@Override
public void characters(char[] ch, int start, int length) throws SAXException {
if (inCell) {
StringBuffer t = new StringBuffer(new String(ch, start, length));
// Quote if not all numbers
if (all_nums.matcher(t).matches()) {
super.characters(ch, start, length);
} else {
for (int i = t.length() - 1; i >= 0; i--) {
if (t.charAt(i) == '\"') {
// Double up double quotes
t.insert(i, '\"');
i--;
}
}
t.insert(0, '\"');
t.append('\"');
char[] c = t.toString().toCharArray();
super.characters(c, 0, c.length);
}
} else {
super.characters(ch, start, length);
}
}
@Override
public void startElement(String uri, String localName, String name, Attributes atts)
throws SAXException {
if (localName.equals("td")) {
inCell = true;
if (needsComma) {
super.characters(comma, 0, 1);
needsComma = true;
}
} else {
super.startElement(uri, localName, name, atts);
}
}
@Override
public void ignorableWhitespace(char[] ch, int start, int length) throws SAXException {
if (length == 1 && ch[0] == '\t') {
// Ignore tabs, as they mess up the CSV output
} else {
super.ignorableWhitespace(ch, start, length);
}
}
@Override
public void endElement(String uri, String localName, String name) throws SAXException {
if (localName.equals("td")) {
needsComma = true;
inCell = false;
} else {
if (localName.equals("tr")) {
needsComma = false;
}
super.endElement(uri, localName, name);
}
}
/**
* Common implementation -- take an input stream and return a ConvertedDoc;
*
* @param input stream for raw file
* @param doc raw file
* @return converted doc
* @throws IOException if underlying Tika parser/writer had an IO problem, an parser problem, or
* MAX_TEXT_SIZE is reached.
*/
@Override
protected ConvertedDocument conversionImplementation(InputStream input, java.io.File doc)
throws IOException {
Metadata metadata = new Metadata();
BodyContentHandler handler = new BodyContentHandler(maxBuffer);
try {
parser.parse(input, handler, metadata, ctx);
} catch (NoClassDefFoundError classErr) {
throw new IOException("Unable to parse content due to Tika misconfiguration", classErr);
} catch (Exception xerr) {
throw new IOException("Unable to parse content", xerr);
} finally {
input.close();
}
ConvertedDocument textdoc = new ConvertedDocument(doc);
textdoc.addTitle(metadata.get(TikaCoreProperties.TITLE));
textdoc.setEncoding(metadata.get(Metadata.CONTENT_ENCODING));
textdoc.addCreateDate(metadata.getDate(TikaCoreProperties.CREATED));
textdoc.addAuthor(metadata.get(TikaCoreProperties.CREATOR));
// v1.5: until this version this blank line reducer was in place.
// Using Java6 it appeared to cause StackOverflow when it encountered a document hundreds of
// \n in a row.
// Eg.., a Spreadsheet doc converted to text may have thousands of empty lines following the
// last data row.
// TextUtils.reduce_line_breaks(txt)
String t = handler.toString();
if (t != null) {
if (textdoc.filename != null && FileUtility.isSpreadsheet(textdoc.filename)) {
textdoc.setText(t.trim());
} else {
textdoc.setText(TextUtils.reduce_line_breaks(t));
}
}
return textdoc;
}
