handle pending blob copy operations

subjects may create multiple copy blob operations to backfill the training data - do not attempt to copy one if we've scheduled one already

app/services/storage/training_data_sync.rb

@@ -2,7 +2,7 @@
 
 module Storage
   class TrainingDataSync
-    COPY_OPERATION_SUCCESS_CODE = 'success'
+    COPY_OPERATION_OK_CODES = %w[success pending].freeze
     PROD_CONTAINER_URL_PREFIX = 'https://panoptesuploads.blob.core.windows.net/public'
     STAGING_CONTAINER_URL_PREFIX = 'https://panoptesuploadsstaging.blob.core.windows.net/public'
 
@@ -13,7 +13,7 @@ def initialize(src_image_url)
     end
 
     def run
-      return if image_url_blob_already_copied
+      return if image_url_blob_already_copied_or_pending
 
       _copy_id, _copy_status = blob_service_client.copy_blob_from_uri(
         Rails.env,
@@ -22,9 +22,12 @@ def run
       )
     end
 
-    def image_url_blob_already_copied
+    def image_url_blob_already_copied_or_pending
       response = blob_service_client.get_blob_properties(Rails.env, blob_destination_path)
-      response.properties[:copy_status] == COPY_OPERATION_SUCCESS_CODE
+      # skip requesting a copy operation if the blob is already copied or pending / scheduled for copy
+      # failure / aborts will return false and schedule a new copy
+      # https://docs.microsoft.com/en-us/rest/api/storageservices/Copy-Blob?redirectedfrom=MSDN#working-with-a-pending-copy-operation-version-2012-02-12-and-newer
+      COPY_OPERATION_OK_CODES.include?(response.properties[:copy_status])
     rescue Azure::Core::Http::HTTPError => _e
       # treat all errors as a failure and attempt to re-copy
       # we can refine this via the response status code

spec/services/storage/training_data_sync_spec.rb

@@ -60,5 +60,29 @@
         expect(blob_service_client_double).not_to have_received(:copy_blob_from_uri)
       end
     end
+
+    context 'when the blob is pending copy' do
+      let(:blob_copy_properties) do
+        {
+          blob_type: 'BlockBlob',
+          copy_id: '6569935c-a8a1-4f90-819a-e4258271e163',
+          copy_status: 'pending',
+          copy_source: src_image_url,
+          copy_progress: '0/551722',
+          copy_completion_time: nil,
+          copy_status_description: nil
+        }
+      end
+
+      before do
+        allow(blob_instance_double).to receive(:properties).and_return(blob_copy_properties)
+        allow(blob_service_client_double).to receive(:get_blob_properties).and_return(blob_instance_double)
+      end
+
+      it 'does not call the blob copy uri function' do
+        data_syncer.run
+        expect(blob_service_client_double).not_to have_received(:copy_blob_from_uri)
+      end
+    end
   end
 end