wip/py-dimarray 0.1.8 NetBSD

dimarray is a package to handle numpy arrays with labelled dimensions and axes Inspired from pandas, it includes advanced alignment and reshaping features and as well as missing-value (NaN) handling.

The main difference with pandas is that it is generalized to N dimensions, and behaves more closely to a numpy array. The axes do not have fixed names ('index', 'columns', etc...) but are given a meaningful name by the user (e.g. 'time', 'items', 'lon' ...). This is especially useful for high dimensional problems such as sensitivity analyses.

A natural I/O format for such an array is netCDF, common in geophysics, which relies on the netCDF4 package, and supports metadata.