Bottleneck

----------------

*Release date: 2010-01-19*

The third release of Bottleneck is twice as fast for small input arrays and
contains 10 new functions.

**Faster**

- All functions are faster (less overhead in selector functions)

**New functions**

- nansum()
- move_sum()
- move_nansum()
- move_mean()
- move_std()
- move_nanstd()
- move_min()
- move_nanmin()
- move_max()
- move_nanmax()

**Enhancements**

- You can now specify the dtype and axis to use in the benchmark timings
- Improved documentation and more unit tests

----------------

*Release date: 2010-12-27*

The second release of Bottleneck is faster, contains more functions, and
supports more dtypes.

**Faster**

- All functions faster (less overhead) when output is not a scalar
- Faster nanmean() for 2d, 3d arrays containing NaNs when axis is not None

**New functions**

- nanargmin()
- nanargmax()
- nanmedian()

**Enhancements**

- Added support for float32
- Fallback to slower, non-Cython functions for unaccelerated ndim/dtype
- Scipy is no longer a dependency
- Added support for older versions of NumPy (1.4.1)
- All functions are now templated for dtype and axis
- Added a sandbox for prototyping of new Bottleneck functions
- Rewrote benchmarking code

----------------

*Release date: 2010-12-10*

Initial release. The three categories of Bottleneck functions:

- Faster replacement for NumPy and SciPy functions
- Moving window functions
- Group functions that bin calculations by like-labeled elements