From 1a0e5db1596802596ea75006af0bfa80968948bc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Frank=20L=C3=BCtzenkirchen?=
 <frank.luetzenkirchen@uni-due.de>
Date: Mon, 30 May 2022 10:35:55 +0200
Subject: [PATCH 1/3] UBO-161 Added basic functionality to import publications
 from Web of Science

---
 .../java/org/mycore/ubo/wos/WoSResolver.java  |  87 +++++++
 .../config/ubo-common/mycore.properties       |  14 +-
 .../xsl/import/simplify-json-xml.xsl          |   2 +-
 .../main/resources/xsl/import/wos2mods.xsl    | 225 ++++++++++++++++++
 4 files changed, 326 insertions(+), 2 deletions(-)
 create mode 100644 ubo-common/src/main/java/org/mycore/ubo/wos/WoSResolver.java
 create mode 100644 ubo-common/src/main/resources/xsl/import/wos2mods.xsl

diff --git a/ubo-common/src/main/java/org/mycore/ubo/wos/WoSResolver.java b/ubo-common/src/main/java/org/mycore/ubo/wos/WoSResolver.java
new file mode 100644
index 000000000..98b9c4f21
--- /dev/null
+++ b/ubo-common/src/main/java/org/mycore/ubo/wos/WoSResolver.java
@@ -0,0 +1,87 @@
+package org.mycore.ubo.wos;
+
+import java.io.InputStream;
+
+import javax.ws.rs.client.ClientBuilder;
+import javax.ws.rs.client.WebTarget;
+import javax.ws.rs.core.Response;
+import javax.ws.rs.core.Response.Status.Family;
+import javax.xml.transform.Source;
+import javax.xml.transform.TransformerException;
+import javax.xml.transform.URIResolver;
+
+import org.jdom2.transform.JDOMSource;
+import org.mycore.common.config.MCRConfiguration2;
+import org.mycore.common.content.MCRContent;
+import org.mycore.common.content.MCRStreamContent;
+import org.mycore.ubo.importer.JSON2XMLTransformer;
+
+/**
+ * URI Resolver to get publication data from Web of Science.
+ * 
+ * To retrieve via Web of Science ID, call wos:[ID]
+ * To retrieve via a user query, append query, e.g. wos:DO=10.1039/c4tb01010h
+ * 
+ * UBO.WebOfScience.API.BaseURL
+ * UBO.WebOfScience.API.Key
+ * 
+ * must be set.
+ * 
+ * @author Frank L\u00FCtzenkirchen
+ */
+public class WoSResolver implements URIResolver {
+
+    private static final String ACCEPT_TYPE = "application/json";
+
+    private static final String API_KEY_HEADER = "X-ApiKey";
+
+    private static final String[] PARAMS = { "databaseId", "WOS", "count", "1", "firstRecord", "1",
+        "optionView", "FR" };
+
+    private static final String USR_QUERY = "usrQuery";
+
+    private static final String WOS_ID_PREFIX = "WOS:";
+
+    private String apiKey;
+
+    private WebTarget baseTarget;
+
+    public WoSResolver() {
+        String prefix = "UBO.WebOfScience.API.";
+        String baseURL = MCRConfiguration2.getStringOrThrow(prefix + "BaseURL");
+
+        this.baseTarget = ClientBuilder.newClient().target(baseURL);
+        this.apiKey = MCRConfiguration2.getStringOrThrow(prefix + "Key");
+    }
+
+    @Override
+    public Source resolve(String href, String base) throws TransformerException {
+        String ref = href.substring(href.indexOf(":") + 1);
+
+        WebTarget target = baseTarget;
+        if (!ref.contains("=")) {
+            String id = WOS_ID_PREFIX + ref;
+            target = target.path("id").path(id);
+        } else {
+            target = target.queryParam(USR_QUERY, ref);
+        }
+
+        for (int i = 0; i < PARAMS.length;) {
+            target = target.queryParam(PARAMS[i++], PARAMS[i++]);
+        }
+
+        Response r = target.request().accept(ACCEPT_TYPE).header(API_KEY_HEADER, apiKey).get();
+
+        if (r.getStatusInfo().getFamily() != Family.SUCCESSFUL) {
+            throw new TransformerException(r.getStatus() + " " + r.getStatusInfo().getReasonPhrase());
+        }
+
+        try {
+            MCRContent json = new MCRStreamContent(r.readEntity(InputStream.class));
+            MCRContent xml = new JSON2XMLTransformer().transform(json);
+            return new JDOMSource(xml.asXML());
+        } catch (Exception ex) {
+            throw new TransformerException(ex);
+        }
+    }
+}
diff --git a/ubo-common/src/main/resources/config/ubo-common/mycore.properties b/ubo-common/src/main/resources/config/ubo-common/mycore.properties
index eb2f319f2..42cc9d287 100644
--- a/ubo-common/src/main/resources/config/ubo-common/mycore.properties
+++ b/ubo-common/src/main/resources/config/ubo-common/mycore.properties
@@ -610,6 +610,18 @@ MCR.MODS.EnrichmentResolver.DataSource.Scopus.scopus.URI=xslStyle:import/scopus2
 
