NewLineToNewParagraph.java

//Dstl (c) Crown Copyright 2017
package uk.gov.dstl.baleen.contentmanipulators;

import java.util.Collections;
import java.util.HashSet;
import java.util.LinkedList;
import java.util.List;
import java.util.Set;

import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.nodes.Node;

import uk.gov.dstl.baleen.contentmanipulators.helpers.ContentManipulator;

/**
 * Converts BR tags into new paragraphs.
 * 
 * If BR tags exist in a paragraph or other tag we probably want Baleen to process these are as
 * separate block of text. This manipulator uses BR tags to introduce new paragraphs.
 * 
 * If the BR tag occurs within a paragraph, then the paragraph is split into multiple sub
 * paragraphs. If the br occurs elsewhere (eg in a td or li) then the a set of paragraphs are
 * introduced into the element.
 */
public class NewLineToNewParagraph implements ContentManipulator {

	@Override
	public void manipulate(Document document) {

		// Find elements which need to be spilt up
		Set<Element> elementsWithBr = new HashSet<>();
		document.select("br").forEach(e -> elementsWithBr.add(e.parent()));

		// For each parent
		elementsWithBr.forEach(e -> {
			List<Element> runs = collectRuns(document, e);
			if (!runs.isEmpty()) {
				addRunsToDom(e, runs);
			}
		});
	}

	/**
	 * Collect tags which are on the same line (unbroken by BRs)
	 *
	 * @param document the document
	 * @param e the e
	 * @return the list
	 */
	private List<Element> collectRuns(Document document, Element e) {
		List<Element> runs = new LinkedList<>();
		Element run = null;
		for (Node c : e.childNodesCopy()) {

			if (c instanceof Element && ("br".equalsIgnoreCase(((Element) c).tagName()))) {
				// If we hit a br then add the old run and start a new one
				if (run != null) {
					runs.add(run);
					run = null;
				}
			} else {
				// If not a br then add this node to the other
				if (run == null) {
					run = document.createElement("p");
				}
				run.appendChild(c);
			}
		}

		// Add the last run
		if (run != null) {
			runs.add(run);
		}

		return runs;
	}

	/**
	 * Adds each new line (a run) to the documnet as a paragraph.
	 *
	 * @param e the element at which to add the runs.
	 * @param runs the runs
	 */
	private void addRunsToDom(Element e, List<Element> runs) {
		// Add these new spans into the DOM
		if ("p".equalsIgnoreCase(e.tagName())) {
			// If this is a p, then just add below it
			// reverse order so the first element of runs ends up closest to p as it should be
			Collections.reverse(runs);
			runs.forEach(e::after);
			// Delete the old paragraph
			e.remove();
		} else {
			// If we aren't in a p (eg in a li) then lets add paragraphs to this element
			// But first clear it out
			e.children().remove();
			runs.forEach(e::appendChild);
		}
	}

}