spotlight-dor-resources 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b86fa8d9250353055a9f68632944a996fa4e0523
-  data.tar.gz: d53448b9b23303463ed7db3c54d7f02bff763739
+  metadata.gz: 31130b0ef44659b04e226ce3f7d54afeb2fab9bd
+  data.tar.gz: 773201c4913261fb010cbd06a7bd027fb550acc1
 SHA512:
-  metadata.gz: 7df884fe6f4617270b56b97dccb83630f6cae9ac2640c6a28d4a7763c702ae1e3b77f4ca300647b952b185fefd577096742ceedd4e17f584b34ffee1068ec676
-  data.tar.gz: a9757073e8b1e7a92df0e2cc39d00ba8df3d028aed15c9f776a907c0b546b19405b08a46535742bd09700819f1e8480c673443ad59e0589912c6ce70beb07608
+  metadata.gz: dd69ded57b4a0daaa03b544218a95edb3b23a503ae54c56274f7f8cc06bb8949120f19af29b9047b81d743914dd2e601fbeb5b4a133edfe2718084457290a096
+  data.tar.gz: 90f34128fd6494c255f69f9cf4b430cdae7c060618ba648e99a982c3fdfdf2d41ade20b73d6febe48a8eae46b7dbf5823996b5469e25b756340a84fb217a2f04

data/.gitignore

@@ -12,6 +12,7 @@ capybara-*.html
 /db/*.sqlite3
 /public/system/*
 /coverage/
+/pkg/
 /spec/tmp/*
 **.orig
 rerun.txt

data/.rubocop.yml

@@ -6,3 +6,7 @@ AllCops:
   Exclude:
     - 'spec/internal/**/*'
     - 'spec/test_app_templates/**/*'
+
+Metrics/ClassLength:
+  Exclude:
+    - 'lib/spotlight/dor/indexer.rb'

data/.rubocop_todo.yml

@@ -6,9 +6,6 @@
 # Note that changes in the inspected code, or installation of new
 # versions of RuboCop, may require this file to be generated again.
 
-# Offense count: 2
-Metrics/AbcSize:
-  Max: 33
 
 # Offense count: 1
 # Configuration parameters: CountComments.

data/lib/spotlight/dor/indexer.rb

@@ -31,11 +31,12 @@ module Spotlight::Dor
     end
 
     # add fields from raw mods
-    # see comment with add_donor_tags about Feigenbaum specific donor tags data
     before_index :add_box
+    # see comment with add_donor_tags about Feigenbaum specific donor tags data
     before_index :add_donor_tags
     before_index :add_genre
     before_index :add_folder
+    before_index :add_folder_name
     before_index :add_series
     before_index :mods_cartographics_indexing
 
@@ -72,7 +73,7 @@ module Spotlight::Dor
 
     # This new donor_tags_sim field was added in October 2015 specifically for the Feigenbaum exhibit.  It is very likely
     #  it will go ununsed by other projects, but should be benign (since this field will not be created if this specific MODs note is not found.)
-    #  Later refactoring could include project specific fields.   Peter Mangiafico
+    #  Later refactoring could include exhibit specific fields.   Peter Mangiafico
     def add_donor_tags sdb, solr_doc
       donor_tags = sdb.smods_rec.note.select { |n| n.displayLabel == 'Donor tags' }.map(&:content)
       insert_field solr_doc, 'donor_tags', donor_tags, :symbol # this is a _ssim field
@@ -86,21 +87,30 @@ module Spotlight::Dor
       #   _location.physicalLocation should find top level and relatedItem
       folder_num = sdb.smods_rec._location.physicalLocation.map do |node|
         val = node.text
-        # folder may be text with commas
-        match_data = val.match(/Folder:? ?(.+)/i)
-        next if match_data.blank?
-        result = match_data[1].strip
-        # Menuez collection may have folder followed by Sleeve then Frame
-        match2_data = result.match(/(.*),? ?Sleeve/i)
-        if match2_data
-          match2_data[1].strip.sub(/,$/, '')
-        else
-          result
-        end
+
+        match_data = if val =~ /\|/
+                       # we assume the data is pipe-delimited, and may contain commas within values
+                       val.match(/Folder ?:? ?([^|]+)/)
+                     else
+                       # the data should be comma-delimited, and may not contain commas within values
+                       val.match(/Folder ?:? ?([^,]+)/)
+                     end
+
+        match_data[1].strip if match_data.present?
       end
+
       solr_doc['folder_ssi'] = folder_num.first if folder_num.present?
     end
 
+    # add the folder name to solr_doc as folder_name_ssi field (note: single valued!)
+    #   data is specific to Feigenbaum collection and is in <note type='preferred citation'>
+    def add_folder_name(sdb, solr_doc)
+      # see spec for data examples
+      preferred_citation = sdb.smods_rec.note.select { |n| n.type_at == 'preferred citation' }.map(&:content)
+      match_data = preferred_citation.first.match(/Title: +(.+)/i) if preferred_citation.present?
+      solr_doc['folder_name_ssi'] = match_data[1].rstrip if match_data.present?
+    end
+
     # add plain MODS <genre> element data, not the SearchWorks genre values
     def add_genre sdb, solr_doc
       insert_field solr_doc, 'genre', sdb.smods_rec.genre.content, :symbol # this is a _ssim field
@@ -121,6 +131,7 @@ module Spotlight::Dor
       solr_doc['series_ssi'] = series_num.first if series_num.present?
     end
 
+    # rubocop:disable Metrics/AbcSize
     def mods_cartographics_indexing sdb, solr_doc
       insert_field(solr_doc, "coordinates", Array(sdb.smods_rec.subject.cartographics.coordinates).map { |n| n.text }, :stored_searchable)
 
@@ -138,6 +149,7 @@ module Spotlight::Dor
         solr_doc["point_bbox"] << "#{minX} #{minY} #{maxX} #{maxY}"
       end
     end
+    # rubocop:enable Metrics/AbcSize
 
     def coord_to_decimal point
       regex = /(?<dir>[NESW])\s*(?<deg>\d+)°(?:(?<sec>\d+)ʹ)?/

data/lib/spotlight/dor/resources/version.rb

@@ -1,7 +1,7 @@
 module Spotlight
   module Dor
     module Resources
-      VERSION = "0.0.1"
+      VERSION = "0.0.2"
     end
   end
 end

data/spec/unit/spotlight/dor/indexer_spec.rb

@@ -172,7 +172,9 @@ describe Spotlight::Dor::Indexer do
       'Flat-box 228 | Volume 1': nil,
       # shpc (actually in <relatedItem><location><physicalLocation>)
       'Series Biographical Photographs | Box 1 | Folder Abbot, Nathan': 'Abbot, Nathan',
-      'Series General Photographs | Box 1 | Folder Administration building--Outer Quad': 'Administration building--Outer Quad'
+      'Series General Photographs | Box 1 | Folder Administration building--Outer Quad': 'Administration building--Outer Quad',
+      # hypothetical
+      'Folder: 42, Sheet: 15': '42'
     }.each do |example, expected|
       describe "for example '#{example}'" do
         let(:example) { example }
@@ -197,4 +199,64 @@ describe Spotlight::Dor::Indexer do
       end # for example
     end # each
   end # add_folder
+
+  let(:mods_note_plain) do
+    Nokogiri::XML <<-EOF
+      <mods xmlns="#{Mods::MODS_NS}">
+        <note>#{example}</note>
+      </mods>
+    EOF
+  end
+  let(:mods_note_preferred_citation) do
+    Nokogiri::XML <<-EOF
+      <mods xmlns="#{Mods::MODS_NS}">
+        <note type="preferred citation">#{example}</note>
+      </mods>
+    EOF
+  end
+  describe "#add_folder_name" do
+    # example string as key, expected folder name as value
+    # all from feigenbaum (or based on feigenbaum), as that is only coll
+    {
+      'Call Number: SC0340, Accession: 1986-052, Box: 20, Folder: 40, Title: S': 'S',
+      'Call Number: SC0340, Accession: 1986-052, Box: 54, Folder: 25, Title: Balzer': 'Balzer',
+      'Call Number: SC0340, Accession: 1986-052, Box : 30, Folder: 21, Title: Feigenbaum, Publications. 2 of 2.': 'Feigenbaum, Publications. 2 of 2.',
+      # colon in name
+      'Call Number: SC0340, Accession 2005-101, Box: 10, Folder: 26, Title: Gordon Bell Letter rdf:about blah (AI) 1987': 'Gordon Bell Letter rdf:about blah (AI) 1987',
+      'Call Number: SC0340, Accession 2005-101, Box: 11, Folder: 74, Title: Microcomputer Systems Proposal: blah blah': 'Microcomputer Systems Proposal: blah blah',
+      'Call Number: SC0340, Accession 2005-101, Box: 14, Folder: 20, Title: blah "bleah: blargW^"ugh" seriously?.': 'blah "bleah: blargW^"ugh" seriously?.',
+      # quotes in name
+      'Call Number: SC0340, Accession 2005-101, Box: 29, Folder: 18, Title: "bleah" blah': '"bleah" blah',
+      'Call Number: SC0340, Accession 2005-101, Box: 11, Folder: 58, Title: "M": blah': '"M": blah',
+      'Call Number: SC0340, Accession 2005-101, Box : 32A, Folder: 19, Title: blah "bleah" blue': 'blah "bleah" blue',
+      # not parseable
+      'Call Number: SC0340, Accession 2005-101': nil,
+      'Call Number: SC0340, Accession: 1986-052': nil,
+      'Call Number: SC0340, Accession: 1986-052, Box 36 Folder 38': nil,
+      'blah blah ... with the umbrella title Feigenbaum and Feldman, Computers and Thought II. blah blah': nil,
+      'blah blah ... Title ... blah blah': nil
+    }.each do |example, expected|
+      describe "for example '#{example}'" do
+        let(:example) { example }
+        context 'in preferred citation note' do
+          before do
+            allow(r).to receive(:mods).and_return(mods_note_preferred_citation)
+            subject.send(:add_folder_name, sdb, solr_doc)
+          end
+          it "has the expected folder name '#{expected}'" do
+            expect(solr_doc['folder_name_ssi']).to eq expected
+          end
+        end
+        context 'in plain note' do
+          before do
+            allow(r).to receive(:mods).and_return(mods_note_plain)
+            subject.send(:add_folder_name, sdb, solr_doc)
+          end
+          it 'does not have a folder name' do
+            expect(solr_doc['folder_name_ssi']).to be_falsey
+          end
+        end
+      end # for example
+    end # each
+  end # add_folder_name
 end

data/spotlight-dor-resources.gemspec

@@ -9,7 +9,7 @@ Gem::Specification.new do |spec|
   spec.authors       = ["Chris Beer"]
   spec.email         = ["cabeer@stanford.edu"]
   spec.summary       = "Spotlight resource indexer for DOR resources."
-  spec.homepage      = ""
+  spec.homepage      = "https://github.com/sul-dlss/spotlight-dor-resources"
   spec.license       = "Apache 2"
 
   spec.files         = `git ls-files -z`.split("\x0")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: spotlight-dor-resources
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Chris Beer
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-10-29 00:00:00.000000000 Z
+date: 2015-10-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: faraday
@@ -292,7 +292,7 @@ files:
 - spec/vcr_cassettes/indexer_integration_tests/genre/no_genre_ssim_field_when_genre_not_in_MODS.yml
 - spec/vcr_cassettes/indexer_integration_tests/genre/solr_doc_has_genre_ssim_field_when_genre_in_MODS.yml
 - spotlight-dor-resources.gemspec
-homepage: ''
+homepage: https://github.com/sul-dlss/spotlight-dor-resources
 licenses:
 - Apache 2
 metadata: {}