Example usage for java.text Normalizer normalize

List of usage examples for java.text Normalizer normalize

  1. HOME
  2. Java
  3. java
  4. java.text.*
  5. Normalizer
  6. normalize

Introduction

In this page you can find the example usage for java.text Normalizer normalize.

Prototype

public static String normalize(CharSequence src, Form form)

Source Link

Document

Normalize a sequence of char values.

Usage

From source file:cn.org.once.cstack.model.Snapshot.java

public void setApplicationName(String applicationName) {
    applicationName = applicationName.toLowerCase();
    applicationName = Normalizer.normalize(applicationName, Normalizer.Form.NFD);
    applicationName = applicationName.replaceAll("[\\p{InCombiningDiacriticalMarks}]", "");
    this.applicationName = applicationName.replaceAll("[^a-z0-9]", "");
}

From source file:tr.edu.gsu.nerwip.retrieval.reader.wikipedia.WikipediaReader.java

@Override
public String getName(URL url) {
    String address = url.toString();

    // get the last part of the URL as the page name
    String temp[] = address.split("/");
    String result = temp[temp.length - 1];

    // remove diacritics
    result = Normalizer.normalize(result, Form.NFD).replaceAll("\\p{InCombiningDiacriticalMarks}+", "");

    return result;
}

From source file:fr.insalyon.creatis.vip.datamanager.server.rpc.FileUploadServiceImpl.java

@Override
protected void doPost(HttpServletRequest request, HttpServletResponse response)
        throws ServletException, IOException {

    User user = (User) request.getSession().getAttribute(CoreConstants.SESSION_USER);
    if (user != null && ServletFileUpload.isMultipartContent(request)) {

        FileItemFactory factory = new DiskFileItemFactory();
        ServletFileUpload upload = new ServletFileUpload(factory);
        try {//from  ww w. j a va 2  s  .c  om
            List items = upload.parseRequest(request);
            Iterator iter = items.iterator();
            String fileName = null;
            FileItem fileItem = null;
            String path = null;
            String target = "uploadComplete";
            String operationID = "no-id";

            while (iter.hasNext()) {
                FileItem item = (FileItem) iter.next();

                if (item.getFieldName().equals("path")) {
                    path = item.getString();
                } else if (item.getFieldName().equals("file")) {
                    fileName = item.getName();
                    fileItem = item;
                } else if (item.getFieldName().equals("target")) {
                    target = item.getString();
                }
            }
            if (fileName != null && !fileName.equals("")) {

                boolean local = path.equals("local") ? true : false;
                String rootDirectory = DataManagerUtil.getUploadRootDirectory(local);
                fileName = new File(fileName).getName().trim().replaceAll(" ", "_");
                fileName = Normalizer.normalize(fileName, Normalizer.Form.NFD)
                        .replaceAll("\\p{InCombiningDiacriticalMarks}+", "");
                File uploadedFile = new File(rootDirectory + fileName);

                try {
                    fileItem.write(uploadedFile);
                    response.getWriter().write(fileName);

                    if (!local) {
                        // GRIDA Pool Client
                        logger.info("(" + user.getEmail() + ") Uploading '" + uploadedFile.getAbsolutePath()
                                + "' to '" + path + "'.");
                        GRIDAPoolClient client = CoreUtil.getGRIDAPoolClient();
                        operationID = client.uploadFile(uploadedFile.getAbsolutePath(),
                                DataManagerUtil.parseBaseDir(user, path), user.getEmail());

                    } else {
                        operationID = fileName;
                        logger.info(
                                "(" + user.getEmail() + ") Uploaded '" + uploadedFile.getAbsolutePath() + "'.");
                    }
                } catch (Exception ex) {
                    logger.error(ex);
                }
            }

            response.setContentType("text/html");
            response.setHeader("Pragma", "No-cache");
            response.setDateHeader("Expires", 0);
            response.setHeader("Cache-Control", "no-cache");
            PrintWriter out = response.getWriter();
            out.println("<html>");
            out.println("<body>");
            out.println("<script type=\"text/javascript\">");
            out.println("if (parent." + target + ") parent." + target + "('" + operationID + "');");
            out.println("</script>");
            out.println("</body>");
            out.println("</html>");
            out.flush();

        } catch (FileUploadException ex) {
            logger.error(ex);
        }
    }
}

From source file:org.haedus.datatypes.phonetic.FeatureModel.java

public String getBestSymbol(List<Double> featureArray) {

    List<Double> bestFeatures = new ArrayList<Double>();
    String bestSymbol = "";
    double minimum = Double.MAX_VALUE;

    for (Map.Entry<String, List<Double>> entry : featureMap.entrySet()) {
        List<Double> features = entry.getValue();

        double difference = getDifferenceValue(featureArray, features);
        if (difference < minimum) {
            bestSymbol = entry.getKey();
            minimum = difference;//w  w w  .ja  v  a2 s  .  c  o  m
            bestFeatures = features;
        }
    }

    String bestDiacritic = "";
    if (minimum > 0.0) {
        bestDiacritic = getBestDiacritic(featureArray, bestFeatures);
    }
    return Normalizer.normalize(bestSymbol + bestDiacritic, Normalizer.Form.NFC);
}

From source file:pl.edu.icm.cermine.metadata.model.DocumentAffiliation.java

public String getOrganization() {
    for (Token<AffiliationLabel> token : tokens) {
        if (token.getLabel().equals(AffiliationLabel.INST)) {
            return Normalizer.normalize(token.getText(), Normalizer.Form.NFC);
        }//from w w  w .ja  va2 s.co  m
    }
    return null;
}