 UBO.WebOfScience.Link=http://ws.isiknowledge.com/cps/openurl/service?url_ver=Z39.88-2004&rft_id=info:ut/
 
+UBO.WebOfScience.API.BaseURL=https://wos-api.clarivate.com/api/wos/
+# Set in local mycore.properties:
+# UBO.WebOfScience.API.Key=...
+
+MCR.URIResolver.ModuleResolver.wos=org.mycore.ubo.wos.WoSResolver
+
+# Configuration to import data via EnrichmentResolver
+MCR.MODS.EnrichmentResolver.DataSource.WebOfScience.IdentifierTypes=isi doi pubmed
+MCR.MODS.EnrichmentResolver.DataSource.WebOfScience.isi.URI=xslStyle:import/simplify-json-xml,import/wos2mods,import/genre2genre:wos:{0}
+MCR.MODS.EnrichmentResolver.DataSource.WebOfScience.doi.URI=xslStyle:import/simplify-json-xml,import/wos2mods,import/genre2genre:wos:DO={0}
+MCR.MODS.EnrichmentResolver.DataSource.WebOfScience.pubmed.URI=xslStyle:import/wos2mods,import/genre2genre:wos:PMID={0}
+
 ######################################################################
 #                                                                    #
 #                              CrossRef                              #
@@ -792,7 +804,7 @@ MCR.MODS.EnrichmentResolver.DataSource.GOLD.issn.URI=gold:{1}
 MCR.MODS.EnrichmentResolver.IdentifierType.shelfmark=mods:location/mods:shelfLocator
 
 # Configuration to import new publication by given ID:
-MCR.MODS.EnrichmentResolver.DataSources.import=DuEPublico1 DuEPublico2 EVALuna (Scopus CrossRef PubMed IEEE DataCite arXiv) Unpaywall (LOBID GBV Alma) ZDB JOP
+MCR.MODS.EnrichmentResolver.DataSources.import=DuEPublico1 DuEPublico2 EVALuna (Scopus WebOfScience CrossRef PubMed IEEE DataCite arXiv) Unpaywall (LOBID GBV Alma) ZDB JOP
 
 # Configuration to enrich publications imported from BibTeX/ORCID/CSV (import-list.xed):
 MCR.MODS.EnrichmentResolver.DataSources.import-list=(Scopus CrossRef PubMed IEEE DataCite arXiv) Unpaywall (LOBID GBV Alma) ZDB JOP
diff --git a/ubo-common/src/main/resources/xsl/import/simplify-json-xml.xsl b/ubo-common/src/main/resources/xsl/import/simplify-json-xml.xsl
index 83a08dfb2..49420f9b6 100644
--- a/ubo-common/src/main/resources/xsl/import/simplify-json-xml.xsl
+++ b/ubo-common/src/main/resources/xsl/import/simplify-json-xml.xsl
@@ -9,7 +9,7 @@
   
