Search in sources :

Example 1 with UnstructuredField

use of org.apache.james.mime4j.dom.field.UnstructuredField in project tika by apache.

the class MailContentHandler method field.

/**
     * Header for the whole message or its parts
     *
     * @see <a href="http://james.apache.org/mime4j/apidocs/org/apache/james/mime4j/parser/">
     *     http://james.apache.org/mime4j/apidocs/org/apache/james/mime4j/parser/</a>
     * Field.html
     */
public void field(Field field) throws MimeException {
    // whole message or its parts
    if (inPart) {
        return;
    }
    try {
        String fieldname = field.getName();
        ParsedField parsedField = LenientFieldParser.getParser().parse(field, DecodeMonitor.SILENT);
        if (fieldname.equalsIgnoreCase("From")) {
            MailboxListField fromField = (MailboxListField) parsedField;
            MailboxList mailboxList = fromField.getMailboxList();
            if (fromField.isValidField() && mailboxList != null) {
                for (Address address : mailboxList) {
                    String from = getDisplayString(address);
                    MailUtil.setPersonAndEmail(from, Message.MESSAGE_FROM_NAME, Message.MESSAGE_FROM_EMAIL, metadata);
                    metadata.add(Metadata.MESSAGE_FROM, from);
                    metadata.add(TikaCoreProperties.CREATOR, from);
                }
            } else {
                String from = stripOutFieldPrefix(field, "From:");
                MailUtil.setPersonAndEmail(from, Message.MESSAGE_FROM_NAME, Message.MESSAGE_FROM_EMAIL, metadata);
                if (from.startsWith("<")) {
                    from = from.substring(1);
                }
                if (from.endsWith(">")) {
                    from = from.substring(0, from.length() - 1);
                }
                metadata.add(Metadata.MESSAGE_FROM, from);
                metadata.add(TikaCoreProperties.CREATOR, from);
            }
        } else if (fieldname.equalsIgnoreCase("Subject")) {
            metadata.set(TikaCoreProperties.TRANSITION_SUBJECT_TO_DC_TITLE, ((UnstructuredField) parsedField).getValue());
        } else if (fieldname.equalsIgnoreCase("To")) {
            processAddressList(parsedField, "To:", Metadata.MESSAGE_TO);
        } else if (fieldname.equalsIgnoreCase("CC")) {
            processAddressList(parsedField, "Cc:", Metadata.MESSAGE_CC);
        } else if (fieldname.equalsIgnoreCase("BCC")) {
            processAddressList(parsedField, "Bcc:", Metadata.MESSAGE_BCC);
        } else if (fieldname.equalsIgnoreCase("Date")) {
            DateTimeField dateField = (DateTimeField) parsedField;
            Date date = dateField.getDate();
            if (date == null) {
                date = tryOtherDateFormats(field.getBody());
            }
            metadata.set(TikaCoreProperties.CREATED, date);
        } else {
            metadata.add(Metadata.MESSAGE_RAW_HEADER_PREFIX + parsedField.getName(), field.getBody());
        }
    } catch (RuntimeException me) {
        if (strictParsing) {
            throw me;
        }
    }
}
Also used : Address(org.apache.james.mime4j.dom.address.Address) ParsedField(org.apache.james.mime4j.dom.field.ParsedField) MailboxListField(org.apache.james.mime4j.dom.field.MailboxListField) UnstructuredField(org.apache.james.mime4j.dom.field.UnstructuredField) MailboxList(org.apache.james.mime4j.dom.address.MailboxList) DateTimeField(org.apache.james.mime4j.dom.field.DateTimeField) Date(java.util.Date)

Aggregations

Date (java.util.Date)1 Address (org.apache.james.mime4j.dom.address.Address)1 MailboxList (org.apache.james.mime4j.dom.address.MailboxList)1 DateTimeField (org.apache.james.mime4j.dom.field.DateTimeField)1 MailboxListField (org.apache.james.mime4j.dom.field.MailboxListField)1 ParsedField (org.apache.james.mime4j.dom.field.ParsedField)1 UnstructuredField (org.apache.james.mime4j.dom.field.UnstructuredField)1