/**
* Opens a FileInputStream and detects its encoding. If no encoding is detected, the default
* system encoding will be returned. {@link http://code.google.com/p/juniversalchardet/} {@link
* http://www-archive.mozilla.org/projects/intl/UniversalCharsetDetection.html}
*
* @param filename case-insensitive absolute filename path.
* @return String, the file encoding, it might be null.
* @see FileUtilities#detectFileEncoding(String)
*/
public static String detectFileEncoding(String filename) {
byte[] buf = new byte[4096];
int nread = 0;
int totalReadBytes = 0;
String encoding = null;
FileInputStream fis = null;
try {
fis = new FileInputStream(new CaseInsensitiveFile(filename).toFile());
UniversalDetector detector = new UniversalDetector(null);
// Feed detector with bytes until it detect the encoding or reach the max bytes to read
// if the detector can't get the encoding after reading some bytes, we should stop to save
// time.
while ((nread = fis.read(buf)) > 0
&& !detector.isDone()
&& totalReadBytes < MAX_BTYES_TO_READ) {
totalReadBytes += nread;
detector.handleData(buf, 0, nread);
}
if (!detector.isDone()) {
// IndependantLog.warn("Didn't detect file encoding after reading "+totalReadBytes+"
// bytes.");
}
detector.dataEnd();
// Get the file encoding string (defined in org.mozilla.universalchardet.Constants)
encoding = detector.getDetectedCharset();
detector.reset();
} catch (Exception e) {
IndependantLog.warn(StringUtils.debugmsg(false) + StringUtils.debugmsg(e));
} finally {
if (fis != null)
try {
fis.close();
} catch (Exception e) {
}
}
// If no encoding is detected, test if it is "utf-8"
try {
if (encoding == null || encoding.trim().isEmpty())
if (FileUtilities.isFileUTF8(filename)) encoding = "UTF-8";
} catch (Exception e) {
IndependantLog.warn(StringUtils.debugmsg(false) + StringUtils.debugmsg(e));
}
// If no encoding is detected, get the default file encoding
try {
if (encoding == null || encoding.trim().isEmpty())
encoding = System.getProperty("file.encoding");
if (!Charset.isSupported(encoding)) {
String error =
"The detected encoding is '"
+ encoding
+ "', it is not supported by java.nio.charset.Charset";
IndependantLog.warn(
error); // We need to map org.mozilla.universalchardet.Constants and
// java.nio.charset.Charset ?
throw new IOException(error);
}
} catch (Exception e) {
IndependantLog.warn(StringUtils.debugmsg(false) + StringUtils.debugmsg(e));
}
return encoding;
}
