Application: gvSIG desktop

package org.gvsig.fmap.dal.fileutils.impl;

import java.io.File;

import java.io.IOException;

import org.apache.commons.io.FileUtils;

import org.apache.commons.io.FilenameUtils;

import org.apache.commons.lang3.StringUtils;

import org.slf4j.Logger;

import org.slf4j.LoggerFactory;

import org.gvsig.fmap.dal.fileutils.CPGFile;

public class DefaultCPGFile implements CPGFile {

    private static final Logger logger = LoggerFactory.getLogger(DefaultPRJFile.class);

    private File source;

    private String charsetName = null;

    /**

     * Define the valid code pages (equivalent to MSDOS code pages). 

     * This codes are used on the byte 29 of the DBF header to define the DBF

     * codepage.

     *

     * The equivalences of these charsets using Java NIO charset names are

     * defined on the {@link #charsetNames} array (so 0x01 is

     * equivalent to IBM437, 0x02 to IBM850, etc)

     *

     * See some other equivalences in:

     * https://github.com/infused/dbf/blob/master/docs/supported_encodings.csv

     * https://github.com/olemb/dbfread/blob/master/dbfread/codepages.py

     * https://joinup.ec.europa.eu/svn/gvsig-desktop/trunk/libraries/libFMap/src/com/iver/cit/gvsig/fmap/drivers/dbf/DbfEncodings.java

     */

    private static final short[] codePages = {

        0x01, 0x02, 0x03, 0x04,

        0x08, 0x09, 0x0a, 0x0b,

        0x0d, 0x0e, 0x0f, 0x10,

        0x11, 0x12, 0x13, 0x14,

        0x15, 0x16, 0x17, 0x18,

        0x19, 0x1a, 0x1b, 0x1c,

        0x1d, 0x1f, 0x22, 0x23,

        0x24, 0x25, 0x26, 0x37,

        0x40, 0x4d, 0x4e, 0x4f,

        0x50, 0x57, 0x58, 0x59,

        0x64, 0x65, 0x66, 0x67,

        0x68, 0x69, 0x6a, 0x6b,

        0x6c, 0x78, 0x79, 0x7a,

        0x7b, 0x7c, 0x7d, 0x7d,

        0x86, 0x87, 0x88, 0xc8,

        0xc9, 0xca, 0xcb, 0xcc};

    /**

     * Equivalent Java charset names to the code pages defined in

     * {@link #codePages}, using Java NIO Charset names (which differ

     * from JAVA IO names, see

     * https://docs.oracle.com/javase/8/docs/technotes/guides/intl/encoding.doc.html)

     */

    private static final String[] charsetNames = new String[]{

        "IBM437", "IBM850", "windows-1252", "x-MacRoman",

        "IBM865", "IBM437", "IBM850", "IBM437",

        "IBM437", "IBM850", "IBM437", "IBM850",

        "IBM437", "IBM850", "x-IBM943", "IBM850",

        "IBM437", "IBM850", "IBM865", "IBM437",

        "IBM437", "IBM850", "IBM437", "IBM863",

        "IBM850", "IBM852", "IBM852", "IBM852",

        "IBM860", "IBM850", "IBM866", "IBM850",

        "IBM852", "x-mswin-936", "x-IBM949", "IBM950",

        "x-IBM874", "windows-1252", "windows-1252", "windows-1252",

        "IBM852", "IBM866", "IBM865", "IBM861",

        // 0x68 and 0x69 are unofficial "Codepage 895 Kamenicky (Czech) MS-DOS" and "Codepage 620  Mazovia (Polish) MS-DOS",

        // but there is no Java equivalent

        // so we use CP437 which is the closest charset for the latin characters part

        "IBM437", "IBM437", "x-IBM737", "IBM857",

        "IBM863", "x-IBM950", "x-IBM949", "x-mswin-936",

        "x-IBM942", "x-IBM874", "windows-1255", "windows-1256",

        "x-IBM737", "IBM852", "IBM857", "windows-1250",

        "windows-1251", "windows-1254", "windows-1253", "windows-1257"};

    public DefaultCPGFile() {

        this.charsetName = null;

        this.source = null;

