Netica error or compression error? "buffer to decompress gzip 'internal' is too small"

I'm using the Netica Java library in my code to read in a bayes net, enter findings, and save the updated bayes net in a byte array. The byte array saves perfectly fine (as far as I can tell) but I can't seem to load it back in using the ByteArrayInputStream.

My code is:

Environ env = new EnvironI( null );
ByteArrayInputStream bais = new ByteArrayInputStream( myBinaryData );
bais.close();
Net model = new Net( new Streamer( bais, "templateFile.neta", env );

It fails on the last line with the following error:

stderr:  norsys.netica.NeticaException: 
**************  E R R O R  ***************
** ErrorNumber   = 2734
** ErrorSeverity = ERROR_ERR
** ErrorCategory = 
** ErrorMessage  = In function Net(Streamer inStream) constructor: buffer to decompress gzip 'internal' is too small (is 3082 but -981804177 needed)

Note that there are specific applications of the Netica library here. I should also note that this works fine on my OSX machine but seems to break on a Linux box.

Any ideas?

Solution

I have determined the issue.

I mentioned that I read in a bayes net prior and that code looks like this:

ByteArrayOutputStream baos = new ByteArrayOutputStream();
Streamer stream = new Streamer( baos, "myNetaFile.neta", env );
net.write( stream );
byte[] bytesToEncode = baos.toByteArray();
stream.flush();
String encoded = Base64.encodeBase64String( bytesToEncode );

The above code is actually correct. The binary was getting corrupted somehow when I was reading it back in. My issue was I had the .toByteArray() and .flush() lines swapped.

If anybody else runs into this issue using Netica, keep in mind correct order of operations!