Merge origin/master

# Conflicts: # pom.xml
zumult-org · Dec 8, 2024 · c67d3f5 · c67d3f5
2 parents 1c200e3 + 4fed1c8
commit c67d3f5
Show file tree

Hide file tree

Showing 16 changed files with 82 additions and 72 deletions.
diff --git a/pom.xml b/pom.xml
@@ -8,9 +8,16 @@
     <!-- <packaging>war</packaging> -->
     <packaging>jar</packaging>
     <name>ZuMultMVN</name>
-    <url>https://zumult.org/</url>
+
+    <!-- related to #214 -->
+    <repositories>
+        <repository>
+          <id>github-manual</id>
+          <url>https://raw.githubusercontent.com/Exmaralda-Org/exmaralda/master/src/build/webjar/</url>
+        </repository>        
+    </repositories>    <url>https://zumult.org/</url>
         <build>
-        <finalName>${project.artifactId}</finalName>
+        <finalName>${project.artifactId}</finalName>       
         <resources>
             <resource>
               <directory>src/main/resources</directory>
@@ -20,8 +27,12 @@
             </resource>
             <resource>
                  <directory>src/main/java</directory>
+                  <excludes>
+                        <exclude>/data/corpora/**</exclude>
+                        <exclude>/data/indices/**</exclude>
+                 </excludes>
             </resource>   
-        </resources>
+        </resources>      
         <plugins>
             <plugin>
                 <groupId>org.apache.maven.plugins</groupId>
@@ -68,10 +79,11 @@
             <artifactId>ojdbc6</artifactId>
             <version>11.2.0.4</version>
         </dependency> -->
+        <!-- see #214 -->
         <dependency>
             <groupId>org.exmaralda</groupId>
-            <artifactId>EXMARaLDA</artifactId>
-            <version>Preview-20220623</version>
+            <artifactId>EXMARaLDA_WEB</artifactId>
+            <version>1.0.0</version>
             <type>jar</type>
         </dependency>
         <dependency>
@@ -258,4 +270,4 @@
             </plugin>
         </plugins>
     </reporting>
-</project>
+</project>
diff --git a/src/main/java/org/zumult/backend/implementations/COMAFileSystem.java b/src/main/java/org/zumult/backend/implementations/COMAFileSystem.java
@@ -359,8 +359,9 @@ public IDList getTranscripts4SpeechEvent(String speechEventID) throws IOExceptio
             /*NodeList allTranscripts = 
                     (NodeList) xPath.evaluate("//Transcription[substring(Filename, string-length(Filename)-3)='.exb' or substring(Filename, string-length(Filename)-3)='.EXB']", 
                             communication.getDocument().getDocumentElement(), XPathConstants.NODESET);*/
+            String xpath = "//Transcription[substring(Filename, string-length(Filename)-3)='.xml' or substring(Filename, string-length(Filename)-3)='.XML']";
             NodeList allTranscripts = 
-                    (NodeList) xPath.evaluate("//Transcription[substring(Filename, string-length(Filename)-3)='.xml' or substring(Filename, string-length(Filename)-3)='.XML']", 
+                    (NodeList) xPath.evaluate(xpath, 
                             communication.getDocument().getDocumentElement(), XPathConstants.NODESET);
             for (int i=0; i<allTranscripts.getLength(); i++){
                 Element transcriptElement = ((Element)(allTranscripts.item(i)));

diff --git a/src/main/java/org/zumult/backend/implementations/TestCOMABackend.java b/src/main/java/org/zumult/backend/implementations/TestCOMABackend.java
@@ -11,6 +11,7 @@
 import java.util.logging.Level;
 import java.util.logging.Logger;
 import org.zumult.backend.BackendInterface;
+import org.zumult.backend.BackendInterfaceFactory;
 import org.zumult.backend.Configuration;
 import org.zumult.backend.MetadataFinderInterface;
 import org.zumult.io.IOHelper;
@@ -51,7 +52,7 @@ private void doit() {
                     "--- Metadata path : " +
                     Configuration.getMetadataPath()
             );
-            BackendInterface bi = new COMAFileSystem();
+            BackendInterface bi = new COMAFileSystem(); 
             SearchResultPlus searchResult = bi.search("[word=\"juste\"]", null, null, "EXMARaLDA-DemoKorpus", null, 1000, null, null, null, null, null);
             long t1 = System.currentTimeMillis();
             KWIC kwic = bi.getKWIC(searchResult, "3-t,3-t");

