Example usage for org.apache.commons.lang3 StringEscapeUtils unescapeHtml4

List of usage examples for org.apache.commons.lang3 StringEscapeUtils unescapeHtml4

  1. HOME
  2. Java
  3. org.apache.commons
  4. org.apache.commons.lang3.*
  5. StringEscapeUtils
  6. unescapeHtml4

Introduction

In this page you can find the example usage for org.apache.commons.lang3 StringEscapeUtils unescapeHtml4.

Prototype

public static final String unescapeHtml4(final String input)

Source Link

Document

Unescapes a string containing entity escapes to a string containing the actual Unicode characters corresponding to the escapes.

Usage

From source file:nya.miku.wishmaster.chans.nullchan.AbstractInstant0chan.java

@Override
public SimpleBoardModel[] getBoardsList(ProgressListener listener, CancellableTask task,
        SimpleBoardModel[] oldBoardsList) throws Exception {
    String url = getUsingUrl() + "boards10.json";
    try {/* w w  w .  java  2  s .  c om*/
        JSONArray json = downloadJSONArray(url, oldBoardsList != null, listener, task);
        if (json == null)
            return oldBoardsList;
        List<SimpleBoardModel> list = new ArrayList<>();
        for (int i = 0; i < json.length(); ++i) {
            String currentCategory = json.getJSONObject(i).optString("name");
            JSONArray boards = json.getJSONObject(i).getJSONArray("boards");
            for (int j = 0; j < boards.length(); ++j) {
                SimpleBoardModel model = new SimpleBoardModel();
                model.chan = getChanName();
                model.boardName = boards.getJSONObject(j).getString("dir");
                model.boardDescription = StringEscapeUtils
                        .unescapeHtml4(boards.getJSONObject(j).optString("desc", model.boardName));
                model.boardCategory = currentCategory;
                model.nsfw = model.boardName.equals("b") || currentCategory.equalsIgnoreCase("adult");
                list.add(model);
            }
        }
        return list.toArray(new SimpleBoardModel[list.size()]);
    } catch (JSONException e) {
        return new SimpleBoardModel[0];
    }
}

From source file:nya.miku.wishmaster.chans.nullchan.AbstractInstant0chan.java

private ThreadModel mapCatalogThreadModel(JSONObject json, String boardName) {
    ThreadModel model = new ThreadModel();
    model.threadNumber = json.optString("id", null);
    if (model.threadNumber == null)
        throw new RuntimeException();
    model.postsCount = json.optInt("reply_count", -2) + 1;
    model.attachmentsCount = json.optInt("images", -2) + 1;
    model.isClosed = json.optInt("locked", 0) != 0;
    model.isSticky = json.optInt("stickied", 0) != 0;

    PostModel opPost = new PostModel();
    opPost.number = model.threadNumber;/*ww  w . j  a va 2 s  .  c om*/
    opPost.name = StringEscapeUtils.unescapeHtml4(RegexUtils.removeHtmlSpanTags(json.optString("name")));
    opPost.subject = StringEscapeUtils.unescapeHtml4(json.optString("subject"));
    opPost.comment = json.optString("message");
    opPost.trip = json.optString("tripcode");
    opPost.timestamp = json.optLong("timestamp") * 1000;
    opPost.parentThread = model.threadNumber;

    String ext = json.optString("file_type", "");
    if (!ext.equals("")) {
        AttachmentModel attachment = new AttachmentModel();
        switch (ext) {
        case "jpg":
        case "jpeg":
        case "png":
            attachment.type = AttachmentModel.TYPE_IMAGE_STATIC;
            break;
        case "gif":
            attachment.type = AttachmentModel.TYPE_IMAGE_GIF;
            break;
        case "mp3":
        case "ogg":
            attachment.type = AttachmentModel.TYPE_AUDIO;
            break;
        case "webm":
        case "mp4":
            attachment.type = AttachmentModel.TYPE_VIDEO;
            break;
        case "you":
            attachment.type = AttachmentModel.TYPE_OTHER_NOTFILE;
            break;
        default:
            attachment.type = AttachmentModel.TYPE_OTHER_FILE;
        }
        attachment.width = json.optInt("image_w", -1);
        attachment.height = json.optInt("image_h", -1);
        attachment.size = -1;
        String fileName = json.optString("file", "");
        if (!fileName.equals("")) {
            if (ext.equals("you")) {
                attachment.thumbnail = (useHttps() ? "https" : "http") + "://img.youtube.com/vi/" + fileName
                        + "/default.jpg";
                attachment.path = (useHttps() ? "https" : "http") + "://youtube.com/watch?v=" + fileName;
            } else {
                attachment.thumbnail = "/" + boardName + "/thumb/" + fileName + "s." + ext;
                attachment.path = "/" + boardName + "/src/" + fileName + "." + ext;
            }
            opPost.attachments = new AttachmentModel[] { attachment };
        }
    }
    model.posts = new PostModel[] { opPost };
    return model;
}

