protected String blobsToText(List<Blob> blobs, String docId) {
List<String> strings = new LinkedList<String>();
for (Blob blob : blobs) {
try {
SimpleBlobHolder bh = new SimpleBlobHolder(blob);
BlobHolder result = convert(bh);
if (result == null) {
continue;
}
blob = result.getBlob();
if (blob == null) {
continue;
}
String string = new String(blob.getByteArray(), "UTF-8");
// strip '\0 chars from text
if (string.indexOf('\0') >= 0) {
string = string.replace("\0", " ");
}
strings.add(string);
} catch (Exception e) {
String msg =
"Could not extract fulltext of file '"
+ blob.getFilename()
+ "' for document: "
+ docId
+ ": "
+ e;
log.warn(msg);
log.debug(msg, e);
continue;
}
}
return StringUtils.join(strings, " ");
}
protected static String guessEncoding(Blob blob) throws IOException {
// encoding already known?
if (blob.getEncoding() != null) {
return null;
}
// bad mime type?
String mimeType = blob.getMimeType();
if (mimeType == null) {
return null;
}
if (!mimeType.startsWith("text/") && !mimeType.startsWith("application/xhtml")) {
// not a text file, we shouldn't be in the Note importer
return null;
}
byte[] bytes = blob.getByteArray();
List<String> charsets = Arrays.asList("utf-8", "iso-8859-1");
// charset specified in MIME type?
String CSEQ = "charset=";
int i = mimeType.indexOf(CSEQ);
if (i > 0) {
String onlyMimeType = mimeType.substring(0, i).replace(";", "").trim();
blob.setMimeType(onlyMimeType);
String charset = mimeType.substring(i + CSEQ.length());
i = charset.indexOf(";");
if (i > 0) {
charset = charset.substring(0, i);
}
charset = charset.trim().replace("\"", "");
charsets = new ArrayList<String>(charsets);
charsets.add(0, charset);
}
// resort to auto-detection
for (String charset : charsets) {
try {
Charset cs = Charset.forName(charset);
CharsetDecoder d =
cs.newDecoder()
.onMalformedInput(CodingErrorAction.REPORT)
.onUnmappableCharacter(CodingErrorAction.REPORT);
CharBuffer cb = d.decode(ByteBuffer.wrap(bytes));
return cb.toString();
} catch (IllegalArgumentException e) {
// illegal charset
} catch (CharacterCodingException e) {
// could not decode
}
}
// nothing worked, use platform
return null;
}
/**
* Checks if the {@code inputBlob} string contains a {@code charset} meta tag. If not, add it.
*
* @param inputBlob the input blob
* @throws IOException Signals that an I/O exception has occurred.
*/
protected Blob checkCharsetMeta(Blob inputBlob) throws IOException {
String charset = inputBlob.getEncoding();
if (!StringUtils.isEmpty(charset)) {
Pattern charsetMetaPattern =
Pattern.compile(String.format("content=\"text/html;\\s*charset=%s\"", charset));
Matcher charsetMetaMatcher = charsetMetaPattern.matcher(inputBlob.getString());
if (!charsetMetaMatcher.find()) {
String charsetMetaTag =
String.format(
"<META http-equiv=\"Content-Type\" content=\"text/html; charset=%s\">", charset);
StringBuilder sb = new StringBuilder(charsetMetaTag);
sb.append(new String(inputBlob.getByteArray(), charset));
Blob blobWithCharsetMetaTag =
Blobs.createBlob(sb.toString(), "text/html", charset, inputBlob.getFilename());
return blobWithCharsetMetaTag;
}
}
return inputBlob;
}