   <xsl:template match="xpf:*[@key]">
     <xsl:if test="(count(*) &gt; 0) or (string-length(text()) &gt; 0)">
-      <xsl:element name="{@key}">
+      <xsl:element name="{translate(@key,':','_')}">
         <xsl:apply-templates select="*|text()" />
       </xsl:element>
     </xsl:if>
diff --git a/ubo-common/src/main/resources/xsl/import/wos2mods.xsl b/ubo-common/src/main/resources/xsl/import/wos2mods.xsl
new file mode 100644
index 000000000..3eaeb1e49
--- /dev/null
+++ b/ubo-common/src/main/resources/xsl/import/wos2mods.xsl
@@ -0,0 +1,225 @@
+<?xml version="1.0" encoding="UTF-8"?>
+
+<xsl:stylesheet version="1.0" 
+  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+  xmlns:mods="http://www.loc.gov/mods/v3">
+
+  <xsl:output method="xml" encoding="UTF-8" indent="yes" />
+
+  <xsl:template match="/">
+    <xsl:apply-templates select="/entry/Data/Records/records/REC/entry[1]" />
+  </xsl:template>
+
+  <xsl:template match="REC/entry">
+    <mods:mods>
+      <xsl:apply-templates select="static_data/summary/doctypes/doctype" />
+      <xsl:apply-templates select="static_data/summary/titles/title/entry[type='item']" />
+      <xsl:apply-templates select="static_data/summary/names/name/entry" />
+      <xsl:apply-templates select="static_data/summary/pub_info/pubyear" />
+      <xsl:apply-templates select="static_data/fullrecord_metadata/normalized_languages/language" />
+      <xsl:apply-templates select="dynamic_data/cluster_related/identifiers/identifier/entry[type='doi']" />
+      <xsl:apply-templates select="dynamic_data/cluster_related/identifiers/identifier/entry[type='pmid']" />
+      <xsl:apply-templates select="UID" />
+      <mods:relatedItem type="host">
+        <xsl:apply-templates select="static_data/summary/pub_info/pubtype" />
+        <xsl:apply-templates select="static_data/summary/titles/title/entry[type='source']" />
+        <xsl:apply-templates select="static_data/summary/publishers/publisher" />
+        <xsl:apply-templates
+          select="dynamic_data/cluster_related/identifiers/identifier/entry[(type='issn') or (type='eissn')]" />
+        <xsl:apply-templates select="static_data/summary/pub_info" mode="part" />
+      </mods:relatedItem>
+    </mods:mods>
+  </xsl:template>
+
+  <xsl:variable name="lower">abcdefghijklmnopqrstuvwxyz</xsl:variable>
+  <xsl:variable name="upper">ABCDEFGHIJKLMNOPQRSTUVWXYZ</xsl:variable>
+
+  <xsl:template match="language">
+    <!-- Find language with matching label in any language, or with matching ID in any supported code schema -->
+    <xsl:variable name="given"
+      select="translate(content,$upper,$lower)" />
+    <xsl:for-each select="document('classification:metadata:-1:children:rfc4646')/mycoreclass/categories">
+      <xsl:for-each select="category[@ID=$given or label[translate(@text,$upper,$lower)=$given]][1]">
+        <mods:language>
+          <mods:languageTerm authority="rfc4646" type="code">
+            <xsl:value-of select="@ID" />
+          </mods:languageTerm>
+        </mods:language>
+      </xsl:for-each>
+    </xsl:for-each>
+  </xsl:template>
+
+  <xsl:template match="doctype|pubtype">
+    <mods:genre>
+      <xsl:value-of select="." />
+    </mods:genre>
+  </xsl:template>
+
+  <xsl:template match="title/entry">
+    <mods:titleInfo>
+      <mods:title>
+        <xsl:value-of select="content" />
+      </mods:title>
+    </mods:titleInfo>
+  </xsl:template>
+
+  <xsl:template match="name/entry">
+    <mods:name type="personal">
+      <xsl:apply-templates select="last_name" />
+      <xsl:apply-templates select="first_name" />
+      <xsl:apply-templates select="role" />
+      <xsl:apply-templates select="daisng_id" mode="affiliation" />
+      <xsl:apply-templates select="orcid_id" />
+      <xsl:apply-templates select="r_id" />
+    </mods:name>
+  </xsl:template>
+
+  <xsl:template match="last_name">
+    <mods:namePart type="family">
+      <xsl:value-of select="." />
+    </mods:namePart>
+  </xsl:template>
+
+  <xsl:template match="first_name">
