From 9f2f6005d476b0bc0c88b8f1bb5954a628fd3c91 Mon Sep 17 00:00:00 2001
From: Kirk Wang <kirk.wang@scientist.com>
Date: Fri, 10 May 2024 08:09:39 -0700
Subject: [PATCH] =?UTF-8?q?=F0=9F=90=9B=20Fix=20FileSet=20csv=20rows=20for?=
 =?UTF-8?q?=20AF=20imports?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This commit will fix the functionality for adding a FileSet row in the
CSV to add additional metadata to FileSets.  The problem was located in
`Bulkrax::ObjectFactory#create_file_set`.  When it tries to find a
record, it actually comes back as a `FalseClass` because of the new
logic introduced recently which breaks further down.  If we add raise an
error in our `Bulkrax::ImportFileSetJob#find_parent_record` which is at
a higher point than before, then we can requeue the job and check again.
If the parent record is created by then, we will execute the rest of the
job.
---
 app/jobs/bulkrax/import_file_set_job.rb | 35 +++++++++++++------------
 1 file changed, 18 insertions(+), 17 deletions(-)

diff --git a/app/jobs/bulkrax/import_file_set_job.rb b/app/jobs/bulkrax/import_file_set_job.rb
index e0bc7175..65780a6e 100644
--- a/app/jobs/bulkrax/import_file_set_job.rb
+++ b/app/jobs/bulkrax/import_file_set_job.rb
@@ -16,7 +16,12 @@ def perform(entry_id, importer_run_id)
       # e.g. "parents" or "parents_1"
       parent_identifier = (entry.raw_metadata[entry.related_parents_raw_mapping] || entry.raw_metadata["#{entry.related_parents_raw_mapping}_1"])&.strip
 
-      validate_parent!(parent_identifier)
+      begin
+        validate_parent!(parent_identifier)
+      rescue MissingParentError => e
+        handle_retry(entry, importer_run_id, e)
+        return
+      end
 
       entry.build
       if entry.succeeded?
@@ -32,17 +37,6 @@ def perform(entry_id, importer_run_id)
       entry.save!
       entry.importer.current_run = ImporterRun.find(importer_run_id)
       entry.importer.record_status
-
-    rescue MissingParentError => e
-      # try waiting for the parent record to be created
-      entry.import_attempts += 1
-      entry.save!
-      if entry.import_attempts < 5
-        ImportFileSetJob.set(wait: (entry.import_attempts + 1).minutes).perform_later(entry_id, importer_run_id)
-      else
-        ImporterRun.decrement_counter(:enqueued_records, importer_run_id) # rubocop:disable Rails/SkipsModelValidations
-        entry.set_status_info(e)
-      end
     end
 
     private
@@ -54,14 +48,9 @@ def validate_parent!(parent_identifier)
       return if parent_identifier.blank?
 
       find_parent_record(parent_identifier)
-      check_parent_exists!(parent_identifier)
       check_parent_is_a_work!(parent_identifier)
     end
 
-    def check_parent_exists!(parent_identifier)
-      raise MissingParentError, %(Unable to find a record with the identifier "#{parent_identifier}") if parent_record.nil?
-    end
-
     def check_parent_is_a_work!(parent_identifier)
       case parent_record
       when Bulkrax.collection_model_class, Bulkrax.file_model_class
@@ -72,6 +61,18 @@ def check_parent_is_a_work!(parent_identifier)
 
     def find_parent_record(parent_identifier)
       _, @parent_record = find_record(parent_identifier, importer_run_id)
+      raise MissingParentError, %(Unable to find a record with the identifier "#{parent_identifier}") unless parent_record
+    end
+
+    def handle_retry(entry, importer_run_id, e)
+      entry.import_attempts += 1
+      entry.save!
+      if entry.import_attempts < 5
+        ImportFileSetJob.set(wait: (entry.import_attempts + 1).minutes).perform_later(entry.id, importer_run_id)
+      else
+        ImporterRun.decrement_counter(:enqueued_records, importer_run_id) # rubocop:disable Rails/SkipsModelValidations
+        entry.set_status_info(e)
+      end
     end
   end
 end