From 994bdf34f8cf849b206839879c5686f4658de0b8 Mon Sep 17 00:00:00 2001
From: Graham Hukill <ghukill@gmail.com>
Date: Wed, 7 Feb 2024 10:14:58 -0500
Subject: [PATCH] Set single URL for part of dct_references_s

Why these changes are being introduced:
It was discovered that the Aardvark format expects only a single URL for the dct_references_s
key 'http://schema.org/url', where previously we had set this as a list of URLs like
'http://schemal.org/downloadUrl'.

This also updates the hostname for where this URL points back to 'geodata.libraries.mit.edu' which
is the anticipated TIMDEX UI for geo resources.

How this addresses that need:
* Sets single URL for 'http://schema.org/url'
* Updates domain name

Side effects of this change:
* Requires an associated PR in Transmogirifer to expect a single value and not an array; forthcoming

Relevant ticket(s):
* https://mitlibraries.atlassian.net/browse/GDT-68
---
 harvester/records/record.py       | 15 +++++----------
 tests/test_records/test_record.py | 10 +++-------
 2 files changed, 8 insertions(+), 17 deletions(-)

diff --git a/harvester/records/record.py b/harvester/records/record.py
index ad87aae..f222210 100644
--- a/harvester/records/record.py
+++ b/harvester/records/record.py
@@ -424,18 +424,13 @@ def _dct_references_s_mit(self) -> dict:
                 "url": f"{cdn_root}/{cdn_folder}/{self.identifier}.zip",
             },
         ]
-        website_urls = [
-            {
-                "label": "Website",
-                "url": (
-                    "https://search.libraries.mit.edu/record/"
-                    f"gismit:{self.identifier.removeprefix('mit:')}"
-                ),
-            }
-        ]
+        website_url = (
+            "https://geodata.libraries.mit.edu/record/"
+            f"gismit:{self.identifier.removeprefix('mit:')}"
+        )
         return {
             "http://schema.org/downloadUrl": download_urls,
-            "http://schema.org/url": website_urls,
+            "http://schema.org/url": website_url,
         }
 
     def _dct_references_s_ogm(self) -> dict:
diff --git a/tests/test_records/test_record.py b/tests/test_records/test_record.py
index ce7f19f..c8c1679 100644
--- a/tests/test_records/test_record.py
+++ b/tests/test_records/test_record.py
@@ -175,13 +175,9 @@ def test_record_shared_field_method_dct_references_s_success(
                 "/SDE_DATA_AE_A8GNS_2003.zip",
             },
         ],
-        "http://schema.org/url": [
-            {
-                "label": "Website",
-                "url": "https://search.libraries.mit.edu/record/"
-                "gismit:SDE_DATA_AE_A8GNS_2003",
-            },
-        ],
+        "http://schema.org/url": (
+            "https://geodata.libraries.mit.edu/record/gismit:SDE_DATA_AE_A8GNS_2003"
+        ),
     }
     assert fgdc_source_record_from_zip._dct_references_s() == json.dumps(references)