Fixes error when indexing with N+1 keys where one of the keys is an Ellipsis.
0.9.0
Highlights: - Adds `Array.transpose()` (using TransposedArray). - Support for indexing with `np.newaxis` to add new axes. - Support automatic numpy array conversion via `__array__()`. - Initial broadcasting support via `BroadcastArray`. - A more human-readable format for `str()`.
Other improvements: - Tests work on 32-bit platforms.
0.8.0
Highlights: - Support for NumPy 1.9 - Create LinearMosaic and ArrayStack from arrays with inhomogeneous fill_values. - Extend `biggus.save()` to allow masked array output.
Bug fixes: - Masked array evaluation of variance. (107) - LinearMosaic indexing with non-zero axis. (113)
0.7.0
Highlights include: - `min`, `max` and `sum` operators - boolean indexing - fixed indexing of aggregations with a non-zero axis
0.6.0
Highlights include: - support for Python 3 - a `count` aggregator - a `mdtol` keyword for the mean aggregator - `zeros` and `ones`
0.5.0
Adds support for aggregations (mean, std, var) over any single axis. First draft docs on: http://biggus.readthedocs.org/