public void autoDetectClues(Content content, boolean filter) {
byte[] data = content.getContent();
if (minConfidence >= 0
&& DETECTABLES.contains(content.getContentType())
&& data.length > MIN_LENGTH) {
CharsetMatch[] matches = null;
// do all these in a try/catch; setText and detect/detectAll
// will sometimes throw exceptions
try {
detector.enableInputFilter(filter);
if (data.length > MIN_LENGTH) {
detector.setText(data);
matches = detector.detectAll();
}
} catch (Exception e) {
LOG.debug("Exception from ICU4J (ignoring): ");
e.printStackTrace(LogUtil.getDebugStream(LOG));
}
if (matches != null) {
for (CharsetMatch match : matches) {
addClue(match.getName(), "detect", match.getConfidence());
}
}
}
// add character encoding coming from HTTP response header
addClue(parseCharacterEncoding(content.getMetadata().get(Response.CONTENT_TYPE)), "header");
}
private String getContentCharset(byte[] content, Metadata metadata) {
String charset = null;
// check if the server specified a charset
String specifiedContentType = metadata.getFirstValue(HttpHeaders.CONTENT_TYPE);
try {
if (specifiedContentType != null) {
ContentType parsedContentType = ContentType.parse(specifiedContentType);
charset = parsedContentType.getCharset().name();
}
} catch (Exception e) {
charset = null;
}
// filter HTML tags
CharsetDetector detector = new CharsetDetector();
detector.enableInputFilter(true);
// give it a hint
detector.setDeclaredEncoding(charset);
detector.setText(content);
try {
CharsetMatch charsetMatch = detector.detect();
if (charsetMatch != null) {
charset = charsetMatch.getName();
}
} catch (Exception e) {
// ignore and leave the charset as-is
}
return charset;
}
protected String detectEncoding(InputStream in) throws IOException {
CharsetDetector detector = new CharsetDetector();
detector.setText(in);
return detector.detect().getName();
}
