Version 1.2.0 - Easy Indexing + Modern CMake

New Features:

• Easier indexing (variable argument index method + negative indexes)
• Easier access to shape

Improvements:

• Cmake upgraded to "modern" usage, i.e. you use the library by adding npy::npy as a link library

Version 1.1.1

Improvements:

• Minor cmake change

Version 1.1.0 - ZIP64 Support

New Features:

• Zip64 compatibility

Improvements:

• Can use numpy style lookup for tensors (i.e. dropping the .npy from the name)
• Added a crc32 test

Version 1.0.0

New Features:

• There is no longer a dependency on zlib

Improvements:

• Better packaging (NuGet packages are now produced for C++ and C#)

v0.2.0

Breaking changes:

• Renamed endian => endian_t
• Renamed data_type => data_type_t
• Renamed compression_method => compression_method_t

New Features:

• Cleaned up exception handling. There are now tests for exceptions being correctly thrown, and the exceptions are properly wrapped for .NET-
• Added peeking for NPY files to get the header information, and contains/peek functionality for inpzstream.

Improvements:

• Removed the unnecessary copies in the compression/decompression process

v0.1.0 - Initial Release

This is the first release of the library. Initial functionality is:

• Read/Write NPY files
• Read/Write NPZ files
• Built-in tensor class + support for user-implemented tensor classes
• .NET interface (Windows only)

Notably, support for files > 2GB is not yet support. This is planned for a future release.