Validate that solr doc has title

pulibrary · Nov 25, 2024 · 32ac1bb · 32ac1bb
1 parent 2fb2e09
commit 32ac1bb
Show file tree

Hide file tree

Showing 2 changed files with 46 additions and 0 deletions.
diff --git a/lib/dpul_collections/indexing_pipeline/figgy/hydration_cache_entry.ex b/lib/dpul_collections/indexing_pipeline/figgy/hydration_cache_entry.ex
@@ -34,8 +34,16 @@ defmodule DpulCollections.IndexingPipeline.Figgy.HydrationCacheEntry do
       page_count_i: page_count(metadata),
       image_service_urls_ss: image_service_urls(metadata, related_data)
     }
+    |> validate_solr_document()
   end
 
+  defp validate_solr_document(%{id: id, title_txtm: nil}) do
+    Logger.warning("Solr document is missing title: #{id}")
+    nil
+  end
+
+  defp validate_solr_document(doc), do: doc
+
   defp image_service_urls(%{"member_ids" => member_ids}, related_data) do
     member_ids
     |> Enum.map(&extract_service_url(&1, related_data))

diff --git a/test/dpul_collections/indexing_pipeline/figgy/hydration_cache_entry_test.exs b/test/dpul_collections/indexing_pipeline/figgy/hydration_cache_entry_test.exs
@@ -282,5 +282,43 @@ defmodule DpulCollections.IndexingPipeline.Figgy.HydrationCacheEntryTest do
       assert capture_log(fn -> HydrationCacheEntry.to_solr_document(entry) end) =~
                "couldn't parse date"
     end
+
+    @tag capture_log: true
+    test "a nil value is returned when a solr document is missing a title" do
+      {:ok, entry} =
+        IndexingPipeline.write_hydration_cache_entry(%{
+          cache_version: 0,
+          record_id: "f134f41f-63c5-4fdf-b801-0774e3bc3b2d",
+          source_cache_order: ~U[2018-03-09 20:19:36.465203Z],
+          data: %{
+            "id" => "f134f41f-63c5-4fdf-b801-0774e3bc3b2d",
+            "internal_resource" => "EphemeraFolder",
+            "metadata" => %{
+              "date_created" => ["2022"]
+            }
+          }
+        })
+
+      assert HydrationCacheEntry.to_solr_document(entry) == nil
+    end
+
+    test "a message is logged when a solr document is missing a title" do
+      {:ok, entry} =
+        IndexingPipeline.write_hydration_cache_entry(%{
+          cache_version: 0,
+          record_id: "f134f41f-63c5-4fdf-b801-0774e3bc3b2d",
+          source_cache_order: ~U[2018-03-09 20:19:36.465203Z],
+          data: %{
+            "id" => "f134f41f-63c5-4fdf-b801-0774e3bc3b2d",
+            "internal_resource" => "EphemeraFolder",
+            "metadata" => %{
+              "date_created" => ["2022"]
+            }
+          }
+        })
+
+      assert capture_log(fn -> HydrationCacheEntry.to_solr_document(entry) end) =~
+               "missing title: f134f41f-63c5-4fdf-b801-0774e3bc3b2d"
+    end
   end
 end