+    <mods:namePart type="given">
+      <xsl:value-of select="." />
+    </mods:namePart>
+  </xsl:template>
+
+  <xsl:template match="role">
+    <mods:role>
+      <mods:roleTerm authority="marcrelator" type="code">aut</mods:roleTerm>
+    </mods:role>
+  </xsl:template>
+
+  <xsl:template match="daisng_id" mode="affiliation">
+    <xsl:variable name="id" select="text()" />
+    <xsl:for-each
+      select="ancestor::static_data/fullrecord_metadata/addresses/address_name/entry[names/name[daisng_id=$id]]/address_spec">
+      <mods:affiliation>
+        <xsl:for-each select="organizations/organization/entry[pref='Y']">
+          <xsl:value-of select="content" />
+        </xsl:for-each>
+        <xsl:for-each select="suborganizations/suborganization">
+          <xsl:text>, </xsl:text>
+          <xsl:value-of select="text()" />
+        </xsl:for-each>
+      </mods:affiliation>
+    </xsl:for-each>
+  </xsl:template>
+
+  <xsl:template match="orcid_id">
+    <mods:nameIdentifier type="orcid">
+      <xsl:value-of select="." />
+    </mods:nameIdentifier>
+  </xsl:template>
+
+  <xsl:template match="r_id">
+    <mods:nameIdentifier type="researcherid">
+      <xsl:value-of select="." />
+    </mods:nameIdentifier>
+  </xsl:template>
+
+  <xsl:template match="pubyear">
+    <mods:originInfo>
+      <mods:dateIssued encoding="w3cdtf">
+        <xsl:value-of select="." />
+      </mods:dateIssued>
+    </mods:originInfo>
+  </xsl:template>
+
+  <xsl:template match="publisher">
+    <mods:originInfo>
+      <xsl:apply-templates select="address_spec" />
+      <xsl:apply-templates select="names/name[1]" />
+    </mods:originInfo>
+  </xsl:template>
+
+  <xsl:template match="publisher/names/name">
+    <mods:publisher>
+      <xsl:value-of select="unified_name" />
+    </mods:publisher>
+  </xsl:template>
+
+  <xsl:template match="publisher/address_spec">
+    <mods:place>
+      <mods:placeTerm type="text">
+        <xsl:value-of select="city" />
+      </mods:placeTerm>
+    </mods:place>
+  </xsl:template>
+
+  <xsl:template match="pub_info" mode="part">
+    <mods:part>
+      <xsl:apply-templates select="vol" />
+      <xsl:apply-templates select="issue" />
+      <xsl:apply-templates select="page" />
+    </mods:part>
+  </xsl:template>
+
+  <xsl:template match="vol">
+    <mods:detail type="volume">
+      <mods:number>
+        <xsl:value-of select="." />
+      </mods:number>
+    </mods:detail>
+  </xsl:template>
+
+  <xsl:template match="issue">
+    <mods:detail type="issue">
+      <mods:number>
+        <xsl:value-of select="." />
+      </mods:number>
+    </mods:detail>
+  </xsl:template>
+
+  <xsl:template match="page">
+    <mods:extent unit="pages">
+      <xsl:apply-templates select="begin" />
+      <xsl:apply-templates select="end" />
+    </mods:extent>
+  </xsl:template>
+
+  <xsl:template match="page/begin">
+    <mods:start>
+      <xsl:value-of select="." />
+    </mods:start>
+  </xsl:template>
+
+  <xsl:template match="page/end">
+    <mods:end>
+      <xsl:value-of select="." />
+    </mods:end>
+  </xsl:template>
+
+  <xsl:variable name="wosPrefix">
+    WOS:
+  </xsl:variable>
+
+  <xsl:template match="UID[starts-with(text(),$wosPrefix)]">
+    <mods:identifier type="isi">
+      <xsl:value-of select="substring-after(text(),$wosPrefix)" />
+    </mods:identifier>
+  </xsl:template>
+
+  <xsl:template match="identifier/entry[type='doi']">
+    <mods:identifier type="doi">
+      <xsl:value-of select="value" />
+    </mods:identifier>
+  </xsl:template>
+
+  <xsl:template match="identifier/entry[type='pmid']">
+    <mods:identifier type="pubmed">
+      <xsl:value-of select="substring-after(text(),':')" />
+    </mods:identifier>
+  </xsl:template>
+
+  <xsl:template match="identifier/entry[(type='issn') or (type='eissn')]">
+    <mods:identifier type="issn">
+      <xsl:value-of select="value" />
+    </mods:identifier>
+  </xsl:template>
+
+  <xsl:template match="*" />
+
+</xsl:stylesheet>

