public static List<String> tokenizeString(Analyzer analyzer, String string) {
List<String> result = new ArrayList<String>();
try {
TokenStream stream = analyzer.tokenStream(null, new StringReader(string));
stream.reset();
while (stream.incrementToken()) {
result.add(stream.getAttribute(CharTermAttribute.class).toString());
}
} catch (IOException e) {
// not thrown b/c we're using a string reader...
throw new RuntimeException(e);
}
return result;
}
public static List<String> getText(BytesWritable value, Boolean tokenizep)
throws InterruptedException {
Session s = Session.getDefaultInstance(new Properties());
InputStream is = new ByteArrayInputStream(value.getBytes());
List<String> out = new ArrayList<String>();
try {
MimeMessage message = new MimeMessage(s, is);
message.getAllHeaderLines();
Analyzer standard_analyzer = new StandardAnalyzer(Version.LUCENE_43);
Analyzer email_analyzer = new UAX29URLEmailAnalyzer(Version.LUCENE_43);
Address[] fromAddrs = message.getFrom();
String fromAddrstr = "";
if (fromAddrs != null) {
for (Address addr : fromAddrs) {
fromAddrstr += (addr.toString() + " ");
}
}
Address[] toAddrs = message.getAllRecipients();
String toAddrstr = "";
if (toAddrs != null) {
for (Address addr : toAddrs) {
toAddrstr += (addr.toString() + " ");
}
}
String subject = message.getSubject();
String body = "";
try {
Object content = message.getContent();
// System.err.println(content.getContentType());
if (content instanceof String) {
body = (String) content;
} else if (content instanceof Multipart) {
Multipart mp = (Multipart) content;
for (int i = 0; i < mp.getCount(); i++) {
BodyPart bp = mp.getBodyPart(i);
// System.err.println(bp.getContentType());
Object c = bp.getContent();
if (c instanceof String) {
body = (String) c;
}
}
}
// people do really evil things with email, we're not sorting through it all now
} catch (DecodingException e) {
System.err.println("DecodingException");
} catch (UnsupportedEncodingException e) {
System.err.println("UnsuportedEncodingException");
} catch (IOException e) {
System.err.println("IOException");
}
if (tokenizep) {
List<String> fromData = new ArrayList<String>();
List<String> toData = new ArrayList<String>();
List<String> subjectData = new ArrayList<String>();
List<String> bodyData = new ArrayList<String>();
if (fromAddrstr != null) {
fromData = tokenizeString(email_analyzer, fromAddrstr);
}
if (toAddrstr != null) {
toData = tokenizeString(email_analyzer, toAddrstr);
}
if (subject != null) {
subjectData = tokenizeString(standard_analyzer, subject);
}
if (body != null) {
bodyData = tokenizeString(standard_analyzer, body);
}
out.add("FROM ");
out.addAll(fromData);
out.add("TO ");
out.addAll(toData);
out.add("SUBJECT ");
out.addAll(subjectData);
out.add("BODY ");
out.addAll(bodyData);
} else {
// if not tokenizep, return list with from and subject fields only
out.add(fromAddrstr);
out.add(subject);
}
} catch (MessagingException e) {
System.err.println("MessagineException");
}
return out;
}