implementation of the graph visualization tool

dnet-dedup-test/src/test/java/eu/dnetlib/pace/DedupTestUtils.java

@@ -1,37 +1,57 @@
 package eu.dnetlib.pace;
 
+import eu.dnetlib.pace.config.Type;
+import eu.dnetlib.pace.model.FieldListImpl;
 import eu.dnetlib.pace.model.MapDocument;
 import eu.dnetlib.support.ConnectedComponent;
+import org.apache.commons.io.IOUtils;
 import org.apache.spark.api.java.JavaRDD;
+import org.codehaus.jackson.map.ObjectMapper;
+import scala.Tuple2;
 
+import java.io.*;
+import java.util.Arrays;
 import java.util.List;
+import java.util.Map;
+import java.util.stream.Collectors;
 
 public abstract class DedupTestUtils {
 
+    public static String prepareTable(MapDocument doc) {
 
-//    public static void printStatistics(JavaRDD<ConnectedComponent> ccs){
-//        final JavaRDD<ConnectedComponent> connectedComponents = ccs.filter(cc -> cc.getDocs().size()>1);
-//        final JavaRDD<ConnectedComponent> nonDeduplicated = ccs.filter(cc -> cc.getDocs().size()==1);
-//
-//        //print deduped
-//        connectedComponents.map(cc -> {
-//            StringBuilder sb = new StringBuilder();
-//            for (MapDocument m : cc.getDocs()){
-//                sb.append(m.getFieldMap().get("originalId").stringValue() + " - "+ m.getFieldMap().get("legalname").stringValue() + "\n");
-//            }
-//            return sb.toString();
-//        }).foreach(s -> System.out.println("*******\n" + s + "*******\n"));
-//
-//        //print nondeduped
-//        nonDeduplicated.foreach(cc -> {
-//            System.out.println(cc.getId() + " - " + cc.getFieldMap().get("legalname").stringValue());
-//        });
-//
-//        System.out.println("Non duplicates: " + nonDeduplicated.count());
-//        System.out.println("Duplicates: " + connectedComponents.flatMap(cc -> cc.getDocs().iterator()).count());
-//        System.out.println("Connected Components: " + connectedComponents.count());
-//
-//    }
+        String ret = "<table>";
+
+        for(String fieldName: doc.getFieldMap().keySet()) {
+            if (doc.getFieldMap().get(fieldName).getType().equals(Type.String)) {
+                ret += "<tr><th>" + fieldName + "</th><td>" + doc.getFieldMap().get(fieldName).stringValue() + "</td></tr>";
+            }
+            else if (doc.getFieldMap().get(fieldName).getType().equals(Type.List)) {
+                ret += "<tr><th>" + fieldName + "</th><td>[" + ((FieldListImpl)doc.getFieldMap().get(fieldName)).stringList().stream().collect(Collectors.joining(";")) + "]</td></tr>";
+            }
+        }
+
+        return ret + "</table>";
+
+    }
+
+    public static void prepareGraphParams(List<String> vertexes, List<Tuple2<String, String>> edgesTuple, String filePath, String templateFilePath, Map<String, MapDocument> mapDocuments) {
+
+        List<Node> nodes = vertexes.stream().map(v -> new Node(v.substring(3, 20).replaceAll("_", ""), vertexes.indexOf(v), prepareTable(mapDocuments.get(v)))).collect(Collectors.toList());
+        List<Edge> edges = edgesTuple.stream().map(e -> new Edge(vertexes.indexOf(e._1()), vertexes.indexOf(e._2()))).collect(Collectors.toList());
+
+        try(FileWriter fw = new FileWriter(filePath)) {
+            String fullText = IOUtils.toString(new FileReader(templateFilePath));
+
+            String s = fullText
+                    .replaceAll("%nodes%", new ObjectMapper().writeValueAsString(nodes))
+                    .replaceAll("%edges%", new ObjectMapper().writeValueAsString(edges));
+
+            IOUtils.write(s, fw);
+        } catch (IOException e) {
+            e.printStackTrace();
+        }
+
+    }
 
     public static String getOrganizationLegalname(MapDocument mapDocument){
         return mapDocument.getFieldMap().get("legalname").stringValue();
@@ -47,3 +67,65 @@ public abstract class DedupTestUtils {
     }
 
 }
+
+class Node{
+    String label;
+    int id;
+    String title;
+
+    public Node(String label, int id, String title) {
+        this.label = label;
+        this.id = id;
+        this.title = title;
+    }
+
+    public String getLabel() {
+        return label;
+    }
+
+    public void setLabel(String label) {
+        this.label = label;
+    }
+
+    public int getId() {
+        return id;
+    }
+
+    public void setId(int id) {
+        this.id = id;
+    }
+
+    public String getTitle() {
+        return title;
+    }
+
+    public void setTitle(String title) {
+        this.title = title;
+    }
+}
+
+class Edge{
+    int from;
+    int to;
+
+    public Edge(int from, int to) {
+        this.from = from;
+        this.to = to;
+    }
+
+    public int getFrom() {
+        return from;
+    }
+
+    public void setFrom(int from) {
+        this.from = from;
+    }
+
+    public int getTo() {
+        return to;
+    }
+
+    public void setTo(int to) {
+        this.to = to;
+    }
+}

dnet-dedup-test/src/test/java/eu/dnetlib/pace/GraphDraw.java

@@ -1,111 +0,0 @@
-package eu.dnetlib.pace;
-
-import java.util.*;
-import java.awt.*;
-import java.awt.event.*;
-import javax.swing.*;
-
-public class GraphDraw extends JFrame {
-    int width;
-    int height;
-
-    ArrayList<Node> nodes;
-    ArrayList<edge> edges;
-
-    public GraphDraw() { //Constructor
-        this.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
-        nodes = new ArrayList<Node>();
-        edges = new ArrayList<edge>();
-        width = 15;
-        height = 15;
-    }
-
-    public GraphDraw(String name) { //Construct with label
-        this.setTitle(name);
-        this.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
-        nodes = new ArrayList<Node>();
-        edges = new ArrayList<edge>();
-        width = 15;
-        height = 15;
-    }
-
-    class Node {
-        int x, y;
-        String name;
-        Color color;
-
-        public Node(String myName, int myX, int myY) {
-            x = myX;
-            y = myY;
-            name = myName;
-            color = Color.white;
-        }
-
-        public Node(String myName, int myX, int myY, Color myColor) {
-            x = myX;
-            y = myY;
-            name = myName;
-            color = myColor;
-        }
-    }
-
-    class edge {
-        int i,j;
-
-        public edge(int ii, int jj) {
-            i = ii;
-            j = jj;
-        }
-    }
-
-    public void addNode(String name, int x, int y) {
-        //add a node at pixel (x,y)
-        nodes.add(new Node(name,x,y));
-        this.repaint();
-    }
-    public void addEdge(int i, int j) {
-        //add an edge between nodes i and j
-        edges.add(new edge(i,j));
-        this.repaint();
-    }
-
-    public void paint(Graphics g) { // draw the nodes and edges
-        FontMetrics f = g.getFontMetrics();
-        int nodeHeight = Math.max(height, f.getHeight());
-        g.setColor(Color.black);
-        for (edge e : edges) {
-            g.drawLine(nodes.get(e.i).x, nodes.get(e.i).y,
-                    nodes.get(e.j).x, nodes.get(e.j).y);
-        }
-
-        for (Node n : nodes) {
-            n.name = n.name.substring(0, 20);
-            int nodeWidth = Math.max(width, f.stringWidth(n.name)+width/2);
-            g.setColor(n.color);
-            g.fillRoundRect(n.x-nodeWidth/2, n.y-nodeHeight/2,
-                    nodeWidth, nodeHeight, 2, 2);
-            g.setColor(Color.black);
-            g.drawRoundRect(n.x-nodeWidth/2, n.y-nodeHeight/2, nodeWidth, nodeHeight, 2, 2);
-
-            g.drawString(n.name, n.x-f.stringWidth(n.name)/2,
-                    n.y+f.getHeight()/5);
-        }
-    }
-}
-
-class testGraphDraw {
-    //Here is some example syntax for the GraphDraw class
-    public static void main(String[] args) {
-        GraphDraw frame = new GraphDraw("Test Window");
-
-        frame.setSize(400,300);
-
-        frame.setVisible(true);
-
-        frame.addNode("a", 50,50);
-        frame.addNode("b", 100,100);
-        frame.addNode("longNode", 200,200);
-        frame.addEdge(0,1);
-        frame.addEdge(0,2);
-    }
-}

dnet-dedup-test/src/test/java/eu/dnetlib/pace/clustering/ClusteringCombinerTest.java