From 930226a9541a1c38001836233c788d8461ffe04c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Frank=20L=C3=BCtzenkirchen?=
 <frank.luetzenkirchen@uni-due.de>
Date: Mon, 30 May 2022 13:34:26 +0200
Subject: [PATCH 2/3] UBO-163 Add a default property for the WoS API key

---
 .../src/main/resources/config/ubo-common/mycore.properties      | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/ubo-common/src/main/resources/config/ubo-common/mycore.properties b/ubo-common/src/main/resources/config/ubo-common/mycore.properties
index 42cc9d287..6f523aa71 100644
--- a/ubo-common/src/main/resources/config/ubo-common/mycore.properties
+++ b/ubo-common/src/main/resources/config/ubo-common/mycore.properties
@@ -612,7 +612,7 @@ UBO.WebOfScience.Link=http://ws.isiknowledge.com/cps/openurl/service?url_ver=Z39
 
 UBO.WebOfScience.API.BaseURL=https://wos-api.clarivate.com/api/wos/
 # Set in local mycore.properties:
-# UBO.WebOfScience.API.Key=...
+UBO.WebOfScience.API.Key=this_is_a_dummy-set-in-local-mycore.properties
 
 MCR.URIResolver.ModuleResolver.wos=org.mycore.ubo.wos.WoSResolver
 

From 0ed0db850d9a1fc85a2e7660d3de45aa1ab6f684 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Frank=20L=C3=BCtzenkirchen?=
 <frank.luetzenkirchen@uni-due.de>
Date: Mon, 30 May 2022 13:39:01 +0200
Subject: [PATCH 3/3] Added web of science as data source

