cdm_migrator 3.2.0 → 3.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 330afa0ea8f2de397dc822fafef651778e039b0b422ec8ab93e61bd4ed6e9ced
-  data.tar.gz: 6860866888bf49430fbbb4eace6e38bb7b2da09f6b8633608e7508cac374dc4c
+  metadata.gz: 8bc6e448326179791de466ba5dc1d35a26fd48fa2a85a8fbabeb7d8dad60339e
+  data.tar.gz: f3cf46dec46cd8166bc0555eef28e2868db6fbbf77963083f93fca1958b00e19
 SHA512:
-  metadata.gz: f263a7da82234a60ac6ac4f73529b226762b483dee9ce093712f8dd3c34dc97b3b1b96b5b62cbc98d0de5f2dc1157367b5c1352a5cb79bd7703f45a707ceb034
-  data.tar.gz: f4f095574c88c63970d2de9032ad96ad134ff34952b78068a2ddceee404d9998134ccfac15ee29ca75d2ac9c7b8f4a3a4db33030b6218d8b100afa64ea357cd3
+  metadata.gz: 9d8b8bc4d318269375c82429cbcd4bb2e3d07b1761ef54854b262746c09e46430ee3d9da02122567f5244599bb3f312b7248b1b0a610b12abc2a049b0b7470dc
+  data.tar.gz: 4dfc8a920f09e1e3c84b5eb09116adadcbf44556c3abe4b4daa1dbce7e87e767ea380de9fbdc5368c548032bf4b6f074a5ab0268fd6600dff75a9530445a7210

data/app/controllers/cdm_migrator/csv_controller.rb

@@ -117,6 +117,7 @@ module CdmMigrator
       solr = RSolr.connect url: Account.find_by(tenant: Apartment::Tenant.current).solr_endpoint.url
       response = solr.get 'select', params: {
           q: "member_of_collection_ids_ssim:#{params[:collection_id]}",
+          rows: 3400,
           fl: "id"
       }
       unless response['response']['docs'].empty? || response['response']['docs'][0].empty?
@@ -269,13 +270,49 @@ module CdmMigrator
     def check_edtf(row_number, row)
       edtf_fields = @edtf_fields
       edtf_errors = edtf_fields.each_with_object({}) do |field, hash|
-        if Date.edtf(row[field]) == nil and row[field] != "unknown"
-          hash[field.to_s] = "Blank or not a valid EDTF date."
+        temp_date = row[field]
+        # modify date so that the interval encompasses the years on the last interval date
+        temp_date = temp_date.gsub('/..','').gsub('%','?~').gsub(/\/$/,'')
+        date = temp_date.include?("/") ? temp_date.gsub(/([0-9]+X+\/)([0-9]+)(X+)/){"#{$1}"+"#{$2.to_i+1}"+"#{$3}"}.gsub("X","u") : temp_date
+        date = date.gsub("XX-","uu-").gsub("X-", "u-").gsub('XX?','uu').gsub('X?', 'u').gsub('u?','u').gsub('?','')
+        # edtf has trouble with year-month (e.g. "19uu-12") or year-season strings (e.g. "190u-23")
+        # that contain unspecified years, or intervals containing the above ("19uu-22/19uu-23", etc.).
+        # So we check for/create exceptions.
+        # Check for season interval
+        if Date.edtf(date) == nil and date != "unknown" # Accept season intervals
+          unless is_season?(date.split("/").first) and is_season?(date.split("/").second)
+            # If an interval then, check each date individually
+            if date.include?("/")
+              dates = date.split("/")
+            else
+              dates = [date]
+            end
+            #byebug
+            dates.each do |d|
+              # Dates with 'u' in the last digit of the year return invalid when in format YYYY-MM
+              # So we flub day specifity before checking again if the date is valid
+              unless Date.edtf(d + '-01') # Date.edtf('193u-03-01') returns valid
+                if match = d[/\d{3}u/] or match = d[/\d{2}u{2}-[2][1-4]/] # edtf can't parse single u in year (e.g. 192u) or uu in YYYY-SS (e.g. 19uu-21), so we replace it
+                  d.gsub!(match, match.gsub("u","0"))
+                  unless Date.edtf(d)
+                    hash[field.to_s] = "Blank or not a valid EDTF date."
+                  end
+                else
+                  hash[field.to_s] = "Blank or not a valid EDTF date."
+                end
+              end
+            end
+          end
         end
+
       end
       @error_list[row_number] = edtf_errors
     end
 