@@ -1,39 +0,0 @@
-package eu.dnetlib.pace.clustering;
-
-import eu.dnetlib.pace.config.Config;
-import eu.dnetlib.pace.model.MapDocument;
-import org.apache.commons.logging.Log;
-import org.apache.commons.logging.LogFactory;
-
-import org.junit.jupiter.api.*;
-
-public class ClusteringCombinerTest  {
-
-
-	// TODO RE IMPLEMENT Tests with the new configuration
-//	private static final Log log = LogFactory.getLog(ClusteringCombinerTest.class);
-//
-//	private Config config;
-//
-//	@Before
-//	public void setUp() {
-//		config = getOrganizationTestConf();
-//	}
-//
-//	@Test
-//	public void testCombine() {
-//
-//		final MapDocument organization = organization(config, "A", "University of Turin", "UNITO");
-//		log.info("University of Turin");
-//		log.info(ClusteringCombiner.combine(organization, config));
-//	}
-//
-//	@Test
-//	public void testCombineBlacklistAware() {
-//
-//		final MapDocument organization = organization(config, "A", "University of Turin", "UNITO");
-//		log.info("University of Turin");
-//		log.info(BlacklistAwareClusteringCombiner.filterAndCombine(organization, config));
-//	}
-
-}

dnet-dedup-test/src/test/resources/eu/dnetlib/pace/config/pub.prod.tree.conf.json

@@ -28,9 +28,26 @@
     "idPath": "$.id"
   },
   "pace": {
-    "clustering" : [
-      { "name" : "wordsStatsSuffixPrefixChain", "fields" : [ "title" ], "params" : { "mod" : "10" } },
-      { "name" : "lowercase", "fields" : [ "doi", "altdoi" ], "params" : { "collapseOn:pid": "0"} }
+    "clustering": [
+      {
+        "name": "wordsStatsSuffixPrefixChain",
+        "fields": [
+          "title"
+        ],
+        "params": {
+          "mod": "10"
+        }
+      },
+      {
+        "name": "lowercase",
+        "fields": [
+          "doi",
+          "altdoi"
+        ],
+        "params": {
+          "collapseOn:pid": "0"
+        }
+      }
     ],
     "decisionTree": {
       "start": {
@@ -393,7 +410,9 @@
         "^(Measurement of the top quark\\-pair production cross section with ATLAS in pp collisions at).*$",
         "^(Measurement of the spin\\-dependent structure function).*",
         "(?i)^.*authors['’′]? reply\\.?$",
-        "(?i)^.*authors['’′]? response\\.?$"
+        "(?i)^.*authors['’′]? response\\.?$",
+        "^Data [mM]anagement [sS]ervices\\.$",
+        "Research and Advanced Technology for Digital Libraries"
       ]
     },
     "synonyms": {}

dnet-dedup-test/src/test/resources/eu/dnetlib/pace/config/test.pub.tree.conf.json

@@ -0,0 +1,403 @@
+{
+  "wf": {
+    "threshold": "0.99",
+    "dedupRun": "001",
+    "entityType": "result",
+    "subEntityType": "resulttype",
+    "subEntityValue": "publication",
+    "orderField": "title",
+    "queueMaxSize": "200",
+    "groupMaxSize": "100",
+    "maxChildren": "100",
+    "slidingWindowSize": "50",
+    "rootBuilder": [
+      "result",
+      "resultProject_outcome_isProducedBy",
+      "resultResult_publicationDataset_isRelatedTo",
+      "resultResult_similarity_isAmongTopNSimilarDocuments",
+      "resultResult_similarity_hasAmongTopNSimilarDocuments",
+      "resultOrganization_affiliation_isAffiliatedWith",
+      "resultResult_part_hasPart",
+      "resultResult_part_isPartOf",
+      "resultResult_supplement_isSupplementTo",
+      "resultResult_supplement_isSupplementedBy",
+      "resultResult_version_isVersionOf"
+    ],
+    "includeChildren": "true",
+    "maxIterations": 20,
+    "idPath": "$.id"
+  },
+  "pace": {
+    "clustering" : [
+      { "name" : "wordsStatsSuffixPrefixChain", "fields" : [ "title" ], "params" : { "mod" : "10" } },
+      { "name" : "lowercase", "fields" : [ "doi", "altdoi" ], "params" : { "collapseOn:pid": "0"} }
+    ],
+    "decisionTree": {
+      "start": {
+        "fields": [
+          {
+            "field": "title",
+            "comparator": "titleVersionMatch",
+            "weight": 1.0,
+            "countIfUndefined": "false",
+            "params": {}
+          },
+          {
+            "field": "authors",
+            "comparator": "sizeMatch",
+            "weight": 1.0,
+            "countIfUndefined": "true",
+            "params": {}
+          }
+        ],
+        "threshold": 1.0,
+        "aggregation": "AND",
+        "positive": "pid_check",
+        "negative": "NO_MATCH",
+        "undefined": "NO_MATCH",
+        "ignoreUndefined": "false"
+      },
+      "pid_check": {
+        "fields": [
+          {
+            "field": "pid",
+            "comparator": "jsonListMatch",
+            "weight": 1.0,
+            "countIfUndefined": "false",
+            "params": {
+              "jpath_value": "$.value",
+              "jpath_classid": "$.qualifier.classid"
+            }
+          },
+          {
+            "field": "pid",
+            "comparator": "jsonListMatch",
+            "weight": 1.0,
+            "countIfUndefined": "false",
+            "params": {
+              "jpath_value": "$.value",
+              "jpath_classid": "$.qualifier.classid",
+              "crossCompare": "alternateid"
+            }
+          }
+        ],
+        "threshold": 0.5,
+        "aggregation": "MAX",
+        "positive": "title_check_low_th",
+        "negative": "title_check_high_th",
+        "undefined": "title_check_high_th",
+        "ignoreUndefined": "true"
+      },
+      "title_check_low_th": {
+        "fields": [
+          {
+            "field": "title",
+            "comparator": "levensteinTitle",
+            "weight": 1.0,
+            "countIfUndefined": "true",
+            "params": {}
+          }
+        ],
+        "threshold": 0.9,
+        "aggregation": "AVG",
+        "positive": "MATCH",
+        "negative": "NO_MATCH",
+        "undefined": "NO_MATCH",
+        "ignoreUndefined": "true"
+      },
+      "title_check_high_th": {
+        "fields": [
+          {
+            "field": "title",
+            "comparator": "levensteinTitle",
+            "weight": 1.0,
+            "countIfUndefined": "true",
+            "params": {}
+          }
+        ],
+        "threshold": 0.99,
+        "aggregation": "AVG",
+        "positive": "MATCH",
+        "negative": "NO_MATCH",
+        "undefined": "NO_MATCH",
+        "ignoreUndefined": "true"
+      }
+    },
+    "model": [
+      {
+        "name": "doi",
+        "type": "String",
+        "path": "$.instance[*].pid[?(@.qualifier.classid == 'doi')].value"
+      },
+      {
+        "name": "altdoi",
+        "type": "String",
+        "path": "$.instance[*].alternateIdentifier[?(@.qualifier.classid == 'doi')].value"
+      },
+      {
