Java tutorial
/******************************************************************************* * Copyright (c) 2010 Torsten Zesch. All rights reserved. This program and the * accompanying materials are made available under the terms of the GNU Lesser * Public License v3 which accompanies this distribution, and is available at * http://www.gnu.org/licenses/lgpl.html * * Contributors: * Torsten Zesch - initial API and implementation * Samy Ateia - provided a patch via the JWPL mailing list ******************************************************************************/ package org.dbpedia.spotlight.mediawiki; import java.util.ArrayList; import java.util.LinkedList; import java.util.List; import java.util.Stack; import org.apache.commons.logging.Log; import org.apache.commons.logging.LogFactory; import de.tudarmstadt.ukp.wikipedia.parser.*; import de.tudarmstadt.ukp.wikipedia.parser.Content.FormatType; import de.tudarmstadt.ukp.wikipedia.parser.mediawiki.MediaWikiTemplateParser; import de.tudarmstadt.ukp.wikipedia.parser.mediawiki.ResolvedTemplate; import de.tudarmstadt.ukp.wikipedia.parser.mediawiki.SpanManager; import de.tudarmstadt.ukp.wikipedia.parser.mediawiki.SrcPosRangeChecker; /** * This is a Parser for MediaWiki Source.<br/> * It exist a MediaWikiParserFactory, to get an Instance of this Parser.<br/> * But, if you want to, you can create a parser by yourself. * * @author CJacobi * */ public class ModularParser implements MediaWikiParser, MediaWikiContentElementParser { private final Log logger = LogFactory.getLog(getClass()); // Options, set by the ParserFactory private String lineSeparator; private List<String> categoryIdentifers; private List<String> languageIdentifers; private List<String> imageIdentifers; private MediaWikiTemplateParser templateParser; private boolean showImageText = false; private boolean deleteTags = true; private boolean showMathTagContent = true; private boolean calculateSrcSpans = true; private String lastTemplate; /** * Creates a unconfigurated Parser... */ public ModularParser() { } /** * Creates a fully configurated parser... */ public ModularParser(String lineSeparator, List<String> languageIdentifers, List<String> categoryIdentifers, List<String> imageIdentifers, boolean showImageText, boolean deleteTags, boolean showMathTagContent, boolean calculateSrcSpans, MediaWikiTemplateParser templateParser) { setLineSeparator(lineSeparator); setLanguageIdentifers(languageIdentifers); setCategoryIdentifers(categoryIdentifers); setImageIdentifers(imageIdentifers); setShowImageText(showImageText); setDeleteTags(deleteTags); setShowMathTagContent(showMathTagContent); setCalculateSrcSpans(calculateSrcSpans); setTemplateParser(templateParser); } /** * Look at MediaWikiParserFactory for a description... */ @Override public String getLineSeparator() { return lineSeparator; } /** * Look at MediaWikiParserFactory for a description... */ public void setLineSeparator(String lineSeparator) { this.lineSeparator = lineSeparator; } /** * Look at MediaWikiParserFactory for a description... */ public List<String> getLanguageIdentifers() { return languageIdentifers; } /** * Look at MediaWikiParserFactory for a description... */ public void setLanguageIdentifers(List<String> languageIdentifers) { this.languageIdentifers = listToLowerCase(languageIdentifers); } /** * Look at MediaWikiParserFactory for a description... */ public List<String> getCategoryIdentifers() { return categoryIdentifers; } /** * Look at MediaWikiParserFactory for a description... */ public void setCategoryIdentifers(List<String> categoryIdentifers) { this.categoryIdentifers = listToLowerCase(categoryIdentifers); } /** * Look at MediaWikiParserFactory for a description... */ public List<String> getImageIdentifers() { return imageIdentifers; } /** * Look at MediaWikiParserFactory for a description... */ public void setImageIdentifers(List<String> imageIdentifers) { this.imageIdentifers = listToLowerCase(imageIdentifers); } /** * Look at MediaWikiParserFactory for a description... */ public MediaWikiTemplateParser getTemplateParser() { return templateParser; } /** * Look at MediaWikiParserFactory for a description... */ public void setTemplateParser(MediaWikiTemplateParser templateParser) { this.templateParser = templateParser; } /** * Look at MediaWikiParserFactory for a description... */ public boolean showImageText() { return showImageText; } /** * Look at MediaWikiParserFactory for a description... */ public void setShowImageText(boolean showImageText) { this.showImageText = showImageText; } /** * Look at MediaWikiParserFactory for a description... */ public boolean deleteTags() { return deleteTags; } /** * Look at MediaWikiParserFactory for a description... */ public void setDeleteTags(boolean deleteTags) { this.deleteTags = deleteTags; } /** * Look at MediaWikiParserFactory for a description... */ public boolean showMathTagContent() { return showMathTagContent; } /** * Look at MediaWikiParserFactory for a description... */ public void setShowMathTagContent(boolean showMathTagContent) { this.showMathTagContent = showMathTagContent; } /** * Look at MediaWikiParserFactory for a description... */ public boolean calculateSrcSpans() { return calculateSrcSpans; } /** * Look at MediaWikiParserFactory for a description... */ public void setCalculateSrcSpans(boolean calculateSrcSpans) { this.calculateSrcSpans = calculateSrcSpans; } @Override public String getLastTemplate() { return lastTemplate; } public void setLastTemplate(String lastTemplate) { this.lastTemplate = lastTemplate; } /** * Converts a List of Strings to lower case Strings. */ private List<String> listToLowerCase(List<String> l) { List<String> result = new ArrayList<String>(); for (String