    }

    @SuppressWarnings("OverridableMethodCallInConstructor")

    public DefaultCPGFile(File file) {

        this();

        read(file);

    }

    @Override

    public File getFile(File file) {

        File f = new File(FilenameUtils.removeExtension(file.getAbsolutePath()) + "." + FILE_EXTENSION);

        return f;

    }

    @Override

    public File getFile() {

        return source;

    }

    @Override

    public String getCharsetName() {

        return this.charsetName;

    }

    @Override

    public void setCharsetName(String charsetName) {

        this.charsetName = charsetName;

    }

    @Override

    public String toCharsetName(String codePageName) {

        if (codePageName.equals("UTF8")) {

            return "UTF-8";

        }

        if (codePageName.equals("SJIS")) {

            return "Shift_JIS";

        }

        if (StringUtils.isNumeric(codePageName)) {

            if (codePageName.startsWith("8859") && codePageName.length() > 4) {

                return "ISO-8859-" + codePageName.substring(4);

            }

            if (codePageName.startsWith("125") && codePageName.length() == 4) {

                return "windows-" + codePageName;

            }

            if (codePageName.length() == 3) {

                return "IBM-" + codePageName;

            }

            for (int i = 0; i < charsetNames.length; i++) {

                if (charsetNames[i].contains(codePageName)) {

                    return codePageName;

                }

            }

        }

        if (codePageName.equals("65001")) {

            return "UTF-8";

        }

        return codePageName;

    }

    /**

     * Gets the Java NIO charset name equivalent to the provided code page.

     * Gets null if the provided code page is not recognised

     * as a valid code

     *

     * @param codePage

     * @return

     */

    @Override

    public String toCharsetName(int codePage) {

        if (codePage != 0) {

            for (int i = 0; i < codePages.length; i++) {

                if (codePages[i] == codePage) {

                    return charsetNames[i];

                }

            }

        }

        return null;

    }

    @Override

    public String toCPGName(String charsetName) {

        if (charsetName.startsWith("windows-")

                || charsetName.startsWith("ISO-8859")

                || charsetName.startsWith("IBM-")

                || charsetName.startsWith("x-IBM")

                || charsetName.startsWith("x-mswin-")) {

            return charsetName.replaceAll("[^\\d]", "");

        }

        if (charsetName.equals("Shift_JIS")) {

            return "SJIS";

        }

        // For the rest of the charsets, we'll directly write the Java NIO Charset

        // Probably they will only be recognized by gvSIG, but it's better than nothing

        return charsetName;

    }

    /**

     * Returns the code page corresponding to the

     * provided charset name

     *

     * @param charsetName

     * @return The code page, or 0x00 if no equivalent code page was found for

     * the provided charsetName

     */

    @Override

    public int toCPG(String charsetName) {

        for (int i = 0; i < charsetNames.length; i++) {

            if (charsetNames[i].equals(charsetName)) {

                return codePages[i];

            }

        }

        // default

        return 0x00;

    }

    @Override

    public void read(File file) {

        File f = this.getFile(file);

        if (f.exists()) {

            try {

                String theContents = FileUtils.readFileToString(f);

                theContents = StringUtils.trim(theContents);

                if (StringUtils.isNotEmpty(theContents)) {

                    String theCharset = toCharsetName(theContents);

                    this.charsetName = theCharset;

                    this.source = f.getAbsoluteFile();

                }

            } catch (IOException e) {

                logger.warn("Couldn't read " + FILE_EXTENSION + " file (" + f.getAbsolutePath() + ").", e);

            }

        }

    }

    @Override

    public void write(File file) throws IOException {

        File f = this.getFile(file);

        try {

            String export = toCPGName(this.charsetName) + "\n";

            FileUtils.writeStringToFile(f, export, "ISO-8859-1");

            this.source = f;

        } catch (Exception e) {

            logger.warn("Couldn't write " + FILE_EXTENSION + " file (" + f.getAbsolutePath() + ").", e);

            throw e;

        }

    }

}