+        "name": "pid",
+        "type": "JSON",
+        "path": "$.instance[*].pid[*]",
+        "overrideMatch": "true"
+      },
+      {
+        "name": "alternateid",
+        "type": "JSON",
+        "path": "$.instance[*].alternateIdentifier[*]",
+        "overrideMatch": "true"
+      },
+      {
+        "name": "title",
+        "type": "String",
+        "path": "$.title[?(@.qualifier.classid == 'main title')].value",
+        "length": 250,
+        "size": 5
+      },
+      {
+        "name": "authors",
+        "type": "List",
+        "path": "$.author[*].fullname",
+        "size": 200
+      },
+      {
+        "name": "resulttype",
+        "type": "String",
+        "path": "$.resulttype.classid"
+      }
+    ],
+    "blacklists": {
+      "title": [
+        "(?i)^Data Management Plan",
+        "^Inside Front Cover$",
+        "(?i)^Poster presentations$",
+        "^THE ASSOCIATION AND THE GENERAL MEDICAL COUNCIL$",
+        "^Problems with perinatal pathology\\.?$",
+        "(?i)^Cases? of Puerperal Convulsions$",
+        "(?i)^Operative Gyna?ecology$",
+        "(?i)^Mind the gap\\!?\\:?$",
+        "^Chronic fatigue syndrome\\.?$",
+        "^Cartas? ao editor Letters? to the Editor$",
+        "^Note from the Editor$",
+        "^Anesthesia Abstract$",
+        "^Annual report$",
+        "(?i)^“?THE RADICAL PREVENTION OF VENEREAL DISEASE\\.?”?$",
+        "(?i)^Graph and Table of Infectious Diseases?$",
+        "^Presentation$",
+        "(?i)^Reviews and Information on Publications$",
+        "(?i)^PUBLIC HEALTH SERVICES?$",
+        "(?i)^COMBINED TEXT-?BOOK OF OBSTETRICS AND GYN(Æ|ae)COLOGY$",
+        "(?i)^Adrese autora$",
+        "(?i)^Systematic Part .*\\. Catalogus Fossilium Austriae, Band 2: Echinoidea neogenica$",
+        "(?i)^Acknowledgement to Referees$",
+        "(?i)^Behçet's disease\\.?$",
+        "(?i)^Isolation and identification of restriction endonuclease.*$",
+        "(?i)^CEREBROVASCULAR DISEASES?.?$",
+        "(?i)^Screening for abdominal aortic aneurysms?\\.?$",
+        "^Event management$",
+        "(?i)^Breakfast and Crohn's disease.*\\.?$",
+        "^Cálculo de concentraciones en disoluciones acuosas. Ejercicio interactivo\\..*\\.$",
+        "(?i)^Genetic and functional analyses of SHANK2 mutations suggest a multiple hit model of Autism spectrum disorders?\\.?$",
+        "^Gushi hakubutsugaku$",
+        "^Starobosanski nadpisi u Bosni i Hercegovini \\(.*\\)$",
+        "^Intestinal spirocha?etosis$",
+        "^Treatment of Rodent Ulcer$",
+        "(?i)^\\W*Cloud Computing\\W*$",
+        "^Compendio mathematico : en que se contienen todas las materias mas principales de las Ciencias que tratan de la cantidad$",
+        "^Free Communications, Poster Presentations: Session [A-F]$",
+        "^“The Historical Aspects? of Quackery\\.?”$",
+        "^A designated centre for people with disabilities operated by St John of God Community Services (Limited|Ltd), Louth$",
+        "^P(er|re)-Mile Premiums for Auto Insurance\\.?$",
+        "(?i)^Case Report$",
+        "^Boletín Informativo$",
+        "(?i)^Glioblastoma Multiforme$",
+        "(?i)^Nuevos táxones animales descritos en la península Ibérica y Macaronesia desde 1994 \\(.*\\)$",
+        "^Zaměstnanecké výhody$",
+        "(?i)^The Economics of Terrorism and Counter-Terrorism: A Survey \\(Part .*\\)$",
+        "(?i)^Carotid body tumours?\\.?$",
+        "(?i)^\\[Españoles en Francia : La condición Emigrante.*\\]$",
+        "^Avant-propos$",
+        "(?i)^St\\. Patrick's Cathedral, Dublin, County Dublin - Head(s)? and Capital(s)?$",
+        "(?i)^St\\. Patrick's Cathedral, Dublin, County Dublin - Bases?$",
+        "(?i)^PUBLIC HEALTH VERSUS THE STATE$",
+        "^Viñetas de Cortázar$",
+        "(?i)^Search for heavy neutrinos and W(\\[|_|\\(|_\\{|-)?R(\\]|\\)|\\})? bosons with right-handed couplings in a left-right symmetric model in pp collisions at.*TeV(\\.)?$",
+        "(?i)^Measurement of the pseudorapidity and centrality dependence of the transverse energy density in Pb(-?)Pb collisions at.*tev(\\.?)$",
+        "(?i)^Search for resonances decaying into top-quark pairs using fully hadronic decays in pp collisions with ATLAS at.*TeV$",
+        "(?i)^Search for neutral minimal supersymmetric standard model Higgs bosons decaying to tau pairs in pp collisions at.*tev$",
+        "(?i)^Relatório de Estágio (de|em) Angiologia e Cirurgia Vascular$",
+        "^Aus der AGMB$",
+        "^Znanstveno-stručni prilozi$",
+        "(?i)^Zhodnocení finanční situace podniku a návrhy na zlepšení$",
+        "(?i)^Evaluation of the Financial Situation in the Firm and Proposals to its Improvement$",
+        "(?i)^Hodnocení finanční situace podniku a návrhy na její zlepšení$",
+        "^Finanční analýza podniku$",
+        "^Financial analysis( of business)?$",
+        "(?i)^Textbook of Gyn(a)?(Æ)?(e)?cology$",
+        "^Jikken nihon shūshinsho$",
+        "(?i)^CORONER('|s)(s|') INQUESTS$",
+        "(?i)^(Μελέτη παραγόντων )?risk management( για ανάπτυξη και εφαρμογή ενός πληροφοριακού συστήματος| και ανάπτυξη συστήματος)?$",
+        "(?i)^Consultants' contract(s)?$",
+        "(?i)^Upute autorima$",
+        "(?i)^Bijdrage tot de Kennis van den Godsdienst der Dajaks van Lan(d|f)ak en Tajan$",
+        "^Joshi shin kokubun$",
+        "^Kōtō shōgaku dokuhon nōson'yō$",
+        "^Jinjō shōgaku shōka$",
+        "^Shōgaku shūjichō$",
+        "^Nihon joshi dokuhon$",
+        "^Joshi shin dokuhon$",
+        "^Chūtō kanbun dokuhon$",
+        "^Wabun dokuhon$",
+        "(?i)^(Analysis of economy selected village or town|Rozbor hospodaření vybrané obce či města)$",
+        "(?i)^cardiac rehabilitation$",
+        "(?i)^Analytical summary$",
+        "^Thesaurus resolutionum Sacrae Congregationis Concilii$",
+        "(?i)^Sumario analítico(\\s{1})?(Analitic summary)?$",
+        "^Prikazi i osvrti$",
+        "^Rodinný dům s provozovnou$",
+        "^Family house with an establishment$",
+        "^Shinsei chūtō shin kokugun$",
+        "^Pulmonary alveolar proteinosis(\\.?)$",
+        "^Shinshū kanbun$",
+        "^Viñeta(s?) de Rodríguez$",
+        "(?i)^RUBRIKA UREDNIKA$",
+        "^A Matching Model of the Academic Publication Market$",
+        "^Yōgaku kōyō$",
+        "^Internetový marketing$",
+        "^Internet marketing$",
+        "^Chūtō kokugo dokuhon$",
+        "^Kokugo dokuhon$",
+        "^Antibiotic Cover for Dental Extraction(s?)$",
+        "^Strategie podniku$",
+        "^Strategy of an Enterprise$",
+        "(?i)^respiratory disease(s?)(\\.?)$",
+        "^Award(s?) for Gallantry in Civil Defence$",
+        "^Podniková kultura$",
+        "^Corporate Culture$",
+        "^Severe hyponatraemia in hospital inpatient(s?)(\\.?)$",
+        "^Pracovní motivace$",
+        "^Work Motivation$",
+        "^Kaitei kōtō jogaku dokuhon$",
+        "^Konsolidovaná účetní závěrka$",
+        "^Consolidated Financial Statements$",
+        "(?i)^intracranial tumour(s?)$",
+        "^Climate Change Mitigation Options and Directed Technical Change: A Decentralized Equilibrium Analysis$",
+        "^\\[CERVECERIAS MAHOU(\\.|\\:) INTERIOR\\] \\[Material gráfico\\]$",
+        "^Housing Market Dynamics(\\:|\\.) On the Contribution of Income Shocks and Credit Constraint(s?)$",
+        "^\\[Funciones auxiliares de la música en Radio París,.*\\]$",
+        "^Úroveň motivačního procesu jako způsobu vedení lidí$",
+        "^The level of motivation process as a leadership$",
+        "^Pay-beds in N(\\.?)H(\\.?)S(\\.?) Hospitals$",
+        "(?i)^news and events$",
+        "(?i)^NOVOSTI I DOGAĐAJI$",
+        "^Sansū no gakushū$",
+        "^Posouzení informačního systému firmy a návrh změn$",
+        "^Information System Assessment and Proposal for ICT Modification$",
+        "^Stresové zatížení pracovníků ve vybrané profesi$",
+        "^Stress load in a specific job$",
+        "^Sunday: Poster Sessions, Pt.*$",
+        "^Monday: Poster Sessions, Pt.*$",
+        "^Wednesday: Poster Sessions, Pt.*",
+        "^Tuesday: Poster Sessions, Pt.*$",
+        "^Analýza reklamy$",
+        "^Analysis of advertising$",