s : l) { result.add(s.toLowerCase()); } return result; } /** * Look at the MediaWikiParser interface for a description... */ @Override public String configurationInfo() { StringBuilder result = new StringBuilder(); result.append("MediaWikiParser configuration:\n"); result.append("ParserClass: " + this.getClass() + "\n"); result.append("ShowImageText: " + showImageText + "\n"); result.append("DeleteTags: " + deleteTags + "\n"); result.append("ShowMathTagContent: " + showMathTagContent + "\n"); result.append("CalculateSrcSpans: " + calculateSrcSpans + "\n"); result.append("LanguageIdentifers: "); for (String s : languageIdentifers) { result.append(s + " "); } result.append("\n"); result.append("CategoryIdentifers: "); for (String s : categoryIdentifers) { result.append(s + " "); } result.append("\n"); result.append("ImageIdentifers: "); for (String s : imageIdentifers) { result.append(s + " "); } result.append("\n"); result.append("TemplateParser: " + templateParser.getClass() + "\n"); result.append(templateParser.configurationInfo()); return result.toString(); } /** * Checks if the configuration is runnable. */ private boolean runConfig() { if (lineSeparator == null) { logger.error("Set lineSeparator"); return false; } if (categoryIdentifers == null) { logger.error("Set categoryIdentifers"); return false; } if (languageIdentifers == null) { logger.error("Set languageIdentifers"); return false; } if (imageIdentifers == null) { logger.error("Set imageIdentifers"); return false; } if (templateParser == null) { logger.error("Set templateParser"); return false; } return true; } /** * Look at the MediaWikiParser for a description... */ @Override public ParsedPage parse(String src) { // check if the configuration is runnable. if (!runConfig()) { return null; } // check if the is something to parse. somtimes there is an empty string // due to an error of other clases... if (src == null || src.length() == 0) { return null; } // creates a new span manager with the given source, appending a newline // to avoid errors. SpanManager sm = new SpanManager(src.replace('\t', ' ') + lineSeparator); if (calculateSrcSpans) { sm.enableSrcPosCalculation(); } // Creating a new ParsePage, which will be filled with information in // the parseing process. ParsedPage ppResult = new ParsedPage(); // Creating a new Parameter Container ContentElementParsingParameters cepp = new ContentElementParsingParameters(); // Deletes comments out of the Source deleteComments(sm); // Deletes any TOC Tags, these are not usesd in this parser. deleteTOCTag(sm); // Removing the Content which should not parsed but integrated later in // the resulting text sm.manageList(cepp.noWikiSpans); parseSpecifiedTag(sm, cepp.noWikiSpans, cepp.noWikiStrings, "PRE", " "); parseSpecifiedTag(sm, cepp.noWikiSpans, cepp.noWikiStrings, "NOWIKI"); if (cepp.noWikiSpans.size() == 0) { sm.removeManagedList(cepp.noWikiSpans); } // Parseing the Math Tags... sm.manageList(cepp.mathSpans); parseSpecifiedTag(sm, cepp.mathSpans, cepp.mathStrings, "MATH"); if (cepp.mathSpans.size() == 0) { sm.removeManagedList(cepp.mathSpans); } // Parseing the Templates (the Span List will be added to the managed // lists by the function) parseTemplates(sm, cepp.templateSpans, cepp.templates, ppResult); //Deleting Templates from the Media Text. To get the plain text. for (Span s : cepp.templateSpans) { sm.delete(s); } sm.removeManagedList(cepp.templateSpans); // Parsing all other Tags parseTags(sm, cepp.tagSpans); // Converting <gallery>s to normal Images, this is not beautiful, but // a simple solution.. convertGalleriesToImages(sm, cepp.tagSpans); // Parsing Links and Images. parseImagesAndInternalLinks(sm, cepp.linkSpans, cepp.links); // Creating a list of Line Spans to work with lines in the following // functions LinkedList<Span> lineSpans = new LinkedList<Span>(); getLineSpans(sm, lineSpans); // Removing the Category Links from the Links list, and crating an // ContentElement for these links... ppResult.setCategoryElement(getSpecialLinks(sm, cepp.linkSpans, cepp.links, " - ", categoryIdentifers)); // Removing the Language Links from the Links list, and crating an // ContentElement for these links... ppResult.setLanguagesElement(getSpecialLinks(sm, cepp.linkSpans, cepp.links, " - ", languageIdentifers)); // Parsing and Setting the Sections... the main work is done in parse // sections! ppResult.setSections(EmptyStructureRemover.eliminateEmptyStructures(parseSections(sm, cepp, lineSpans))); // Finding and Setting the paragraph which is concidered as the "First" setFirstParagraph(ppResult); // check the calculated source positions, and reset them if necessary. if (calculateSrcSpans) { SrcPosRangeChecker.checkRange(ppResult); } // So it is done... return ppResult; } /** * Deleting all comments out of the SpanManager...<br/> * <!-- COMMENT --> */ private void deleteComments(SpanManager sm) { int start = 0; while ((start = sm.indexOf("<!