diff --git a/src/main/java/org/zumult/io/isotei2html_table.xsl b/src/main/java/org/zumult/io/isotei2html_table.xsl
@@ -173,8 +173,10 @@
         <xsl:variable name="lemma" select="@lemma"/>
         <xsl:variable name="pos" select="@pos"/>
         <xsl:variable name="trans" select="."/>
-        <xsl:variable name="startAnchor" select="preceding-sibling::tei:anchor[1]/@synch"/>        
-        <xsl:variable name="endAnchor" select="following-sibling::tei:anchor[1]/@synch"/>        
+        <!-- <xsl:variable name="startAnchor" select="preceding-sibling::tei:anchor[1]/@synch"/>        
+        <xsl:variable name="endAnchor" select="following-sibling::tei:anchor[1]/@synch"/>      -->   
+        <xsl:variable name="startAnchor" select="preceding::tei:anchor[1]/@synch"/>        
+        <xsl:variable name="endAnchor" select="following::tei:anchor[1]/@synch"/>
         <xsl:variable name="start" select="//tei:when[@xml:id=$startAnchor]/@interval"/>
         <xsl:variable name="end" select="//tei:when[@xml:id=$endAnchor]/@interval"/>
         <span data-start="{$start}" data-end="{$end}" data-norm="{$norm}" data-lemma="{$lemma}" data-pos="{$pos}" data-id="{$ID}">

diff --git a/src/main/java/org/zumult/objects/implementations/COMACommunication.java b/src/main/java/org/zumult/objects/implementations/COMACommunication.java
@@ -113,7 +113,8 @@ public List<String> getTopics() {
     public IDList getTranscripts() {
         IDList result = new IDList("transcript");
         try {
-            String xPathString = "descendant::Transcription";
+            //String xPathString = "descendant::Transcription";
+            String xPathString = "descendant::Transcription[substring(Filename, string-length(Filename)-3)='.xml' or substring(Filename, string-length(Filename)-3)='.XML']";
             NodeList transcriptionNodeList = (NodeList) xPath.evaluate(xPathString, getDocument().getDocumentElement(), XPathConstants.NODESET);
             for (int i=0; i<transcriptionNodeList.getLength(); i++){
                 Element transcriptionElement = (Element)(transcriptionNodeList.item(i));

diff --git a/src/main/java/org/zumult/query/implementations/COMASearcher.java b/src/main/java/org/zumult/query/implementations/COMASearcher.java
@@ -8,6 +8,7 @@
 import java.io.File;
 import java.io.IOException;
 import java.util.ArrayList;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
 import org.zumult.backend.Configuration;
@@ -67,8 +68,22 @@ public IDList getCorporaForSearch(String searchIndex) {
     @Override
     public ArrayList<String> getIndexPaths(SearchIndexType searchIndex) throws IOException, SearchServiceException {
         ArrayList<String> result = new ArrayList<>();
+        // this is for #219
+        String corpQ = metadataQuery.corpusQuery.replaceAll("corpusSigle=", "").replace("\"", "");
+        // Corpus: corpusSigle="INTV"
+        System.out.println("Corpus: " + corpQ);        
+        String[] corps = (corpQ + "|").split("\\|");
+        Set<String> corpora = new HashSet<>();
+        for (String c : corps){
+            corpora.add(c.trim());
+        }
+
         //for (String indexID : Configuration.getTranscriptBasedIndexIDs()){
         for (String indexID : Configuration.getSpeakerBasedIndexIDs()){
+            // this is for #219
+            String thisCorpus = indexID.substring(indexID.indexOf("_")+1);
+            if (!corpora.contains(thisCorpus)) continue;
+
             File file = new File(Configuration.getSearchIndexPath(), indexID);
             result.add(file.getAbsolutePath());
         }

