        "foreign_landing_url": "https://open.smk.dk/en/artwork/image/KKS1615",
+        "image_url": "https://iip.smk.dk/iiif/jp2/kks1615.tif.jp2/full/!2048,/0/default.jpg",
+        "height": 5141,
+        "width": 3076,
+        "filesize": 47466428,
+        "filetype": "jpg",
+        "license_info": CC0,
+        "thumbnail_url": "https://iip.smk.dk/iiif/jp2/kks1615.tif.jp2/full/!400,/0/default.jpg",
+        "creator": "Altdorfer, Albrecht",
+        "title": "Jomfru Maria med barnet og Sankt Anne ved vuggen",
+        "meta_data": {
+            "created_date": "2020-03-21T10:18:17Z",
+            "collection": "Gammel bestand",
+            "techniques": "Kobberstik",
+        },
+    }
+    assert actual_image == expected_image
+
+
 def test_handle_items_data_failure():
     items = []
     with patch.object(smk.image_store, "add_item", return_value=None) as mock_add_item: