Merge pull request #500 from evidentpoint/doc-equivalence

Use a URN derived from DC meta tags for creating equivalent doc links

src/annotator/plugin/document.coffee

@@ -159,6 +159,21 @@ module.exports = class Document extends Plugin
           if id[0..3] == "doi:"
             @metadata.link.push(href: id)
 
+    # look for a link to identify the resource in dublincore metadata
+    dcRelationValues = @metadata.dc['relation.ispartof']
+    dcIdentifierValues = @metadata.dc['identifier']
+    if dcRelationValues && dcIdentifierValues
+      dcUrnRelationComponent =
+        dcRelationValues[dcRelationValues.length - 1]
+      dcUrnIdentifierComponent =
+        dcIdentifierValues[dcIdentifierValues.length - 1]
+      dcUrn = 'urn:x-dc:' +
+        encodeURIComponent(dcUrnRelationComponent) + '/' +
+        encodeURIComponent(dcUrnIdentifierComponent)
+      @metadata.link.push(href: dcUrn)
+      # set this as the documentFingerprint as a hint to include this in search queries
+      @metadata.documentFingerprint = dcUrn
+
   _getFavicon: =>
     for link in $("link")
       if $(link).prop("rel") in ["shortcut icon", "icon"]
@@ -180,4 +195,4 @@ module.exports = class Document extends Plugin
     if new URL(href).protocol != new URL(baseURI).protocol
       # use the baseURI instead since it's likely what's intended
       href = baseURI
-    return href
\ No newline at end of file
+    return href

src/annotator/plugin/test/document-test.coffee

@@ -41,7 +41,8 @@ describe 'Document', ->
     head.append('<meta name="citation_title" content="Foo">')
     head.append('<meta name="citation_pdf_url" content="foo.pdf">')
     head.append('<meta name="dc.identifier" content="doi:10.1175/JCLI-D-11-00015.1">')
-    head.append('<meta name="dc:identifier" content="isbn:123456789">')
+    head.append('<meta name="dc:identifier" content="foobar-abcxyz">')
+    head.append('<meta name="dc.relation.ispartof" content="isbn:123456789">')
     head.append('<meta name="DC.type" content="Article">')
     head.append('<meta property="og:url" content="http://example.com">')
     head.append('<meta name="twitter:site" content="@okfn">')
@@ -65,7 +66,7 @@ describe 'Document', ->
 
     it 'should have links with absolute hrefs and types', ->
       assert.ok(metadata.link)
-      assert.equal(metadata.link.length, 9)
+      assert.equal(metadata.link.length, 10)
       assert.match(metadata.link[0].href, docBaseUri)
       assert.equal(metadata.link[1].rel, "alternate")
       assert.match(metadata.link[1].href, /^.+foo\.pdf$/)
@@ -84,8 +85,16 @@ describe 'Document', ->
       assert.equal(metadata.link[7].type, "application/pdf")
       assert.equal(metadata.link[8].href, "doi:10.1175/JCLI-D-11-00015.1")
 
+      # Link derived from dc resource identifiers in the form of urn:x-dc:<container>/<identifier>
+      # Where <container> is the percent-encoded value of the last dc.relation.ispartof meta element
+      # and <identifier> is the percent-encoded value of the last dc.identifier meta element.
+      assert.equal(
+        metadata.link[9].href
+        "urn:x-dc:isbn%3A123456789/foobar-abcxyz"
+      )
+
     it 'should ignore atom and RSS feeds and alternate languages', ->
-      assert.equal(metadata.link.length, 9)
+      assert.equal(metadata.link.length, 10)
 
     it 'should have highwire metadata', ->
       assert.ok(metadata.highwire)
@@ -95,7 +104,8 @@ describe 'Document', ->
 
     it 'should have dublincore metadata', ->
       assert.ok(metadata.dc)
-      assert.deepEqual(metadata.dc.identifier, ["doi:10.1175/JCLI-D-11-00015.1", "isbn:123456789"])
+      assert.deepEqual(metadata.dc.identifier, ["doi:10.1175/JCLI-D-11-00015.1", "foobar-abcxyz"])
+      assert.deepEqual(metadata.dc['relation.ispartof'], ["isbn:123456789"])
       assert.deepEqual(metadata.dc.type, ["Article"])
 
     it 'should have facebook metadata', ->
@@ -116,7 +126,7 @@ describe 'Document', ->
 
     it 'should have unique uris', ->
       uris = testDocument.uris()
-      assert.equal(uris.length, 7)
+      assert.equal(uris.length, 8)
 
     it 'uri() returns the canonical uri', ->
       uri = testDocument.uri()
@@ -128,6 +138,10 @@ describe 'Document', ->
         'http://example.com/images/icon.ico'
       )
 
+    it 'should have a documentFingerprint as the dc resource identifiers URN href', ->
+      assert.equal(metadata.documentFingerprint, metadata.link[9].href)
+
+
   describe '#_absoluteUrl', ->
 
     it 'should add the protocol when the url starts with two slashes', ->
@@ -168,4 +182,4 @@ describe 'Document', ->
         'document-base-uri': document.location.href
       })
       altDocument = new Document($('<div></div>')[0], {})
-      assert.equal(altDocument._getDocumentHref(), document.location.href)
\ No newline at end of file
+      assert.equal(altDocument._getDocumentHref(), document.location.href)