Add support for more html tags on the Document/PDF exporter

dmp-backend/web/src/main/java/eu/eudat/logic/utilities/documents/word/HtmlToWorldBuilder.java

@@ -8,34 +8,28 @@ import org.jsoup.nodes.TextNode;
 import org.jsoup.select.Elements;
 import org.jsoup.select.NodeTraversor;
 import org.jsoup.select.NodeVisitor;
-import org.openxmlformats.schemas.wordprocessingml.x2006.main.STHighlightColor;
+import org.openxmlformats.schemas.wordprocessingml.x2006.main.*;
 
-import java.util.LinkedHashMap;
+import java.math.BigInteger;
-import java.util.Map;
+import java.util.*;
 
 public class HtmlToWorldBuilder implements NodeVisitor {
 
     private final Map<String, Boolean> properties = new LinkedHashMap<>();
-    private final XWPFParagraph paragraph;
+    private XWPFParagraph paragraph;
     private XWPFRun run;
     private Boolean dumpRun;
     private final String identation;
     private Boolean isIdentationUsed;
+    private XWPFNumbering numbering;
+    private Queue<BigInteger> abstractNumId;
+    private BigInteger numberingLevel;
 
     public static void convert(XWPFDocument document, Document htmlDocument, String identation) {
-        Elements htmlParagraphs = htmlDocument.select("p");
-        if (!htmlParagraphs.isEmpty()) {
-            for (Element htmlParagraph : htmlParagraphs) {
-                XWPFParagraph paragraph = document.createParagraph();
-                HtmlToWorldBuilder htmlToWorldBuilder = new HtmlToWorldBuilder(paragraph, identation);
-                NodeTraversor.traverse(htmlToWorldBuilder, htmlParagraph);
-            }
-        } else {
         XWPFParagraph paragraph = document.createParagraph();
         HtmlToWorldBuilder htmlToWorldBuilder = new HtmlToWorldBuilder(paragraph, identation);
         NodeTraversor.traverse(htmlToWorldBuilder, htmlDocument);
     }
-    }
 
     public HtmlToWorldBuilder(XWPFParagraph paragraph, String identation) {
         this.paragraph = paragraph;
@@ -44,6 +38,8 @@ public class HtmlToWorldBuilder implements NodeVisitor {
         this.identation = identation;
         this.isIdentationUsed = false;
         this.run.setFontSize(11);
+        this.abstractNumId = new ArrayDeque<>();
+        this.numberingLevel = BigInteger.valueOf(-1);
     }
 
     @Override
@@ -62,11 +58,11 @@ public class HtmlToWorldBuilder implements NodeVisitor {
             this.run.setFontSize(11);
             this.dumpRun = false;
         }
-        parseProperties();
+        parseProperties(node);
         properties.clear();
     }
 