From source file:opengovcrawler.DB.java

public static String EscapeHtml(String htmlText) {
    String escapedText = StringEscapeUtils.unescapeHtml4(htmlText);//escapeHtml(htmlText);
    return escapedText;
}

From source file:org.andstatus.app.msg.MessageViewItem.java

@NonNull
private String getCleanedBody(String body) {
    String out = MyHtml.fromHtml(body).toLowerCase();
    out = StringEscapeUtils.unescapeHtml4(out);
    return out.replaceAll("\n", " ").replaceAll("  ", " ").replaceFirst(".*(favorited something by.*)", "$1");
}

From source file:org.apache.jmeter.extractor.RegexExtractor.java

private String getInputString(SampleResult result) {
    String inputString = useUrl() ? result.getUrlAsString() // Bug 39707
            : useHeaders() ? result.getResponseHeaders()
                    : useRequestHeaders() ? result.getRequestHeaders()
                            : useCode() ? result.getResponseCode() // Bug 43451
                                    : useMessage() ? result.getResponseMessage() // Bug 43451
                                            : useUnescapedBody()
                                                    ? StringEscapeUtils
                                                            .unescapeHtml4(result.getResponseDataAsString())
                                                    : useBodyAsDocument()
                                                            ? Document.getTextFromDocument(
                                                                    result.getResponseData())
                                                            : result.getResponseDataAsString() // Bug 36898
    ;/*w w  w  . j  a  v  a  2  s  . co  m*/
    if (log.isDebugEnabled()) {
        log.debug("Input = " + inputString);
    }
    return inputString;
}

From source file:org.apache.jmeter.functions.UnEscapeHtml.java

/** {@inheritDoc} */
@Override/*from w  ww . j  a  v  a 2s .com*/
public String execute(SampleResult previousResult, Sampler currentSampler) throws InvalidVariableException {

    String escapedString = ((CompoundVariable) values[0]).execute();
    return StringEscapeUtils.unescapeHtml4(escapedString);

}

From source file:org.apache.lucene.analysis.kr.test.KoreanAnalyzerTest.java

public void testJavaEscape() throws Exception {

    String str = StringEscapeUtils.unescapeHtml4("&#48085;");
    System.out.println(str);/*from   w  w  w .  j a v a 2 s .  c  o  m*/

    //??
    String han = StringEscapeUtils.unescapeJava("0x3400");
    han = StringEscapeUtils.escapeJava("?");

    System.out.println(han);

}

From source file:org.apache.mahout.text.wikipedia.WikipediaDatasetCreatorMapper.java

@Override
protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
    String document = value.toString();
    document = StringEscapeUtils.unescapeHtml4(CLOSE_TEXT_TAG_PATTERN
            .matcher(OPEN_TEXT_TAG_PATTERN.matcher(document).replaceFirst("")).replaceAll(""));
    String catMatch = findMatchingCategory(document);
    if (!"Unknown".equals(catMatch)) {
        StringBuilder contents = new StringBuilder(1000);
        TokenStream stream = analyzer.tokenStream(catMatch, new StringReader(document));
        CharTermAttribute termAtt = stream.addAttribute(CharTermAttribute.class);
        stream.reset();/*from w ww .j  a v a 2  s  .com*/
        while (stream.incrementToken()) {
            contents.append(termAtt.buffer(), 0, termAtt.length()).append(' ');
        }
        context.write(new Text(SPACE_NON_ALPHA_PATTERN.matcher(catMatch).replaceAll("_")),
                new Text(contents.toString()));
        stream.end();
        Closeables.close(stream, true);
    }
}

From source file:org.apache.mahout.text.wikipedia.WikipediaMapper.java

@Override
protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

    String content = value.toString();
    if (content.contains(REDIRECT)) {
        return;/*from  w w  w  .j  ava  2s  .  c o m*/
    }
    String document;
    String title;
    try {
        document = getDocument(content);
        title = getTitle(content);
    } catch (RuntimeException e) {
        // TODO: reporter.getCounter("Wikipedia", "Parse errors").increment(1);
        return;
    }

    String catMatch = findMatchingCategory(document);
    if (!all) {
        if ("Unknown".equals(catMatch)) {
            return;
        }
    }

    document = StringEscapeUtils.unescapeHtml4(document);
    if (removeLabels) {
        document = removeCategoriesFromText(document);
        // Reject documents with malformed tags
        if (document == null) {
            return;
        }
    }

    // write out in Bayes input style: key: /Category/document_name
    String category = "/" + catMatch.toLowerCase(Locale.ENGLISH) + "/"
            + SPACE_NON_ALPHA_PATTERN.matcher(title).replaceAll("_");

    context.write(new Text(category), new Text(document));
}

From source file:org.apache.tika.parser.pdf.EnhancedPDFParser.java

private String getMatchGroup(Matcher m, int group) {
    return StringEscapeUtils.unescapeHtml4(StringUtils.trimToEmpty(m.group(group)));
}