diff --git a/src/main/java/org/zumult/query/implementations/ISOTEIKWICSnippetCreator.java b/src/main/java/org/zumult/query/implementations/ISOTEIKWICSnippetCreator.java
@@ -95,8 +95,13 @@ public DefaultKWICSnippet apply(Document transcriptDoc,
                 //Node firstAnnotationBlock = (Element) xPath.evaluate(xpathString, TranscriptDoc, XPathConstants.NODE);
                 //Node firstAnnotationBlock = (Node) xPath.compile(xpathString).evaluate(transcriptDoc, XPathConstants.NODE);
                 //Node firstAnnotationBlock = (Node) firstElem.getParentNode().getParentNode().getParentNode();
-                Element firstAnnotationBlockElem = (Element) firstElem.getParentNode().getParentNode().getParentNode();
+                // This is wrong if we have nested <seg>s
+                //Element firstAnnotationBlockElem = (Element) firstElem.getParentNode().getParentNode().getParentNode();
+
+                String xpathString = "ancestor::tei:"+ Constants.ELEMENT_NAME_ANNOTATION_BLOCK + "[1]";
+                Element firstAnnotationBlockElem = (Element) xPath.evaluate(xpathString, firstElem, XPathConstants.NODE);
                 String speaker = firstAnnotationBlockElem.getAttribute(Constants.ATTRIBUTE_NAME_WHO);
+                //System.out.println("Speaker is " + speaker);
                 speaker = getSpeakerInitials(speaker, transcriptDoc);
                 // get left context
                 int index = 0;
@@ -192,10 +197,13 @@ public DefaultKWICSnippet apply(Document transcriptDoc,
 
     private String getSpeakerInitials(String speaker, Document transcriptDoc){
 
+        String xpathString = "//tei:person[@xml:id='" + speaker + "']";
+        //System.out.println("Looking for " + speaker);
+        //Node personNode = null;
         try {
 
             // /tei:TEI/tei:teiHeader[1]/tei:profileDesc[1]/tei:particDesc[1]/tei:person[1]
-            String xpathString = "/tei:TEI/tei:teiHeader[1]/tei:profileDesc[1]/tei:particDesc[1]/tei:person[@xml:id='" + speaker + "']";
+            //String xpathString = "/tei:TEI/tei:teiHeader[1]/tei:profileDesc[1]/tei:particDesc[1]/tei:person[@xml:id='" + speaker + "']";
             Element personElement = (Element) xPath.compile(xpathString).evaluate(transcriptDoc, XPathConstants.NODE);
             if (personElement!=null){
                 return personElement.getAttribute("n");

diff --git a/src/main/resources/resources/MessageBundle.properties b/src/main/resources/resources/MessageBundle.properties
@@ -14,7 +14,7 @@ DocumentsPerPage=Documents per Page
 Download=Download
 DownloadOptions=Download Options
 DownloadKWIC=Download KWIC
-EnterYourCQP=Please enter your CQP query and choose between different search modes
+EnterYourCQP=Please enter your CQP query
 Examples=Examples
 SearchInputFieldPlaceholder=Find...
 First=First
@@ -89,7 +89,7 @@ ViewHitsAsLemmaTable=View hits as lemma table
 ViewHitsAsKWICOnTheQueryTab=View hits as KWIC on the Query-Tab
 Vocabulary=Lemma and Query Lists
 WordTokens=word tokens
-ZuRechtSubtitle=CQP queries to the AGD corpora, using a Lucene index via MTAS
+ZuRechtSubtitle=CQP queries corpora, using a Lucene index via MTAS
 ReferenceWordlist=Reference wordlist
 RefWordlist=Ref wordlist
 WaitWordlistLoading=Please wait -- wordlist is loading.

diff --git a/src/main/resources/resources/MessageBundle_de_DE.properties b/src/main/resources/resources/MessageBundle_de_DE.properties
@@ -2,7 +2,7 @@
 Project=Projekt
 Home=Startseite
 Logout=Ausloggen
-ZuRechtSubtitle=CQP-Suchanfragen in den Korpora des AGD
+ZuRechtSubtitle=CQP-Suchanfragen in Korpora
 AvailableCorpora=Korpora
 Cancel=Abbrechen
 InPreparation=in Vorbereitung
@@ -32,7 +32,7 @@ SearchBtn=Suche starten
 OpenPanelWithSearchOptions=Suchoptionen \u00f6ffnen
 IgnorePunctuation=Satzzeichen ignorieren
 NumberOfHitsForDownload=Anzahl der Treffer f\u00fcr Download
-EnterYourCQP=Bitte geben Sie Ihren CQP-Suchausdruck ein und w\u00e4hlen Sie den Suchmodus (Transkript- vs. Sprecher-basiert)
+EnterYourCQP=Bitte geben Sie Ihren CQP-Suchausdruck ein
 OpenHelp=Hilfe anzeigen
 Loading=wird geladen...
 LoadingPage=Bitte warten Sie. Die Seite wird geladen...

