Gemfile.lock

@@ -133,7 +133,7 @@ GEM
       mime-types (>= 1.16)
       ssrf_filter (~> 1.0)
     chronic (0.10.2)
-    cocina-models (0.59.0)
+    cocina-models (0.59.1)
       activesupport
       dry-struct (~> 1.0)
       dry-types (~> 1.1)
@@ -295,6 +295,7 @@ GEM
     i18n (1.8.10)
       concurrent-ruby (~> 1.0)
     ice_nine (0.11.2)
+    io-wait (0.1.0)
     iso-639 (0.2.10)
     jbuilder (2.11.2)
       activesupport (>= 5.0.0)
@@ -341,7 +342,8 @@ GEM
     mime-types-data (3.2021.0225)
     mini_exiftool (2.10.1)
     mini_mime (1.1.0)
-    mini_portile2 (2.5.1)
+    mini_portile2 (2.5.2)
+      net-ftp (~> 0.1)
     minitest (5.14.4)
     moab-versioning (4.4.1)
       druid-tools (>= 1.0.0)
@@ -356,15 +358,21 @@ GEM
     msgpack (1.4.2)
     multipart-post (2.1.1)
     mysql2 (0.4.10)
+    net-ftp (0.1.2)
+      net-protocol
+      time
+    net-protocol (0.1.1)
+      io-wait
+      timeout
     net-scp (3.0.0)
       net-ssh (>= 2.6.5, < 7.0.0)
     net-ssh (6.1.0)
     netrc (0.11.0)
     nio4r (2.5.7)
-    nokogiri (1.11.5)
+    nokogiri (1.11.6)
       mini_portile2 (~> 2.5.0)
       racc (~> 1.4)
-    nokogiri (1.11.5-x86_64-linux)
+    nokogiri (1.11.6-x86_64-linux)
       racc (~> 1.4)
     nokogiri-happymapper (0.8.1)
       nokogiri (~> 1.5)
@@ -489,7 +497,7 @@ GEM
       rubocop-ast (>= 1.5.0, < 2.0)
       ruby-progressbar (~> 1.7)
       unicode-display_width (>= 1.4.0, < 3.0)
-    rubocop-ast (1.5.0)
+    rubocop-ast (1.7.0)
       parser (>= 3.0.1.1)
     ruby-enum (0.9.0)
       i18n
@@ -556,6 +564,8 @@ GEM
     thor (1.1.0)
     thread_safe (0.3.6)
     tilt (2.0.10)
+    time (0.1.0)
+    timeout (0.1.1)
     twitter-typeahead-rails (0.11.1.pre.corejavascript)
       actionpack (>= 3.1)
       jquery-rails
@@ -574,7 +584,7 @@ GEM
       i18n
     warden (1.2.9)
       rack (>= 2.0.9)
-    websocket-driver (0.7.3)
+    websocket-driver (0.7.4)
       websocket-extensions (>= 0.1.0)
     websocket-extensions (0.1.5)
     whenever (0.11.0)

bin/export-items

@@ -7,6 +7,17 @@
 COLLECTION_DRUIDS_LIST = 'collection_druids.txt'
 GRAVEYARD_APO = 'druid:kg712km1576'
 
+# This list from: https://docs.google.com/spreadsheets/d/1Gu0TIUpNByNgNtFDn5MJARUvtsgcNXJAHiKf5s_NsUc/edit#gid=0
+# NOTE: we don't need to worry about the "hydrus object with DOI", because these are not returned in the query
+# for Hydrus::Items as they have been converted to Dor::Item.
+ITEMS_TO_EXCLUDE = %w[
+  druid:ty334nd6571
+  druid:bx749bs2681
+  druid:zc000fq4044
+  druid:pr213sh5046
+  druid:jr671fk0644
+]
+
 require_relative '../config/environment'
 
 # Monkey-patch AF to allow using HTTP POST (for querying items by their collection)
@@ -128,6 +139,9 @@ list.each do |item|
     elsif item.catkey.present?
       warn "Skipping #{item.pid} because it has a catkey"
       next
+    elsif ITEMS_TO_EXCLUDE.include? item.pid
+      warn "Skipping #{item.pid} because it is on the list of items to exclude"
+      next
     end
     attributes = serialize(item)
     puts attributes.compact.to_json