#3 Improve OPJ files import

[Anonymous]

* Use native byteorder (=) instead of little-endian (<) for unpack expressions.
Should handle little-endian/big-endian files transparently (not tested).

* Added creation of Project Folders and Matrix windows.
Added description of Graphs as Note windows.

NOTE: Import of Matrices is slow mainly because elements have to be
converted one by one. SciDAVis Matrix Python API should have a
replaceValues() method to convert the whole Matrix at once.

To convert multiple files at once, a Python API to delete a SciDAVis
project tree is needed.

[Anonymous]

Add project tree and matrix windows

[Knut Franke]

I think assuming little-endian byte order was correct, since Origin runs only on Windows and Intel Macs[1], so there should not be any big-endian OPJ files. On the other hand, when trying to import (little-endian) OPJ files in a SciDAVis running on a big-endian system, using native byte order is definitely not what one wants.

I've added methods for row-wise and column-wise filling of matrices (r1336). These were already available on the C++ side; implementing filling of a matrix completely in one go would have to be implemented from scratch (if it's still needed).

I'll need some more time to review the rest of the patch (which seems to contain more changes than just handling of folders, matrices and graphs) and think about the issue of batch-converting multiple files.

Knut

[1] http://www.originlab.com/index.aspx?s=8&lm=85

[Knut Franke]

Applied those parts I could make sense of in r1346-r1356.