+    def is_season?(date)
+      Date.edtf(date).class == EDTF::Season
+    end
+
     # <Example: should be http://rightsstatements.org/vocab/etc. NOT https://rightsstatements.org/page/etc.
     def check_uris(row_number, row)
       uri_fields = @uri_fields
@@ -294,7 +331,7 @@ module CdmMigrator
         value = row[field]
         if value.present?
           URI.extract(value).each { |uri| value.gsub!(uri, '') }
-          unless value.split("").all?(character) # Check if remaining characters are the correct separator
+          unless value.split("").all? { |sep| sep == character } # Check if remaining characters are the correct separator
             hash[field.to_s] = "May contain the wrong multi-value separator (i.e. not #{character})."
           end
         end

data/app/views/cdm_migrator/cdm/mappings.html.erb

@@ -30,7 +30,7 @@ border:1px solid black;
     </tr>
   <% end %>
   </table>
-  <%= select_tag "mappings_url", options_for_select(@dirs) if @cdm_dirs %>
+  <%= select_tag "mappings_url", options_for_select(@dirs.sort { |x,y| x[0].downcase <=> y[0].downcase }) if @cdm_dirs %>
   <%= hidden_field_tag "work", params[:work] %>
   <%= submit_tag 'generate csv'%>
 <% end %>

data/app/views/cdm_migrator/csv/_error_list.html.erb

@@ -0,0 +1,21 @@
+<% if @error_list && @error_list.any? %>
+  <%#= @error_list.inspect %>
+  <table class="table table-striped">
+    <thead>
+    <tr>
+      <th scope="col" style="min-width: 100px;">Line No.</th>
+      <th scope="col" style="margin-right: 0.5em;">Column(s)</th>
+      <th scope="col">Issue</th>
+    </tr>
+    </thead>
+    <tbody>
+    <% @error_list.keys.each do |line_number| %>
+      <tr>
+        <td><%= line_number %></td>
+        <td><%= @error_list[line_number].keys.join("<br />").html_safe %></td>
+        <td><%= @error_list[line_number].values.join("<br />").html_safe %></td>
+      </tr>
+    <% end %>
+    </tbody>
+  </table>
+<% end %>

data/lib/cdm_migrator/version.rb

@@ -1,3 +1,3 @@
 module CdmMigrator
-  VERSION = '3.2.0'
+  VERSION = '3.2.1'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: cdm_migrator
 version: !ruby/object:Gem::Version
-  version: 3.2.0
+  version: 3.2.1
 platform: ruby
 authors:
 - sephirothkod
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-08-24 00:00:00.000000000 Z
+date: 2020-11-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -100,9 +100,8 @@ files:
 - app/views/cdm_migrator/cdm/mappings.html.erb
 - app/views/cdm_migrator/csv/_batches_list.html.erb
 - app/views/cdm_migrator/csv/_default_group.html.erb
+- app/views/cdm_migrator/csv/_error_list.html.erb
 - app/views/cdm_migrator/csv/_list_batches.html.erb
-- app/views/cdm_migrator/csv/_path_list.html.erb
-- app/views/cdm_migrator/csv/_results_pagination.html.erb
 - app/views/cdm_migrator/csv/_tabs.html.erb
 - app/views/cdm_migrator/csv/csv_checker.html.erb
 - app/views/cdm_migrator/csv/edit.html.erb
@@ -125,7 +124,7 @@ homepage: https://github.com/UVicLibrary/cdm_migrator
 licenses:
 - MIT
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -140,9 +139,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.7
-signing_key: 
+rubygems_version: 3.1.2
+signing_key:
 specification_version: 4
 summary: ContentDM to Hyrax migrator.
 test_files: []

data/app/views/cdm_migrator/csv/_path_list.html.erb

@@ -1,19 +0,0 @@
-<% if @path_list && @path_list.any? %>
-    <table class="table table-striped">
-      <thead>
-        <tr>
-          <th scope="col" style="min-width: 100px;">Line No.</th>
-          <th scope="col">File Path (url)</th>
-        </tr>
-      </thead>
-      <tbody>
-        <% @path_list.each do |line, path| %>
-          <tr>
-            <td><%= line %></td>
-            <td><%= path %></td>
-          </tr>
-        <% end %>
-      </tbody>
-    </table>
-<% end %>
-

data/app/views/cdm_migrator/csv/_results_pagination.html.erb

@@ -1,9 +0,0 @@
-<% if @response.total_pages > 1 %>
- <div class="row record-padding">
-  <div class="col-md-9">
-    <div class="pagination">
-      <%= paginate @response, outer_window: 2, theme: 'blacklight', route_set: hyrax %>
-    </div>
-  </div>
- </div>
-<% end %>