revert commit 3ea708b deletion of DocxLoader

src/main/java/eu/snik/tag/DocxLoader.java

@@ -0,0 +1,185 @@
+package eu.snik.tag;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.HashSet;
+import java.util.LinkedHashSet;
+import java.util.List;
+
+import org.apache.commons.lang3.StringUtils;
+import org.docx4j.Docx4J;
+import org.docx4j.TextUtils;
+import org.docx4j.openpackaging.exceptions.Docx4JException;
+import org.docx4j.wml.ContentAccessor;
+import org.docx4j.wml.R;
+
+import jakarta.xml.bind.JAXBElement;
+import jakarta.xml.bind.JAXBException;
+
+
+/** Extracts SNIK classes from a tagged DOCX file. */
+public class DocxLoader extends Loader {
+
+	//	private static ObjectFactory factory = Context.getWmlObjectFactory();
+	//	private static int commentId = 10000;
+
+	/**
+	 * Creates a new instance for the DOCX loader to load one DOCX file
+	 * @param in Input stream for the DOCX file to load
+	 * @throws IOException Exceptions while trying to open/read the file
+	 */
+	public DocxLoader(InputStream in) throws IOException {
+		super(in);
+	}
+
+	/** from https://stackoverflow.com/questions/19676282/docx4j-find-and-replace */
+	static List<Object> getAllElementsFromObject(Object obj, Class<?> toSearch) {
+		List<Object> result = new ArrayList<Object>();
+		if (obj instanceof JAXBElement) obj = ((JAXBElement<?>) obj).getValue();
+
+		if (obj.getClass().equals(toSearch)) result.add(obj); else if (obj instanceof ContentAccessor) {
+			List<?> children = ((ContentAccessor) obj).getContent();
+			for (Object child : children) {
+				result.addAll(getAllElementsFromObject(child, toSearch));
+			}
+		}
+		return result;
+	}
+
+	/** @return the complete text from the DOCX file without any formatting */
+	@Override
+	public String getText() {
+		try {
+			var wordMLPackage = Docx4J.load(in());
+
+			var doc = wordMLPackage.getMainDocumentPart();
+			var parts = new ArrayList<String>();
+
+			List<Object> texts = getAllElementsFromObject(doc, org.docx4j.wml.Text.class);
+			for (Object t : texts) {
+				org.docx4j.wml.Text content = (org.docx4j.wml.Text) t;
+				parts.add(content.getValue());
+			}
+			return parts
+				.stream()
+				.reduce(
+					(a, b) -> {
+						if (
+							(a.endsWith(" ") || b.startsWith(" ")) ||
+							(b.startsWith(".")) ||
+							(b.startsWith(",")) ||
+							(b.startsWith(";")) ||
+							(b.startsWith("’")) ||
+							(a.endsWith("(") || b.startsWith(")"))
+						) {
+							return a + b;
+						}
+
+						if (a.endsWith(",") || a.endsWith(".")) {
+							return a + (b.startsWith(" ") ? b : (" " + b));
+						}
+
+						return a + '\n' + b;
+					}
+				)
+				.get();
+		} catch (Docx4JException e) {
+			throw new RuntimeException("Error loading Docx File", e);
+		}
+	}
+
+	record TagClass(String tag, String description, Subtop subtop) {}
+
+	/**	@return all classes extracted from the tagged parts of the DOCX document*/
+	@Override
+	public Collection<Clazz> getClasses() {
+		try {
+			var wordMLPackage = Docx4J.load(in());
+			var doc = wordMLPackage.getMainDocumentPart();
+			//List<Comment> comments = doc.getCommentsPart().getContents().getComment();
+
+			TagClass[] tagClasses = {
+				new TagClass("w:i", "Entity Type", Subtop.EntityType),
+				new TagClass("w:b", "Role", Subtop.Role),
+				new TagClass("w:u", "Function", Subtop.Function),
+			};
+
+			var classes = new LinkedHashSet<Clazz>();
+			var processedRuns = new HashSet<R>();
+			var processedLabels = new HashSet<String>();
+			//var warnings = new HashSet<String>(); // prevent the same warning from showing multiple times
+
+			for (var tc : tagClasses) {
+				String xpath = "//w:r[w:rPr/" + tc.tag + "[not(@w:val='false')]]";
+				// find next bold/italic/underlined passage
+				@SuppressWarnings("unchecked")
+				var runs = (List<R>) (List<?>) doc.getJAXBNodesViaXPath(xpath, false);
+				// remove previously processed passages
+				runs.removeAll(processedRuns); // we cannot handle overlapping tags right now
+
+				for (R run : runs) {
+					String text = TextUtils.getText(run);
+					String label = StringUtils.strip(text, "., ");
+					if (label.length() > 80) {
+						continue;
+					} // too long texts seems to be erroneously detected
+					//label = label.replaceAll("[^A-Za-z0-9 ]", ""); // removing non-alphanumerical characters leads to missing matches in the text tab
+					String filterLabel = label.replaceAll("[^A-Za-z0-9 ]", "").replaceAll("(the)|(and)|(or)", "");
+					if (filterLabel.length() < 4 && !filterLabel.matches("[A-Z]{3}")) {} // abbreviations with 3 letters are OK
+					if (filterLabel.length() < 3) {
+						continue;
+					} // abbreviations
+					processedRuns.add(run);
+
+					/*
+				Comment comment = factory.createCommentsComment();
+				comments.add(comment);
+				comment.setId(BigInteger.valueOf(++commentId));
+				Text commentText = factory.createText();
+				commentText.setValue("this is a comment for "+label);
+				comment.getContent().add(commentText);
+				CommentReference commentRef = factory.createRCommentReference();
+				run.getContent().add(commentRef);
+				commentRef.setId(BigInteger.valueOf(commentId));				
+				 */
+					Clazz clazz = new Clazz(label, labelToLocalName(label), tc.subtop);
+					//System.out.println(text+" "+ clazz);
+					if (processedLabels.contains(label)) {
+						classes
+							.stream()
+							.filter(c -> c.subtop() != clazz.subtop())
+							.findAny()
+							.ifPresent(
+								c -> {
+									System.err.println(
+										"Typenkonflikt. Klasse \"" +
+										label +
+										"\" ist getagged als sowohl " +
+										c.subtop() +
+										" als auch " +
+										clazz.subtop() +
+										". " +
+										"Ignoriere " +
+										clazz.subtop() +
+										". Bitte beheben Sie den Konflikt im Dokument."
+									);
+								}
+							);
+						continue;
+					}
+					classes.add(clazz);
+					processedLabels.add(label);
+				}
+			}
+
+			//warningCallback.ifPresent(c->c.accept(warnings.stream().reduce("", (a,b)->a+"\n"+b)));
+			System.out.println(classes.size() + " classes extracted.");
+
+			return classes;
+		} catch (Docx4JException | JAXBException e) {
+			throw new RuntimeException(e);
+		}
+	}
+}