---
 .../src/main/resources/META-INF/resources/newPublication.xed    | 1 +
 .../src/main/resources/config/ubo-common/messages_de.properties | 2 +-
 .../src/main/resources/config/ubo-common/messages_en.properties | 2 +-
 3 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/ubo-common/src/main/resources/META-INF/resources/newPublication.xed b/ubo-common/src/main/resources/META-INF/resources/newPublication.xed
index fcd09ba5d..6fecba122 100644
--- a/ubo-common/src/main/resources/META-INF/resources/newPublication.xed
+++ b/ubo-common/src/main/resources/META-INF/resources/newPublication.xed
@@ -37,6 +37,7 @@
                       <option value="doi">DOI</option>
                       <option value="pubmed">PubMed ID</option>
                       <option value="scopus">Scopus ID</option>
+                      <option value="isi">Web of Science ID</option>
                       <option value="ieee">IEEE Article No.</option>
                       <option value="arxiv">arXiv.org ID</option>
                       <option value="isbn">ISBN</option>
diff --git a/ubo-common/src/main/resources/config/ubo-common/messages_de.properties b/ubo-common/src/main/resources/config/ubo-common/messages_de.properties
index dd074a213..f9346c6cc 100644
--- a/ubo-common/src/main/resources/config/ubo-common/messages_de.properties
+++ b/ubo-common/src/main/resources/config/ubo-common/messages_de.properties
@@ -985,7 +985,7 @@ ubo.newPublicationWizard.duplicates.noContinue  = Nein, das ist eine andere Publ
 ubo.newPublicationWizard.duplicates.pleaseCheck = Bitte pr\u00FCfen Sie, ob die Publikation, die Sie melden m\u00F6chten, nicht bereits in der unten aufgef\u00FChrten Liste verzeichnet ist.
 ubo.newPublicationWizard.notFound               = Keine Publikationsdaten gefunden
 ubo.newPublicationWizard.notFound.continue      = Alternativ Titel und Autor*in eingeben
-ubo.newPublicationWizard.notFound.info          = Leider konnten wir f\u00FCr die eingegebene ID keine Publikationsdaten finden. Wir k\u00F6nnen derzeit DOIs aus Scopus, PubMed, IEEE, CrossRef und DataCite aufl\u00F6sen. Bitte pr\u00FCfen Sie die ID oder wenden sich ggf. an die Universit\u00E4tsbibliothek:
+ubo.newPublicationWizard.notFound.info          = Leider konnten wir f\u00FCr die eingegebene ID keine Publikationsdaten finden. Wir k\u00F6nnen derzeit DOIs aus Scopus, Web of Science, PubMed, IEEE, CrossRef und DataCite aufl\u00F6sen. Bitte pr\u00FCfen Sie die ID oder wenden sich ggf. an die Universit\u00E4tsbibliothek:
 ubo.newPublicationWizard.publishedIn            = ver\u00F6ffentlicht in
 ubo.newPublicationWizard.required               = Bitte geben Sie einen DOI oder anderen Identifikator oder alternativ Titel UND Autor*innenname ein!
 ubo.newPublicationWizard.selectGenre            = Bitte w\u00E4hlen Sie den Publikationstyp bzw. korrigieren die Auswahl!
diff --git a/ubo-common/src/main/resources/config/ubo-common/messages_en.properties b/ubo-common/src/main/resources/config/ubo-common/messages_en.properties
index b2f0c77fa..b78db6380 100644
--- a/ubo-common/src/main/resources/config/ubo-common/messages_en.properties
+++ b/ubo-common/src/main/resources/config/ubo-common/messages_en.properties
@@ -973,7 +973,7 @@ ubo.newPublicationWizard.duplicates.noContinue  = No, it is another publication,
 ubo.newPublicationWizard.duplicates.pleaseCheck = Please check if the publication you want to register isn't already listed below.
 ubo.newPublicationWizard.notFound               = No publication data found
 ubo.newPublicationWizard.notFound.continue      = Alternatively, continue by entering title and author
-ubo.newPublicationWizard.notFound.info          = Unfortunately we could'nt find publication data for the identifier entered. We currently resolve DOIs from Scopus, PubMed, IEEE, CrossRef and DataCite. Please check the identifier or contact us:
+ubo.newPublicationWizard.notFound.info          = Unfortunately we could'nt find publication data for the identifier entered. We currently resolve DOIs from Scopus, Web of Science, PubMed, IEEE, CrossRef and DataCite. Please check the identifier or contact us:
 ubo.newPublicationWizard.publishedIn            = published in
 ubo.newPublicationWizard.required               = Please enter the DOI or other ID or alternativly BOTH title and author!
 ubo.newPublicationWizard.selectGenre            = Please choose the publication type resp. fix the selection!