--", start)) != -1) { int end = sm.indexOf("-->", start + 4) + 3; if (end == -1 + 3) { end = sm.length(); } // Remove the one lineSeparator too, if the whole line is a comment! try { if (lineSeparator.equals(sm.substring(start - lineSeparator.length(), start)) && lineSeparator.equals(sm.substring(end, end + lineSeparator.length()))) { end += lineSeparator.length(); } } catch (IndexOutOfBoundsException e) { } sm.delete(start, end); } } /** * Deleteing ALL TOC Tags */ private void deleteTOCTag(SpanManager sm) { // delete all __TOC__ from SRC int temp = 0; while ((temp = sm.indexOf("__TOC__", temp)) != -1) { sm.delete(temp, temp + 2 + 3 + 2); } // delete all __NOTOC__ from SRC temp = 0; while ((temp = sm.indexOf("__NOTOC__", temp)) != -1) { sm.delete(temp, temp + 2 + 5 + 2); } } private ContentElement getSpecialLinks(SpanManager sm, List<Span> linkSpans, List<Link> links, String linkSpacer, List<String> identifers) { ContentElement result = new ContentElement(); StringBuilder text = new StringBuilder(); List<Link> localLinks = new ArrayList<Link>(); for (int i = links.size() - 1; i >= 0; i--) { String identifer = getLinkNameSpace(links.get(i).getTarget()); if (identifer != null && identifers.indexOf(identifer) != -1) { Link l = links.remove(i); Span s = linkSpans.remove(i); String linkText = sm.substring(s); sm.delete(s); l.setHomeElement(result); s.adjust(-s.getStart() + text.length()); text.append(linkText + linkSpacer); localLinks.add(l); //TODO add type? } } int len = text.length(); if (len != 0) { text.delete(len - linkSpacer.length(), len); } result.setText(text.toString()); result.setLinks(localLinks); if (result.empty()) { return null; } else { return result; } } private void getLineSpans(SpanManager sm, LinkedList<Span> lineSpans) { sm.manageList(lineSpans); int start = 0; int end; while ((end = sm.indexOf(lineSeparator, start)) != -1) { lineSpans.add(new Span(start, end).trimTrail(sm)); start = end + lineSeparator.length(); } lineSpans.add(new Span(start, sm.length()).trimTrail(sm)); while (!lineSpans.isEmpty() && lineSpans.getFirst().length() == 0) { lineSpans.removeFirst(); } while (!lineSpans.isEmpty() && lineSpans.getLast().length() == 0) { lineSpans.removeLast(); } } private SectionContainer parseSections(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans) { List<SectionContent> contentSections = new ArrayList<SectionContent>(); SectionContent sc = new SectionContent(1); if (calculateSrcSpans) { try { sc.setSrcSpan(new SrcSpan(sm.getSrcPos(lineSpans.getFirst().getStart()), -1)); } catch (Exception e) { System.out.println("Parse error :" + sm.toString()); } } // Identify the Line Type and call the necessary Function for the // further handling... while (!lineSpans.isEmpty()) { Span s = lineSpans.getFirst(); lineType t = getLineType(sm, s); switch (t) { case SECTION: contentSections.add(sc); int level = getSectionLevel(sm, s); sc = new SectionContent( parseContentElement(sm, cepp, new Span(s.getStart() + level, s.getEnd() - level).trim(sm)), level); lineSpans.removeFirst(); if (calculateSrcSpans) { sc.setSrcSpan(new SrcSpan(sm.getSrcPos(s.getStart()), -1)); } break; case HR: // remove the HR (----) and handle the rest as a parapraph line removeHr(sm, s); t = lineType.PARAGRAPH; case PARAGRAPH: case PARAGRAPH_BOXED: case PARAGRAPH_INDENTED: sc.addParagraph(buildParagraph(sm, cepp, lineSpans, t)); break; case NESTEDLIST: case NESTEDLIST_NR: sc.addNestedList(buildNestedList(sm, cepp, lineSpans, t)); break; case DEFINITIONLIST: sc.addDefinitionList(buildDefinitionList(sm, cepp, lineSpans)); break; case TABLE: sc.addTable(buildTable(sm, cepp, lineSpans)); break; case EMPTYLINE: lineSpans.removeFirst(); break; default: logger.error("unknown lineStart!: \"" + sm.substring(s) + "\""); lineSpans.removeFirst(); } } // add the remaining Section to the list. contentSections.add(sc); return buildSectionStructure(contentSections); } private Span removeHr(SpanManager sm, Span s) { int start = s.getStart(); final int end = s.getEnd(); while (sm.charAt(start) == '-' && start < end) { start++; } return s.setStart(start).trim(sm); } /** * The Line Types wich are possible... */ private enum lineType { SECTION, TABLE, NESTEDLIST, NESTEDLIST_NR, DEFINITIONLIST, HR, PARAGRAPH, PARAGRAPH_INDENTED, PARAGRAPH_BOXED, EMPTYLINE } /** * Retunrns the Type of a line, this is mainly done by the First Char of the * Line... */ private lineType getLineType(SpanManager sm, Span lineSpan) { switch (lineSpan.charAt(0, sm)) { case '{': if (lineSpan.charAt(1, sm) == '|') { return lineType.TABLE; } else { return lineType.PARAGRAPH; } case '=': if (lineSpan.length() > 2 && sm.charAt(lineSpan.getEnd() - 1) == '=') { return lineType.SECTION; } else { return lineType.PARAGRAPH; } case '-': if (lineSpan.charAt(1, sm) == '-' && lineSpan.charAt(2, sm) == '-' && lineSpan.charAt(3, sm) == '-') { return lineType.HR; } else { return lineType.PARAGRAPH; } case '*': return lineType.NESTEDLIST; case '#': return lineType.NESTEDLIST_NR; case ';': return lineType.DEFINITIONLIST; case ':': if (lineSpan.length() > 1) { if (lineSpan.length() > 2 && lineSpan.charAt(1, sm) == '{' && lineSpan.charAt(2, sm) == '|') { return lineType.TABLE; } else { return lineType.PARAGRAPH_INDENTED; } } else { return lineType.PARAGRAPH; } case ' ': int nonWSPos = lineSpan.nonWSCharPos(sm); switch (lineSpan.charAt(nonWSPos, sm)) { case Span.ERRORCHAR: return lineType.EMPTYLINE; case '{': if (lineSpan.charAt(nonWSPos + 1, sm) == '|') { return lineType.TABLE; } default: return lineType.PARAGRAPH_BOXED; } case Span.ERRORCHAR: return lineType.EMPTYLINE; default: return lineType.PARAGRAPH; } } /** * Returns the number of Equality Chars which are used to specify the level * of the Section. */ private int getSectionLevel(SpanManager sm, Span sectionNameSpan) { int begin = sectionNameSpan.getStart(); int end = sectionNameSpan.getEnd(); int level = 0; try { while ((sm.charAt(begin + level) == '=') && (sm.charAt(end - 1 - level) == '=')) { level++; } } catch (StringIndexOutOfBoundsException e) { // there is no need to do anything! logger.debug("EXCEPTION IS OK: " + e); } if (begin + level == end) { level = (level - 1) / 2; } return level; } /** * Takes a list of SectionContent and returns a SectionContainer with the * given SectionContent s in the right structure. */ private SectionContainer buildSectionStructure(List<SectionContent> scl) { SectionContainer result = new SectionContainer(0); for (SectionContent sContent : scl) { int contentLevel = sContent.getLevel(); SectionContainer