+        "^Shōgaku shūshinsho$",
+        "^Shōgaku sansū$",
+        "^Shintei joshi kokubun$",
+        "^Taishō joshi kokubun dokuhon$",
+        "^Joshi kokubun$",
+        "^Účetní uzávěrka a účetní závěrka v ČR$",
+        "(?i)^The \"?Causes\"? of Cancer$",
+        "^Normas para la publicación de artículos$",
+        "^Editor('|s)(s|') [Rr]eply$",
+        "^Editor(’|s)(s|’) letter$",
+        "^Redaktoriaus žodis$",
+        "^DISCUSSION ON THE PRECEDING PAPER$",
+        "^Kōtō shōgaku shūshinsho jidōyō$",
+        "^Shōgaku nihon rekishi$",
+        "^(Theory of the flow of action currents in isolated myelinated nerve fibers).*$",
+        "^Préface$",
+        "^Occupational [Hh]ealth [Ss]ervices.$",
+        "^In Memoriam Professor Toshiyuki TAKESHIMA$",
+        "^Účetní závěrka ve vybraném podniku.*$",
+        "^Financial statements in selected company$",
+        "^Abdominal [Aa]ortic [Aa]neurysms.*$",
+        "^Pseudomyxoma peritonei$",
+        "^Kazalo autora$",
+        "(?i)^uvodna riječ$",
+        "^Motivace jako způsob vedení lidí$",
+        "^Motivation as a leadership$",
+        "^Polyfunkční dům$",
+        "^Multi\\-funkcional building$",
+        "^Podnikatelský plán$",
+        "(?i)^Podnikatelský záměr$",
+        "(?i)^Business Plan$",
+        "^Oceňování nemovitostí$",
+        "^Marketingová komunikace$",
+        "^Marketing communication$",
+        "^Sumario Analítico$",
+        "^Riječ uredništva$",
+        "^Savjetovanja i priredbe$",
+        "^Índice$",
+        "^(Starobosanski nadpisi).*$",
+        "^Vzdělávání pracovníků v organizaci$",
+        "^Staff training in organization$",
+        "^(Life Histories of North American Geometridae).*$",
+        "^Strategická analýza podniku$",
+        "^Strategic Analysis of an Enterprise$",
+        "^Sadržaj$",
+        "^Upute suradnicima$",
+        "^Rodinný dům$",
+        "(?i)^Fami(l)?ly house$",
+        "^Upute autorima$",
+        "^Strategic Analysis$",
+        "^Finanční analýza vybraného podniku$",
+        "^Finanční analýza$",
+        "^Riječ urednika$",
+        "(?i)^Content(s?)$",
+        "(?i)^Inhalt$",
+        "^Jinjō shōgaku shūshinsho jidōyō$",
+        "(?i)^Index$",
+        "^Chūgaku kokubun kyōkasho$",
+        "^Retrato de una mujer$",
+        "^Retrato de un hombre$",
+        "^Kōtō shōgaku dokuhon$",
+        "^Shotōka kokugo$",
+        "^Shōgaku dokuhon$",
+        "^Jinjō shōgaku kokugo dokuhon$",
+        "^Shinsei kokugo dokuhon$",
+        "^Teikoku dokuhon$",
+        "^Instructions to Authors$",
+        "^KİTAP TAHLİLİ$",
+        "^PRZEGLĄD PIŚMIENNICTWA$",
+        "(?i)^Presentación$",
+        "^İçindekiler$",
+        "(?i)^Tabl?e of contents$",
+        "^(CODICE DEL BEATO DE LOS REYES FERNANDO I Y SANCHA).*$",
+        "^(\\[MADRID\\. BIBL\\. NAC\\. N.*KING FERDINAND I.*FROM SAN ISIDORO DE LEON\\. FACUNDUS SCRIPSIT DATED.*\\]).*",
+        "^Editorial( Board)?$",
+        "(?i)^Editorial \\(English\\)$",
+        "^Editörden$",
+        "^(Corpus Oral Dialectal \\(COD\\)\\.).*$",
+        "^(Kiri Karl Morgensternile).*$",
+        "^(\\[Eksliibris Aleksandr).*\\]$",
+        "^(\\[Eksliibris Aleksandr).*$",
+        "^(Eksliibris Aleksandr).*$",
+        "^(Kiri A\\. de Vignolles).*$",
+        "^(2 kirja Karl Morgensternile).*$",
+        "^(Pirita kloostri idaosa arheoloogilised).*$",
+        "^(Kiri tundmatule).*$",
+        "^(Kiri Jenaer Allgemeine Literaturzeitung toimetusele).*$",
+        "^(Eksliibris Nikolai Birukovile).*$",
+        "^(Eksliibris Nikolai Issakovile).*$",
+        "^(WHP Cruise Summary Information of section).*$",
+        "^(Measurement of the top quark\\-pair production cross section with ATLAS in pp collisions at).*$",
+        "^(Measurement of the spin\\-dependent structure function).*",
+        "(?i)^.*authors['’′]? reply\\.?$",
+        "(?i)^.*authors['’′]? response\\.?$",
+        "^Data [mM]anagement [sS]ervices\\.$",
+        "Research and Advanced Technology for Digital Libraries"
+      ]
+    },
+    "synonyms": {}
+  }
+}