diff --git a/web/images/talking.png b/web/images/talking.png
diff --git a/web/images/words.jpg b/web/images/words.jpg
diff --git a/web/index.jsp b/web/index.jsp
@@ -8,6 +8,11 @@
 
 <%@page import="java.util.Locale"%>
 <%@page import="java.util.ResourceBundle"%>
+<%@page import="org.zumult.objects.IDList"%>
+<%@page import="org.zumult.backend.BackendInterfaceFactory"%>
+<%@page import="org.zumult.backend.BackendInterface"%>
+
+
 
 
 <!DOCTYPE html>
@@ -36,7 +41,6 @@
 
         <link rel="stylesheet" href="css/overview.css"/>       
 
-        <%@include file="WEB-INF/jspf/matomoTracking.jspf" %>       
 
         <script>
             $(document).ready(function(){
@@ -55,16 +59,22 @@
 
     <%
        String pageTitle = myResources.getString("testApp"); 
+       BackendInterface bi = BackendInterfaceFactory.newBackendInterface();
+       IDList allTranscriptIDs =  bi.getTranscripts4Corpus(bi.getCorpora().get(0));
+       String randomTranscriptID = allTranscriptIDs.get((int)(Math.random() * (allTranscriptIDs.size())));
     %>
 
-    <body style="background-image: url('./images/doors.jpg');">
+    <body style="background-image: url('./images/talking.png');">
 
         <% String pageName = "ZuMult"; %>
         <%@include file="../WEB-INF/jspf/zumultNav.jspf" %>                                                
         <div class="row">
             <div class="col-sm-2">
             </div>
             <div class="col-sm-8" style="background:white;padding-top: 20px; border-radius: 15px;">
+                <div class="m-3 text-center">
+                    <a href="./jsp/corpusoverview.jsp?lang=<%=currentLocale.getLanguage()%>" class="btn btn-outline-info" target="_blank">Corpora</a>                    
+                </div>
 
                 <div class="card-deck">
                     <div class="card mb-3">
@@ -80,7 +90,7 @@
                       <div class="card-body">
                         <h5 class="card-title">ZuViel</h5>
                         <p class="card-text"><%=myResources.getString("ZuVielShort")%></p>
-                        <a href="./jsp/zuViel.jsp?transcriptID=ID2668D68B-60BC-E915-A7D3-68DE4A6AB954" class="btn btn-primary" target="_blank">ZuViel</a>
+                        <a href="./jsp/zuViel.jsp?transcriptID=<%= randomTranscriptID %>" class="btn btn-primary" target="_blank">ZuViel</a>
                       </div>
                     </div>                      
                 </div>

diff --git a/web/js/media.js b/web/js/media.js
@@ -13,7 +13,7 @@ function getCurrentPosition(){
 }
 
 function getVideoImage(videoID){
-    var WEBSERVICE = "/ZumultDataServlet?command=getVideoImage"
+    var WEBSERVICE = "/ZumultDataServlet?command=getVideoImage";
     var PARAMETERS =    "&videoID=" + videoID
                      +  "&position=" + getCurrentPosition();
     $.get(BASE_URL + WEBSERVICE + PARAMETERS, function( data ) {

diff --git a/web/jsp/corpusoverview.jsp b/web/jsp/corpusoverview.jsp
@@ -44,7 +44,7 @@
     %>
     <head>  
         <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
-        <title>ZuMult: AGD Corpus Overview</title>
+        <title>ZuMult: Corpus Overview</title>
         <link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.4.1/css/bootstrap.min.css" integrity="sha384-Vkoo8x4CGsO3+Hhxv8T/Q5PaXtkKtu6ug5TOeNV6gBiFeWPGFN9MuhOf23Q9Ifjh" crossorigin="anonymous"/>
         <script src="https://code.jquery.com/jquery-3.4.1.min.js" integrity="sha256-CSXorXvZcTkaix6Yvo6HppcZGetbYMGWSFlBw8HfCJo="crossorigin="anonymous"></script>        
         <script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/umd/popper.min.js" integrity="sha384-Q6E9RHvbIyZFJoft+2mJbHaEWldlvI9IOYy5n3zV9zzTtmI3UksdQRVvoxMfooAo" crossorigin="anonymous"></script>
@@ -69,9 +69,9 @@
 