sContainer = result; // get the right SectionContainer or create it for (int containerLevel = result.getLevel() + 1; containerLevel < contentLevel; containerLevel++) { int containerSubSections = sContainer.nrOfSubSections(); if (containerSubSections != 0) { Section temp = sContainer.getSubSection(containerSubSections - 1); if (temp.getClass() == SectionContainer.class) { sContainer = (SectionContainer) temp; } else { SectionContainer sct = new SectionContainer(temp.getTitleElement(), containerLevel); sct.addSection(temp); if (calculateSrcSpans) { sct.setSrcSpan(temp.getSrcSpan()); } temp.setTitleElement(null); temp.setLevel(containerLevel + 1); sContainer.removeSection(temp); sContainer.addSection(sct); sContainer = sct; } } else { sContainer = new SectionContainer(null, containerLevel); } } sContainer.addSection(sContent); } if (calculateSrcSpans) { result.setSrcSpan(new SrcSpan(0, -1)); } return result; } private boolean startsWithIgnoreCase(String s1, String s2) { final int s2len = s2.length(); if (s1.length() < s2len) { return false; } return s1.substring(0, s2len).equalsIgnoreCase(s2); } private Span getTag(SpanManager sm, int offset) { int start = sm.indexOf("<", offset); if (start == -1) { return null; } int end = sm.indexOf(">", start); if (end == -1) { return null; } Span s = new Span(start, end + 1); if (calculateSrcSpans) { s.setSrcSpan(new SrcSpan(sm.getSrcPos(start), sm.getSrcPos(end) + 1)); } return s; } private String getTagText(SpanManager sm, Span tag) { return sm.substring(new Span(tag.getStart() + 1, tag.getEnd() - 1).trim(sm)); } private void parseSpecifiedTag(SpanManager sm, List<Span> spans, List<String> strings, String specifier) { parseSpecifiedTag(sm, spans, strings, specifier, ""); } private void parseSpecifiedTag(SpanManager sm, List<Span> spans, List<String> strings, String specifier, String prefix) { int offset = 0; Span s; while ((s = getTag(sm, offset)) != null) { offset = s.getEnd(); String tagText = getTagText(sm, s); if (startsWithIgnoreCase(tagText, specifier)) { Span e; while ((e = getTag(sm, offset)) != null) { offset = e.getEnd(); tagText = getTagText(sm, e); if (startsWithIgnoreCase(tagText, "/" + specifier)) { break; } } if (e == null) { /* * OF: Setting e to sm.length()results in ArrayIndexOutOfBoundsExeption if calculateSrcSpans=true */ //e = new Span(sm.length(), sm.length()); e = new Span(Math.max(0, sm.length() - 1), Math.max(0, sm.length() - 1)); } strings.add(sm.substring(s.getEnd(), e.getStart())); Span tSpan = new Span(s.getStart(), e.getEnd()); if (calculateSrcSpans) { tSpan.setSrcSpan(new SrcSpan(sm.getSrcPos(s.getStart()), sm.getSrcPos(e.getEnd()))); } spans.add(tSpan); sm.replace(tSpan, prefix + "(" + specifier + ")"); tSpan.adjustStart(prefix.length()); offset = tSpan.getEnd(); } } } private void parseTags(SpanManager sm, List<Span> spans) { sm.manageList(spans); Span s = new Span(0, 0); Stack<Span> tagSpans = new Stack<Span>(); while ((s = getTag(sm, s.getEnd())) != null) { //spans.add(s); //Logic to remove ref tags such as <ref> </ref> from the Article text if (sm.substring(s.getStart(), s.getEnd()).contains("/>")) { spans.add(s); continue; } if (sm.substring(s.getStart(), s.getEnd()).contains("/")) { if (!tagSpans.empty()) spans.add(new Span(tagSpans.pop().getStart(), s.getEnd())); } else tagSpans.push(s); } if (spans.size() == 0) { sm.removeManagedList(spans); } for (Span span : spans) sm.delete(span); } private void parseTemplates(SpanManager sm, List<Span> resolvedTemplateSpans, List<ResolvedTemplate> resolvedTemplates, ParsedPage pp) { sm.manageList(resolvedTemplateSpans); int pos = -2; Stack<Integer> templateOpenTags = new Stack<Integer>(); while ((pos = sm.indexOf("{{", pos + 2)) != -1) { if (sm.length() > pos + 3 && sm.charAt(pos + 2) == '{' && sm.charAt(pos + 3) != '{') { pos++; } templateOpenTags.push(pos); } boolean isLastTemplate = true; while (!templateOpenTags.empty()) { int templateOpenTag = templateOpenTags.pop(); int templateCloseTag = sm.indexOf("}}", templateOpenTag); if (templateCloseTag == -1) { continue; } int templateOptionTag = sm.indexOf("|", templateOpenTag, templateCloseTag); int templateNameEnd; List<String> templateOptions; if (templateOptionTag != -1) { templateNameEnd = templateOptionTag; templateOptions = tokenize(sm, templateOptionTag + 1, templateCloseTag, "|"); } else { templateNameEnd = templateCloseTag; templateOptions = new ArrayList<String>(); } Span ts = new Span(templateOpenTag, templateCloseTag + 2); if (isLastTemplate) { lastTemplate = sm.substring(new Span(templateOpenTag + 2, templateCloseTag)); isLastTemplate = false; } Template t = new Template(ts, encodeWikistyle(sm.substring(templateOpenTag + 2, templateNameEnd).trim()), templateOptions); if (calculateSrcSpans) { t.setSrcSpan(new SrcSpan(sm.getSrcPos(templateOpenTag), sm.getSrcPos(templateCloseTag + 2))); } t.setPos(ts); ResolvedTemplate rt = templateParser.parseTemplate(t, pp); resolvedTemplateSpans.add(ts); resolvedTemplates.add(rt); sm.replace(ts, rt.getPreParseReplacement()); } if (resolvedTemplateSpans.isEmpty()) { sm.removeManagedList(resolvedTemplateSpans); } } private void convertGalleriesToImages(SpanManager sm, List<Span> tagSpans) { // Quick Hack, not very efficent, should be improved, wont work with // calculateSrcSpans == true ! for (int i = 0; i < tagSpans.size() - 1; i++) { String openText = getTagText(sm, tagSpans.get(i)); if (startsWithIgnoreCase(openText, "GALLERY")) { if (startsWithIgnoreCase(getTagText(sm, tagSpans.get(i + 1)), "/GALLERY")) { // gallery range is tag(i).end() .. tag(i+1).start() Span startSpan = tagSpans.remove(i); Span endSpan = tagSpans.remove(i); i--; StringBuilder sb = new StringBuilder(); // caption (any option will