Extracting info from binary scenes

[user_deleted]

Hi,
this is not strictly API question, but it's best place to ask.

I'm using fileInfo commands to store some usefull info into the scene (artist name, character lists and many others).
I'd like to extract that info from scene file in external app (ie. python). This would be easy with ASCII format, but how about binary? It seems, each fileInfo entry begins with "FINF" tag string.
Has anybody experience, or knowledge of structure of .mb files ?

thanks

[joojaa]

cgkit has python library for reading maya chunks use it as a guide. FINF is the name of a file info chunk yes.

[joojaa]

A simple cgkit tool that extracts the values might look like this:

class myReader(cgkit.mayabinary.MBReader):
    def onEndGroup(self, chunk):
        if chunk.type == 'HEAD':
            self.abort()
    def onDataChunk(self,chunk):
        if chunk.tag == 'FINF':
            print chunk.read()
 
myReader().read(r'c:\temp\test.mb')

[user_deleted]

thanks !
cgkit seems great way, but I don't have it installed.
I managed to write it myself, easier that I thought. Here's the code if anybody needs it:

def mfExtractMayaBinInfo(filename):
	fh = open(filename, "rb")
	line = fh.readline()
	fh.close()
 
	RESULT = {}
 
	_index = line.find("FINF")
	while(_index != -1):
		_index += 4
		#read integer size of data
		size = struct.unpack(">i", line[_index:_index + 4])
		size = size[0]
		#read data
		_index += 4
		data = line[_index:_index+size-1]
		data = data.strip()
		tokens = string.split(data, "\x00")
		RESULT[tokens[0]] = string.join(tokens[1:],  " ")
		_index = line.find("FINF", _index+1)
 
	return RESULT

Anyway, good to know that cgkit does it.

[joojaa]

">i"should be ">L"

You can avoid going trough the whole file by reading the length of the HEAD chunk that way you know the data is over by the time you reach that chunks end. So you can skip most of the file irrespectively to how loong the file is. This avoids problems more robustly.

You should also chek that its a maya file. the 16 first shars contain that info.

cgkit seems great way, but I don't have it installed.

well if you can put in your own file then you can put in the only one python file form cgkit you need. You can just take the zzip file and copy over the one python file that contains all you need. But yes i digress the format si pretty simple.

Sincce cgkit is LGPL you can pretty much do what you need in any situation.

But yes i digress the MBReader class leaves something to be desired, so i would reengineer one myself too. In fact i have a windows shell extension wirtten in C to do this so i can list theese things just liek for other windows files.