dnet-dedup-test/src/test/resources/graph_visualization_tool/graph_template.html

@@ -0,0 +1,70 @@
+<html>
+<head>
+    <script type="text/javascript" src="https://unpkg.com/vis-network/standalone/umd/vis-network.min.js"></script>
+
+    <style type="text/css">
+        #mynetwork {
+            width: 1000px;
+            height: 700px;
+            border: 1px solid lightgray;
+        }
+
+        th, td {
+            font-size: 10px;
+        }
+    </style>
+</head>
+<body>
+<div id="mynetwork"></div>
+
+<script type="text/javascript">
+
+    // HTML parsing with all XSS goodness
+    function htmlTitle(html) {
+        const container = document.createElement("div");
+        container.innerHTML = html;
+        return container;
+    }
+
+    var nodesArray = %nodes%;
+    var edgesArray = %edges%;
+
+    for (var i = 0; i < nodesArray.length; i++) {
+        nodesArray[i].title = htmlTitle(nodesArray[i].title);
+    };
+
+    // create an array with nodes
+    var nodes = new vis.DataSet(nodesArray);
+
+    // create an array with edges
+    var edges = new vis.DataSet(edgesArray);
+
+    // HTML parsing with all XSS goodness
+    function htmlTitle(html) {
+        const container = document.createElement("div");
+        container.innerHTML = html;
+        return container;
+    }
+
+    // create a network
+    var container = document.getElementById('mynetwork');
+
+    // provide the data in the vis format
+    var data = {
+        nodes: nodes,
+        edges: edges
+    };
+
+    var options = {
+        physics:{enabled: false},
+        edges:{physics:false},
+        nodes:{font:{size:10}},
+        layout: {improvedLayout:true}
+    };
+
+    // initialize your network!
+    var network = new vis.Network(container, data, options);
+
+</script>
+</body>
+</html>