be treated as caption) int eqPos = openText.indexOf('='); if (eqPos != -1) { int captionStart = eqPos + 1; int captionEnd = openText.length(); if (captionStart < captionEnd && openText.charAt(captionStart) == '"' && openText.charAt(captionEnd - 1) == '"') { captionStart++; captionEnd--; } if (captionStart < captionEnd) { sb.append(openText.substring(captionStart, captionEnd) + lineSeparator); } } // images for (String s : tokenize(sm, startSpan.getEnd(), endSpan.getStart(), lineSeparator)) { sb.append("[[" + s + "]]" + lineSeparator); } // replace the source and remove the tags sm.replace(startSpan.getStart(), endSpan.getEnd(), sb.toString()); } else { continue; } } } } private Table buildTable(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans) { Table result = new Table(); int col = -1; int row = 0; int subTables = 0; LinkedList<Span> tableDataSpans = new LinkedList<Span>(); sm.manageList(tableDataSpans); if (calculateSrcSpans) { result.setSrcSpan(new SrcSpan(sm.getSrcPos(lineSpans.getFirst().getStart()), -1)); } lineSpans.removeFirst(); while (!lineSpans.isEmpty()) { Span s = lineSpans.removeFirst(); int pos = s.nonWSCharPos(sm); char c0 = s.charAt(pos, sm); char c1 = s.charAt(pos + 1, sm); if (subTables == 0 && (c0 == '!' || c0 == '|')) { if (!tableDataSpans.isEmpty()) { lineSpans.addFirst(s); SrcSpan ei = null; if (calculateSrcSpans) { ei = new SrcSpan(sm.getSrcPos(tableDataSpans.getFirst().getStart() - 1) + 1, -1); } TableElement te = new TableElement(parseSections(sm, cepp, tableDataSpans), row, col); te.setSrcSpan(ei); result.addTableElement(te); lineSpans.removeFirst(); } col++; if (c1 == '-') { row++; col = -1; continue; } else if (c0 == '|' && c1 == '}') { sm.removeManagedList(tableDataSpans); if (calculateSrcSpans) { result.getSrcSpan().setEnd(sm.getSrcPos(s.getEnd())); } return result; } else if (c0 == '|' && c1 == '+') { result.setTitleElement( parseContentElement(sm, cepp, new Span(s.getStart() + pos + 2, s.getEnd()).trim(sm))); continue; } else { int multipleCols; if ((multipleCols = sm.indexOf("||", s.getStart() + pos + 1, s.getEnd())) != -1) { lineSpans.addFirst(new Span(multipleCols + 1, s.getEnd())); s.setEnd(multipleCols); } int optionTagPos = sm.indexOf("|", s.getStart() + pos + 1, s.getEnd()); if (optionTagPos != -1) { s.setStart(optionTagPos + 1).trim(sm); } else { s.adjustStart(pos + 1).trim(sm); } } } else if (c0 == '|' && c1 == '}') { subTables--; } else if (c0 == '{' && c1 == '|') { subTables++; } tableDataSpans.addLast(s); } if (tableDataSpans.size() != 0) { SrcSpan ei = null; if (calculateSrcSpans) { ei = new SrcSpan(sm.getSrcPos(tableDataSpans.getFirst().getStart() - 1) + 1, -1); } TableElement te = new TableElement(parseSections(sm, cepp, tableDataSpans), row, col); te.setSrcSpan(ei); result.addTableElement(te); } sm.removeManagedList(tableDataSpans); if (calculateSrcSpans) { result.getSrcSpan().setEnd(-1); } return result; } private NestedListContainer buildNestedList(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans, lineType listType) { boolean numbered = listType == lineType.NESTEDLIST_NR; NestedListContainer result = new NestedListContainer(numbered); if (calculateSrcSpans) { result.setSrcSpan(new SrcSpan(sm.getSrcPos(lineSpans.getFirst().getStart()), -1)); } LinkedList<Span> nestedListSpans = new LinkedList<Span>(); while (!lineSpans.isEmpty()) { Span s = lineSpans.getFirst(); if (listType != getLineType(sm, s)) { break; } nestedListSpans.add(new Span(s.getStart() + 1, s.getEnd()).trim(sm)); lineSpans.removeFirst(); } sm.manageList(nestedListSpans); if (calculateSrcSpans) { result.getSrcSpan().setEnd(sm.getSrcPos(nestedListSpans.getLast().getEnd())); } while (!nestedListSpans.isEmpty()) { Span s = nestedListSpans.getFirst(); lineType t = getLineType(sm, s); if (t == lineType.NESTEDLIST || t == lineType.NESTEDLIST_NR) { result.add(buildNestedList(sm, cepp, nestedListSpans, t)); } else { nestedListSpans.removeFirst(); result.add((NestedListElement) parseContentElement(sm, cepp, s, new NestedListElement())); } } sm.removeManagedList(nestedListSpans); return result; } private DefinitionList buildDefinitionList(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans) { List<ContentElement> content = new ArrayList<ContentElement>(); Span s = lineSpans.removeFirst(); int temp = sm.indexOf(":", s); if (temp == -1) { content.add(parseContentElement(sm, cepp, new Span(s.getStart() + 1, s.getEnd()))); } else { content.add(parseContentElement(sm, cepp, new Span(temp + 1, s.getEnd()))); content.add(0, parseContentElement(sm, cepp, new Span(s.getStart() + 1, temp))); } while (!lineSpans.isEmpty()) { Span ns = lineSpans.getFirst(); if (sm.charAt(ns.getStart()) != ':') { break; } lineSpans.removeFirst(); content.add(parseContentElement(sm, cepp, new Span(ns.getStart() + 1, ns.getEnd()))); } DefinitionList result = new DefinitionList(content); if (calculateSrcSpans) { result.setSrcSpan( new SrcSpan(sm.getSrcPos(s.getStart()), content.get(content.size() - 1).getSrcSpan().getEnd())); } return result; } private Paragraph buildParagraph(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans, lineType paragraphType) { LinkedList<Span> paragraphSpans = new LinkedList<Span>(); Paragraph result = new Paragraph(); Span s = lineSpans.removeFirst(); paragraphSpans.add(s); switch (paragraphType) { case PARAGRAPH: result.setType(Paragraph.type.NORMAL); while (!lineSpans.isEmpty()) { if (paragraphType != getLineType(sm, lineSpans.getFirst())) { break; } paragraphSpans.add(lineSpans.removeFirst()); } break; case PARAGRAPH_BOXED: result.setType(Paragraph.type.BOXED); while (!lineSpans.isEmpty()) { lineType lt = getLineType(sm, lineSpans.getFirst()); if (paragraphType != lt && lineType.EMPTYLINE != lt) { break; } paragraphSpans.add(lineSpans.removeFirst()); } break; case PARAGRAPH_INDENTED: result.setType(Paragraph.type.INDENTED); s.trim(sm.setCharAt(s.getStart(), ' ')); break; default: return null; } parseContentElement(sm, cepp, paragraphSpans, result); return result; } private List<String> tokenize(SpanManager sm, int start, int end, String delim) { List<String> result = new ArrayList<String>(); if (start > end) { logger.debug("tokenize(" + start + ", " + end + ") doesn't make sense"); return result; } int s = start; int e; String token; // Span rs; while ((e = sm.indexOf(delim, s, end)) != -1) { // rs = new Span(s, e).trim( sm ); // if( rs.length()>0 ) result.add( sm.substring( rs ) ); token = sm.substring(s, e).trim(); if (token.length() > 0) { result.add(token); } s = e + delim.length(); } // rs = new Span(s, end).trim( sm ); // if( rs.length()>0 ) result.add( sm.substring( rs ) ); token = sm.substring(s, end).trim(); if (token.length() > 0) { result.add(token); } return result; } private void parseExternalLinks(SpanManager sm, Span s, String protocol, List<Span> managedList, List<Link> links, Content home_cc) { int extLinkTargetStart; Span extLinkSpan = new Span(0, s.getStart()); while ((extLinkTargetStart = sm.indexOf(protocol, extLinkSpan.getEnd(), s.getEnd())) != -1) { // Allowed char before the protocol identifer ? if (extLinkTargetStart > s.getStart() && (" [").indexOf(sm.charAt(extLinkTargetStart - 1)) == -1) { extLinkSpan = new Span(0, extLinkTargetStart + 1); continue; } // Target int extLinkTargetEnd = extLinkTargetStart; while ((lineSeparator + " ]").indexOf(sm.charAt(extLinkTargetEnd)) == -1) { extLinkTargetEnd++; } // Open/Close Tags int extLinkOpenTag = extLinkTargetStart - 1; int extLinkCloseTag; int extLinkTextStart = extLinkTargetStart; int extLinkTextEnd = extLinkTargetEnd; while (extLinkOpenTag >= s.getStart() && sm.charAt(extLinkOpenTag) == ' ') { extLinkOpenTag--; } if (extLinkOpenTag >= s.getStart() && sm.charAt(extLinkOpenTag) == '[') { extLinkCloseTag = sm.indexOf("]", extLinkTargetEnd, s.getEnd()); if (extLinkCloseTag != -1) { extLinkTextStart = extLinkTargetEnd; // nicht wie bei "normalen" links durhc | getrennt sondenr // durhc leerzeichen !!! scheie !!! while (sm.charAt(extLinkTextStart) == ' ') { extLinkTextStart++; } extLinkTextEnd = extLinkCloseTag; extLinkCloseTag++; if (extLinkTextStart == extLinkTextEnd) { sm.insert(extLinkTextStart, "[ ]"); extLinkTextEnd += 3; extLinkCloseTag += 3; } } else { extLinkOpenTag = extLinkTargetStart; extLinkCloseTag = extLinkTargetEnd; } } else { extLinkOpenTag = extLinkTargetStart; extLinkCloseTag = extLinkTargetEnd; } extLinkSpan = new Span(extLinkOpenTag, extLinkCloseTag); managedList.add(extLinkSpan); Link l = new Link(home_cc, extLinkSpan, sm.substring(extLinkTargetStart, extLinkTargetEnd), Link.type.EXTERNAL, null); links.add(l); if (calculateSrcSpans) { l.setSrcSpan(new SrcSpan(sm.getSrcPos(extLinkOpenTag), sm.getSrcPos(extLinkCloseTag - 1) + 1)); } sm.delete(extLinkTextEnd, extLinkCloseTag); sm.delete(extLinkOpenTag, extLinkTextStart); } } /** * Returns the LOWERCASE NameSpace of the link target */ private static String getLinkNameSpace(String target) { int pos = target.indexOf(':'); if (pos == -1) { return null; } else { return target.substring(0, pos).replace('_', ' ').trim().toLowerCase(); } } /** * There is not much differences between links an images, so they are parsed * in a single step */ private void parseImagesAndInternalLinks(SpanManager sm, List<Span> linkSpans, List<Link> links) { sm.manageList(linkSpans); int pos = -1; Stack<Integer> linkOpenTags = new Stack<Integer>(); while ((pos = sm.indexOf("[[", pos + 1)) != -1) { linkOpenTags.push(pos); } Span lastLinkSpan = new Span(sm.length() + 1, sm.length() + 1); Link.type linkType = Link.type.INTERNAL; while (!linkOpenTags.empty()) { int linkStartTag = linkOpenTags.pop(); int linkEndTag = sm.indexOf("]]", linkStartTag); if (linkEndTag == -1) { continue; } int linkOptionTag = sm.indexOf("|", linkStartTag, linkEndTag); int linkTextStart; String linkTarget; if (linkOptionTag != -1) { linkTextStart = linkOptionTag + 1; linkTarget = sm.substring(new Span(linkStartTag + 2, linkOptionTag).trim(sm)); } else { linkTextStart = linkStartTag + 2; linkTarget = sm.substring(new Span(linkStartTag + 2, linkEndTag).trim(sm)); } // is is a regular link ? if (linkTarget.indexOf(lineSeparator) != -1) { continue; } linkTarget = encodeWikistyle(linkTarget); // so it is a Link or image!!! List<String> parameters; String namespace = getLinkNameSpace(linkTarget); if (namespace != null) { if (imageIdentifers.indexOf(namespace) != -1) { if (linkOptionTag != -1) { int temp; while ((temp = sm.indexOf("|", linkTextStart, linkEndTag)) != -1) { linkTextStart = temp + 1; } parameters = tokenize(sm, linkOptionTag + 1, linkEndTag, "|"); // maybe there is an external link at the end of the // image description... if (sm.charAt(linkEndTag + 2) == ']' && sm.indexOf("[", linkTextStart, linkEndTag) != -1) { linkEndTag++; } } else { parameters = null; } linkType = Link.type.IMAGE; } else { //Link has namespace but is not image linkType = Link.type.UNKNOWN; parameters = null; } } else { if (linkType == Link.type.INTERNAL && lastLinkSpan.hits(new Span(linkStartTag, linkEndTag + 2))) { continue; } parameters = null; linkType = Link.type.INTERNAL; } Span posSpan = new Span(linkTextStart, linkEndTag).trim(sm); //Ignoring the blank links from the paragraphs if (linkTarget.length() > 0) linkSpans.add(posSpan); Link l = new Link(null, posSpan, linkTarget, linkType, parameters); //Ignoring the blank links from the paragraphs if (linkTarget.length() > 0) { links.add(l); if (calculateSrcSpans) { l.setSrcSpan(new