-    private void parseProperties() {
+    private void parseProperties(Node node) {
         properties.entrySet().forEach(stringBooleanEntry -> {
             switch (stringBooleanEntry.getKey()) {
                 case "i" :
@@ -85,6 +81,9 @@ public class HtmlToWorldBuilder implements NodeVisitor {
                     this.run.setFontSize(stringBooleanEntry.getValue() ? 8 : 11);
                     break;
                 case "del":
+                case "strike":
+                case "strikethrough":
+                case "s":
                     this.run.setStrikeThrough(stringBooleanEntry.getValue());
                     break;
                 case "mark":
@@ -96,6 +95,75 @@ public class HtmlToWorldBuilder implements NodeVisitor {
                 case "sup":
                     this.run.setSubscript(stringBooleanEntry.getValue() ? VerticalAlign.SUPERSCRIPT : VerticalAlign.BASELINE);
                     break;
+                case "div":
+                case "p":
+                    this.paragraph = this.paragraph.getDocument().createParagraph();
+                    this.run = this.paragraph.createRun();
+                    if (stringBooleanEntry.getValue()) {
+                        if (node.hasAttr("align")) {
+                            String alignment = node.attr("align");
+                            this.paragraph.setAlignment(ParagraphAlignment.valueOf(alignment.toUpperCase(Locale.ROOT)));
+                        }
+                    }
+                    break;
+                case "blockquote":
+                    this.paragraph = this.paragraph.getDocument().createParagraph();
+                    this.run = this.paragraph.createRun();
+                    if (stringBooleanEntry.getValue()) {
+                        this.paragraph.setIndentationLeft(720);
+                    }
+                    break;
+                case "ul":
+                    if (stringBooleanEntry.getValue()) {
+                        createNumbering(STNumberFormat.BULLET);
+                    } else {
+                        this.paragraph = this.paragraph.getDocument().createParagraph();
+                        this.run = this.paragraph.createRun();
+                        this.numberingLevel = this.numberingLevel.subtract(BigInteger.ONE);
+                        ((ArrayDeque)this.abstractNumId).removeLast();
+                    }
+                    break;
+                case "ol":
+                    if (stringBooleanEntry.getValue()) {
+                        createNumbering(STNumberFormat.DECIMAL);
+                    } else {
+                        this.paragraph = this.paragraph.getDocument().createParagraph();
+                        this.run = this.paragraph.createRun();
+                        this.numberingLevel = this.numberingLevel.subtract(BigInteger.ONE);
+                        ((ArrayDeque)this.abstractNumId).removeLast();
+                    }
+                    break;
+                case "li":
+                    if (stringBooleanEntry.getValue()) {
+                        this.paragraph = this.paragraph.getDocument().createParagraph();
+                        this.paragraph.setIndentationLeft(720 * numberingLevel.intValue());
+                        this.run = this.paragraph.createRun();
+                        this.paragraph.setNumID(((ArrayDeque<BigInteger>)abstractNumId).getLast());
+                    }
+                    break;
+                case "font":
+                    if (stringBooleanEntry.getValue()) {
+                        if (node.hasAttr("color")) {
+                            this.run.setColor(node.attr("color").substring(1));
+                        }
+                    } else {
+                        this.run.setColor("000000");
+                    }
+                    break;
+                case "a":
+                    if (stringBooleanEntry.getValue()) {
+                        if (node.hasAttr("href")) {
+                            this.run = createHyperLinkRun(node.attr("href"));
+                            this.run.setColor("0000FF");
+                            this.run.setUnderline(UnderlinePatterns.SINGLE);
+                        }
+                    } else {
+                        this.run = paragraph.createRun();
+                    }
+                    break;
+                case "br":
+                    this.run.addCarriageReturn();
+                    break;
             }
         });
     }
@@ -104,7 +172,50 @@ public class HtmlToWorldBuilder implements NodeVisitor {
     public void tail(Node node, int i) {
         String name = node.nodeName();
         properties.put(name, false);
-        parseProperties();
+        parseProperties(node);
         properties.clear();
     }
+
+    //GK: This function creates one numbering.xml for the word document and adds a specific format.
+    //It imitates the numbering.xml that is usually generated by word editors like LibreOffice
+    private void createNumbering(STNumberFormat.Enum format) {
+        CTAbstractNum ctAbstractNum = CTAbstractNum.Factory.newInstance();
+        if (this.numbering == null) this.numbering = this.paragraph.getDocument().createNumbering();
+        BigInteger tempNumId = BigInteger.ONE;
+        boolean found = false;
+        while (!found) {
+            Object o = numbering.getAbstractNum(tempNumId);
+            found = (o == null);
+            if (!found) tempNumId = tempNumId.add(BigInteger.ONE);
+        }
+        ctAbstractNum.setAbstractNumId(tempNumId);
+        CTLvl ctLvl = ctAbstractNum.addNewLvl();
+        this.numberingLevel = numberingLevel.add(BigInteger.ONE);
+        ctLvl.setIlvl(numberingLevel);
+        ctLvl.addNewNumFmt().setVal(format);
+        ctLvl.addNewStart().setVal(BigInteger.ONE);
+        if (format == STNumberFormat.BULLET) {
+            ctLvl.addNewLvlText().setVal("\u2022");
+        } else {
+            ctLvl.addNewLvlText().setVal("%1.");
+        }
+        XWPFAbstractNum xwpfAbstractNum = new XWPFAbstractNum(ctAbstractNum);
+
+        this.abstractNumId.add(this.numbering.addAbstractNum(xwpfAbstractNum));
+        this.numbering.addNum(((ArrayDeque<BigInteger>)abstractNumId).getLast());
+    }
+
+    private XWPFHyperlinkRun createHyperLinkRun(String uri) {
+        String rId = this.paragraph.getDocument().getPackagePart().addExternalRelationship(uri, XWPFRelation.HYPERLINK.getRelation()).getId();
+
+        CTHyperlink cthyperLink=paragraph.getCTP().addNewHyperlink();
+        cthyperLink.setId(rId);
+        cthyperLink.addNewR();
+
+        return new XWPFHyperlinkRun(
+                cthyperLink,
+                cthyperLink.getRArray(0),
+                paragraph
+        );
+    }
 }