From source file:org.sonar.batch.scm.DefaultBlameOutput.java

private static String removeAccents(String inputString) {
    String unicodeDecomposedString = Normalizer.normalize(inputString, Normalizer.Form.NFD);
    return ACCENT_CODES.matcher(unicodeDecomposedString).replaceAll("");
}

From source file:com.continusec.client.ObjectHash.java

private static final byte[] hashString(String s, String r) throws ContinusecException {
    if (r != null && s.startsWith(r)) {
        try {/*from   w w w. j  a  va2s . co m*/
            return Hex.decodeHex(s.substring(r.length()).toCharArray());
        } catch (DecoderException e) {
            throw new InvalidObjectException(e);
        }
    } else {
        try {
            MessageDigest d = DigestUtils.getSha256Digest();
            d.update((byte) 'u');
            d.update(Normalizer.normalize(s, Normalizer.Form.NFC).getBytes("UTF8"));
            return d.digest();
        } catch (UnsupportedEncodingException e) {
            throw new InvalidObjectException(e);
        }
    }
}

From source file:biblivre3.z3950.Z3950Client.java

public List<Record> doSearch(final Z3950ServerDTO server, final Z3950SearchDTO search) {
    List<Record> listRecords = new ArrayList<Record>();

    factory.setHost(server.getUrl());//from w  ww . j av  a2s  .co  m
    factory.setPort(server.getPort());
    factory.setCharsetEncoding("UTF-8");

    factory.setApplicationContext(z3950Context);
    factory.setDefaultRecordSyntax("usmarc");
    factory.setDefaultElementSetName("F");

    factory.setDoCharsetNeg(true);

    factory.getRecordArchetypes().put("Default", "usmarc::F");
    factory.getRecordArchetypes().put("FullDisplay", "usmarc::F");
    factory.getRecordArchetypes().put("BriefDisplay", "usmarc::B");

    final String qry = QUERY_PREFIX + search.getType() + " \"" + TextUtils.removeDiacriticals(search.getValue())
            + "\"";

    IRQuery query = new IRQuery();
    query.collections = new Vector();
    query.collections.add(server.getCollection());
    query.query = new org.jzkit.search.util.QueryModel.PrefixString.PrefixString(qry);

    try {
        Searchable s = factory.newSearchable();
        s.setApplicationContext(z3950Context);
        IRResultSet result = s.evaluate(query);

        // Wait without timeout until result set is complete or failure
        result.waitForStatus(IRResultSetStatus.COMPLETE | IRResultSetStatus.FAILURE, 0);
        if (result.getStatus() == IRResultSetStatus.FAILURE) {
            log.error("IRResultSetStatus == FAILURE");
        }
        if (result.getFragmentCount() == 0) {
            return listRecords;

        }

        String encoding = server.getCharset();
        AnselToUnicode atu = new AnselToUnicode();

        Enumeration e = new ReadAheadEnumeration(result, new ArchetypeRecordFormatSpecification("Default"));
        int errorRecords = 0;
        Record record = null;
        for (int i = 0; e.hasMoreElements(); i++) {
            iso2709 o = (iso2709) e.nextElement();

            try {
                String iso = "";
                if (encoding.equals("MARC-8")) {
                    iso = Normalizer.normalize(
                            atu.convert(new String((byte[]) o.getOriginalObject(), "ISO-8859-1")),
                            Normalizer.Form.NFC);
                } else {
                    iso = new String((byte[]) o.getOriginalObject(), encoding);
                }

                try {
                    record = MarcUtils.iso2709ToRecordAsIso(iso, false);
                } catch (Exception encodeE) {
                }

                if (record == null) {
                    try {
                        record = MarcUtils.iso2709ToRecord(iso, false);
                    } catch (Exception encodeE) {
                    }
                }

                if (record == null) {
                    try {
                        record = MarcUtils.iso2709ToRecordAsIso(
                                new String((byte[]) o.getOriginalObject(), "ISO-8859-1"), false);
                    } catch (Exception encodeE) {
                    }
                }

                if (record == null) {
                    try {
                        record = MarcUtils.iso2709ToRecord(
                                new String((byte[]) o.getOriginalObject(), "ISO-8859-1"), false);
                    } catch (Exception encodeE) {
                    }
                }
            } catch (Exception ex) {
            }

            if (record != null) {
                listRecords.add(record);
            } else {
                ++errorRecords;
            }
        }
        if (errorRecords > 0) {
            log.warn("Total number of records that failed the conversion: " + errorRecords);
        }
        try {
            result.close();
            s.close();
        } catch (Exception closingException) {
            log.error(closingException.getMessage());
        }
    } catch (Exception e) {
        log.error(e.getMessage(), e);
    }
    log.info("returning results");
    return listRecords;
}

From source file:com.silverpeas.tags.navigation.links.CachedLinkGeneratorImpl.java

/**
 * Suppresion des accents d'une chaine de caractres.
 * @param s/*from ww w. j ava  2 s. c om*/
 * @return
 */
public String stripAccents(String s) {
    s = Normalizer.normalize(s, Normalizer.Form.NFD);
    s = s.replaceAll("\\p{InCombiningDiacriticalMarks}+", "");
    return s;
}

From source file:fr.paris.lutece.plugins.profanityfilter.service.ProfanityFilter.java

public static String removeAccent(String source) {
    return Normalizer.normalize(source, Normalizer.Form.NFD).replaceAll("[\u0300-\u036F]", "");
}