SrcSpan(sm.getSrcPos(linkStartTag), sm.getSrcPos(linkEndTag + 2))); } } sm.delete(posSpan.getEnd(), linkEndTag + 2); sm.delete(linkStartTag, posSpan.getStart()); // removing line separators in link text int lsinlink; while ((lsinlink = sm.indexOf(lineSeparator, posSpan)) != -1) { sm.replace(lsinlink, lsinlink + lineSeparator.length(), " "); } lastLinkSpan = posSpan; } } /** * Searches the Range given by the Span s for the double occurence of * "quotation" and puts the results in the List quotedSpans. The Quotation * tags will be deleted. * * @param sm * , the Source in which will be searched * @param s * , the range in which will be searched * @param quotedSpans * , the List where the Spans will be placed, should be managed * by the SpanManager sm * @param quotation * , the start and end tag as String */ private void parseQuotedSpans(SpanManager sm, Span s, List<Span> quotedSpans, String quotation) { final int qlen = quotation.length(); // get the start position int start = sm.indexOf(quotation, s.getStart(), s.getEnd()); while (start != -1) { // get the end position int end = sm.indexOf(quotation, start + qlen, s.getEnd()); if (end == -1) { break; } // build a new span from start and end position. Span qs = new Span(start, end); quotedSpans.add(qs); // calculate the original src positions. if (calculateSrcSpans) { qs.setSrcSpan(new SrcSpan(sm.getSrcPos(start), sm.getSrcPos(end + qlen - 1) + 1)); } // delete the tags. sm.delete(end, end + qlen); sm.delete(start, start + qlen); // get the next start position start = sm.indexOf(quotation, qs.getEnd(), s.getEnd()); } } /** * Searches a line for Bold and Italic quotations, this has to be done * linewhise. */ private void parseBoldAndItalicSpans(SpanManager sm, Span line, List<Span> boldSpans, List<Span> italicSpans) { // Das suchen nach BOLD und ITALIC muss in den Jeweiligen // Zeilen geschenhen, da ein LineSeparator immer BOLD und // Italic Tags schliesst. // Bold Spans parseQuotedSpans(sm, line, boldSpans, "'''"); // Italic Spans parseQuotedSpans(sm, line, italicSpans, "''"); // Maybe there is ONE SINGLE OPEN TAG left... handel these... int openTag = sm.indexOf("''", line); if (openTag != -1) { // build a Span from this Tag. Span qs = new Span(openTag, line.getEnd()); // calculate the original src positions. if (calculateSrcSpans) { qs.setSrcSpan(new SrcSpan(sm.getSrcPos(openTag), sm.getSrcPos(line.getEnd()))); } // is it a Bold or an Italic tag ? if (sm.indexOf("'''", openTag, openTag + 3) != -1) { // --> BOLD boldSpans.add(qs); sm.delete(openTag, openTag + 3); } else { // --> ITALIC italicSpans.add(qs); sm.delete(openTag, openTag + 2); } } } private static String encodeWikistyle(String str) { return str.replace(' ', '_'); } /** * Building a ContentElement from a String */ @Override public ContentElement parseContentElement(String src) { SpanManager sm = new SpanManager(src); ContentElementParsingParameters cepp = new ContentElementParsingParameters(); parseImagesAndInternalLinks(sm, cepp.linkSpans, cepp.links); LinkedList<Span> lineSpans = new LinkedList<Span>(); getLineSpans(sm, lineSpans); sm.removeManagedList(lineSpans); return (parseContentElement(sm, cepp, lineSpans, new ContentElement())); } /** * Building a ContentElement from a single line. */ private ContentElement parseContentElement(SpanManager sm, ContentElementParsingParameters cepp, Span lineSpan) { LinkedList<Span> lineSpans = new LinkedList<Span>(); lineSpans.add(lineSpan); return parseContentElement(sm, cepp, lineSpans, new ContentElement()); } /** * Building a ContentElement from a single line. But the result is given, so * e.g. a NestedListElement can be filled with information... */ private ContentElement parseContentElement(SpanManager sm, ContentElementParsingParameters cepp, Span lineSpan, ContentElement result) { LinkedList<Span> lineSpans = new LinkedList<Span>(); lineSpans.add(lineSpan); return parseContentElement(sm, cepp, lineSpans, result); } /** * Building a ContentElement, this funciton is calles by all the other * parseContentElement(..) functions */ private ContentElement parseContentElement(SpanManager sm, ContentElementParsingParameters cepp, LinkedList<Span> lineSpans, ContentElement result) { List<Link> localLinks = new ArrayList<Link>(); List<Template> localTemplates = new ArrayList<Template>(); List<Span> boldSpans = new ArrayList<Span>(); List<Span> italicSpans = new ArrayList<Span>(); sm.manageList(boldSpans); sm.manageList(italicSpans); List<Span> managedSpans = new ArrayList<Span>(); sm.manageList(managedSpans); Span contentElementRange = new Span(lineSpans.getFirst().getStart(), lineSpans.getLast().getEnd()).trim(sm); managedSpans.add(contentElementRange); // set the SrcSpan if (calculateSrcSpans) { result.setSrcSpan(new SrcSpan(sm.getSrcPos(contentElementRange.getStart()), sm.getSrcPos(contentElementRange.getEnd()))); } sm.manageList(lineSpans); while (!lineSpans.isEmpty()) { Span line = lineSpans.getFirst(); parseBoldAndItalicSpans(sm, line, boldSpans, italicSpans); // External links parseExternalLinks(sm, line, "http://", managedSpans, localLinks, result); parseExternalLinks(sm, line, "https://", managedSpans, localLinks, result); parseExternalLinks(sm, line, "ftp://", managedSpans, localLinks, result); parseExternalLinks(sm, line, "mailto:", managedSpans, localLinks, result); // end of linewhise opperations lineSpans.removeFirst(); } sm.removeManagedList(lineSpans); // Links int i; i = 0; while (i < cepp.linkSpans.size()) { if (contentElementRange.hits(cepp.linkSpans.get(i))) { Span linkSpan = cepp.linkSpans.remove(i); managedSpans.add(linkSpan); Link l = cepp.links.remove(i).setHomeElement(result); localLinks.add(l); if (!showImageText && l.getType() == Link.type.IMAGE) { // deletes the Image