[jira] [Created] (TIKA-1308) Support in memory parse mode(don't create temp file): to support run Tika in GAE

["yuanyun.cn (JIRA)" <ji...@apache.org>]

yuanyun.cn created TIKA-1308:
--------------------------------

             Summary: Support in memory parse mode(don't create temp file): to support run Tika in GAE
                 Key: TIKA-1308
                 URL: https://issues.apache.org/jira/browse/TIKA-1308
             Project: Tika
          Issue Type: Improvement
          Components: parser
    Affects Versions: 1.5
            Reporter: yuanyun.cn
             Fix For: 1.6


I am trying to use Tika in GAE and write a simple servlet to extract meta data info from jpeg:

String urlStr = req.getParameter("imageUrl");
byte[] oldImageData = IOUtils.toByteArray(new URL(urlStr));

ByteArrayInputStream bais = new ByteArrayInputStream(oldImageData);
Metadata metadata = new Metadata();
BodyContentHandler ch = new BodyContentHandler();
AutoDetectParser parser = new AutoDetectParser();
parser.parse(bais, ch, metadata, new ParseContext());
bais.close();

This fails with exception:
Caused by: java.lang.SecurityException: Unable to create temporary file
	at java.io.File.createTempFile(File.java:1986)
	at org.apache.tika.io.TemporaryResources.createTemporaryFile(TemporaryResources.java:66)
	at org.apache.tika.io.TikaInputStream.getFile(TikaInputStream.java:533)
	at org.apache.tika.parser.jpeg.JpegParser.parse(JpegParser.java:56)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242

Checked the code, in org.apache.tika.parser.jpeg.JpegParser.parse(InputStream, ContentHandler, Metadata, ParseContext), it creates a temp file from the input stream.

I can understand why tika create temp file from the stream: so tika can parse it multiple times.

But as GAE and other cloud servers are getting more popular, is it possible to avoid create temp file: instead we can copy the origin stream to a byteArray stream, so tika can also parse it multiple times.
-- This will have a limit on the file size, as tika keeps the whole file in memory, but this can make tika work in GAE and maybe other cloud server.

We can add a parameter in parser.parse to indicate whether do in memory parse only.
 



--
This message was sent by Atlassian JIRA
(v6.2#6252)