     </head>
     <body>
-        <% String pageTitle = "Korpusübersicht Archiv für Gesprochenes Deutsch"; 
+        <% String pageTitle = "Korpusübersicht"; 
            if ("en".equals(language)) {
-               pageTitle = "Corpus overview Archive for Spoken German";
+               pageTitle = "Corpus overview";
            }
            String pageName = "ZuMult";
         %>
@@ -83,13 +83,12 @@
             <div class="col-sm-8">
                 <p class="text-justify">
                 <% if ("en".equals(language)) { %>
-                    This page gives an overview of the <%= corpora.size() %> spoken language corpora of the 
-                    <a href="http://agd.ids-mannheim.de" target="_blank">Archive for Spoken German</a>
-                    which are available online.
+                    This page gives an overview of the <%= corpora.size() %> corpora available in this 
+                    <a href="http://zumult.org" target="_blank">ZuMult</a> instance.
                 <% } else { %>
-                    Diese Seite gibt einen Überblick über die <%= corpora.size() %> mündlichen Korpora des
-                    <a href="http://agd.ids-mannheim.de" target="_blank">Archiv für Gesprochenes Deutsch</a>,
-                    die online zugänglich sind. 
+                    Diese Seite gibt einen Überblick über die <%= corpora.size() %> Korpora, die in dieser
+                    <a href="http://zumult.org" target="_blank">ZuMult</a>-Instanz
+                    zugänglich sind. 
                 <% } %>
                 </p>
             </div>
@@ -112,7 +111,7 @@
                   <div class="row no-gutters">
                     <div class="col-md-2">
                         <figure class="figure">
-                            <img src="../images/corpora/<%= corpusID %>.png" class="card-img-top rounded" alt="...">
+                            <img src="../images/words.jpg" class="card-img-top rounded" alt="...">
                             <% if (IMAGE_COPYRIGHT_MAP.containsKey(corpusID)) { %>
                                 <figcaption class="figure-caption text-right">&copy; <%= IMAGE_COPYRIGHT_MAP.get(corpusID) %></figcaption>                            
                             <% } %>
@@ -123,14 +122,6 @@
                             <h5 class="card-title"><%=acronym%></h5>
                             <h6 class="card-subtitle mb-2 text-muted"><%=name%></h6>
                             <p class="card-text"><%=description%></p>
-                            <a class="card-link" target="_blank" href="eventstable.jsp?corpusID=<%=corpusID%>">
-                                <%= backendInterface.getEvents4Corpus(corpusID).size() %> 
-                                <% if ("en".equals(language)) { %>
-                                    Events
-                                <% } else { %>
-                                    Ereignisse
-                                <% } %>
-                            </a>
                             <a class="card-link"  target="_blank" href="speecheventstable.jsp?corpusID=<%=corpusID%>">
                                 <%= backendInterface.getSpeechEvents4Corpus(corpusID).size() %>
                                 <% if ("en".equals(language)) { %>

diff --git a/web/jsp/speecheventstable.jsp b/web/jsp/speecheventstable.jsp
@@ -106,9 +106,10 @@
                   //  Long dur = System.currentTimeMillis() - time;
                   //  time = System.currentTimeMillis();
                     SpeechEvent speechEvent = backendInterface.getSpeechEvent(speechEventID);
+                    String firstTranscriptID = speechEvent.getTranscripts().get(0);
             %>
                     <tr>
-                        <th><%=speechEventID%></th>
+                        <th><a href="zuViel.jsp?transcriptID=<%= firstTranscriptID %>" target="_blank"><%=speechEventID%></a></th>
                         <% 
                             for (MetadataKey metadataKey : metadataKeys){ 
                             String metadataValue = speechEvent.getMetadataValue(metadataKey);

diff --git a/web/jsp/testvideo.jsp b/web/jsp/testvideo.jsp