Text from the ContentElement Text. sm.delete(linkSpan); } } else { i++; } } // Templates //DBPedia - Spotlight. Removing the boiler plate logic from the wikitext //Commenting the Templates Logic /* i = 0; while (i < cepp.templateSpans.size()) { Span ts = cepp.templateSpans.get(i); if (contentElementRange.hits(ts)) { ResolvedTemplate rt = cepp.templates.remove(i); if (rt.getPostParseReplacement() != null) { sm.replace(ts, rt.getPostParseReplacement()); } cepp.templateSpans.remove(i); Object parsedObject = rt.getParsedObject(); if (parsedObject != null) { managedSpans.add(ts); Class parsedObjectClass = parsedObject.getClass(); if (parsedObjectClass == Template.class) { localTemplates.add((Template) parsedObject); } else if (parsedObjectClass == Link.class) { localLinks.add(((Link) parsedObject) .setHomeElement(result)); } else { localTemplates.add(rt.getTemplate()); } } } else { i++; } } */ // HTML/XML Tags i = 0; List<Span> tags = new ArrayList<Span>(); while (i < cepp.tagSpans.size()) { Span s = cepp.tagSpans.get(i); if (contentElementRange.hits(s)) { cepp.tagSpans.remove(i); if (deleteTags) { sm.delete(s); } else { tags.add(s); managedSpans.add(s); } } else { i++; } } // noWiki i = 0; List<Span> localNoWikiSpans = new ArrayList<Span>(); while (i < cepp.noWikiSpans.size()) { Span s = cepp.noWikiSpans.get(i); if (contentElementRange.hits(s)) { cepp.noWikiSpans.remove(i); sm.replace(s, cepp.noWikiStrings.remove(i)); localNoWikiSpans.add(s); managedSpans.add(s); } else { i++; } } // MATH Tags i = 0; List<Span> mathSpans = new ArrayList<Span>(); while (i < cepp.mathSpans.size()) { Span s = cepp.mathSpans.get(i); if (contentElementRange.hits(s)) { cepp.mathSpans.remove(i); if (showMathTagContent) { mathSpans.add(s); managedSpans.add(s); sm.replace(s, cepp.mathStrings.remove(i)); } else { sm.delete(s); } } else { i++; } } result.setText(sm.substring(contentElementRange)); // managed spans must be removed here and not earlier, because every // change in the SpanManager affects the Spans! sm.removeManagedList(boldSpans); sm.removeManagedList(italicSpans); sm.removeManagedList(managedSpans); // contentElementRange ist auch noch in managedSpans !!! deswegen: final int adjust = -contentElementRange.getStart(); for (Span s : boldSpans) { s.adjust(adjust); } for (Span s : italicSpans) { s.adjust(adjust); } for (Span s : managedSpans) { s.adjust(adjust); } result.setFormatSpans(FormatType.BOLD, boldSpans); result.setFormatSpans(FormatType.ITALIC, italicSpans); result.setFormatSpans(FormatType.TAG, tags); result.setFormatSpans(FormatType.MATH, mathSpans); result.setFormatSpans(FormatType.NOWIKI, localNoWikiSpans); result.setLinks(sortLinks(localLinks)); result.setTemplates(sortTemplates(localTemplates)); return result; } /** * Sorts the Links... */ private static List<Link> sortLinks(List<Link> links) { List<Link> result = new ArrayList<Link>(); for (Link l : links) { int pos = 0; while (pos < result.size() && l.getPos().getStart() > result.get(pos).getPos().getStart()) { pos++; } result.add(pos, l); } return result; } /** * Sorts the Templates... */ private static List<Template> sortTemplates(List<Template> templates) { List<Template> result = new ArrayList<Template>(); for (Template t : templates) { int pos = 0; while (pos < result.size() && t.getPos().getStart() > result.get(pos).getPos().getStart()) { pos++; } result.add(pos, t); } return result; } /** * Algorithm to identify the first paragraph of a ParsedPage */ private void setFirstParagraph(ParsedPage pp) { int nr = pp.nrOfParagraphs(); // the paragraph with the lowest number, must not be the first, maybe it // is only an Image... for (int i = 0; i < nr; i++) { Paragraph p = pp.getParagraph(i); // get the Text from the paragraph SpanManager ptext = new SpanManager(p.getText()); List<Span> delete = new ArrayList<Span>(); ptext.manageList(delete); // getting the spans to remove from the text, for templates List<Template> tl = p.getTemplates(); for (int j = tl.size() - 1; j >= 0; j--) { delete.add(tl.get(j).getPos()); } // getting the spans to remove from the text, for Tags List<Span> sl = p.getFormatSpans(FormatType.TAG); for (int j = sl.size() - 1; j >= 0; j--) { delete.add(sl.get(j)); } // getting the spans to remove from the text, for image text if (showImageText) { List<Link> ll = p.getLinks(Link.type.IMAGE); for (int j = ll.size() - 1; j >= 0; j--) { delete.add(ll.get(j).getPos()); } } // delete the spans in reverse order, the spans are managed, so // there is no need to sort them for (int j = delete.size() - 1; j >= 0; j--) { ptext.delete(delete.remove(j)); } // removing line separators if exist, so the result can be trimmed // in the next step int pos = ptext.indexOf(lineSeparator); while (pos != -1) { ptext.delete(pos, pos + lineSeparator.length()); pos = ptext.indexOf(lineSeparator); } // if the result is not an empty string, we got the number of the // first paragraph if (!ptext.toString().trim().equals("")) { pp.setFirstParagraphNr(i); return; } } } /** * Container for all the Parameters needed in the parseing process * * @author CJacobi * */ class ContentElementParsingParameters { List<Span> noWikiSpans; List<String> noWikiStrings; List<Span> linkSpans; List<Link> links; List<Span> templateSpans; List<ResolvedTemplate> templates; List<Span> tagSpans; List<Span> mathSpans; List<String> mathStrings; ContentElementParsingParameters() { noWikiSpans = new ArrayList<Span>(); noWikiStrings = new ArrayList<String>(); linkSpans = new ArrayList<Span>(); links = new ArrayList<Link>(); templateSpans = new ArrayList<Span>(); templates = new ArrayList<ResolvedTemplate>(); tagSpans = new ArrayList<Span>(); mathSpans = new ArrayList<Span>(); mathStrings = new ArrayList<String>(); } } }