News

v0.5.0

the default missing value representation was switched to DataArrays, so by default you will see a DataArray inside an inner function
if you choose to use :mask as missing value representation the actual data array and the mask will be passed as a tuple to the inner function, instead of being two separate arguments
there were lots of changes in the internals of mapCube please open issues if you find old code not working anymore or you find some serious performance regressions
there is a new method for registering functions available, which avoids all these nested tuples for the input and output dimensions. See

the package was updated to julia 0.6
a new cube type TransformedCube was introduced, to do simple element-wise operations without having to call mapCube
so thing like adding scalars to cube, and elementwise operations between cubes are possible now
see the demo notebook 02 for details
the reduceCube backend was switched to DataArrays, so when calling the function you will see a DataArray inside with missings represented as NA

the plotting module was split out into a separate package called ESDLPlots, so please add this package if you want to use Plots
the time axis is now based on Date - and not on DateTime anymore. This means that some methods are not applicable on old datacubes anymore, so please be careful

a new macro @loadOrGenerate was added to generate and save intermediate results or load them if already existent
a wrapper to calculate Online PCAs was added with cubePCA
a new map plotting function plotMAPRGB to generate RGB, Lab, HSV etc plots
print cube size when showing cube info

a list of pre-defined lon-lat boxes was added that can be accessed when calling getCubeData, so for example getCubeData(c,region="Germany") will extract a Germany from the Cube.
Support for masks was extended, this includes multiple changes:
- every cube now has a properties field, which contains a Dict{String,Any} where additional cube propertyies can be stored
- if an Integer-valued cube is used as a categorical mask one can add a Dict containg the mapping from value to label name to the cubes properties, eg c.properties["labels"]=Dict(1=>"low",2=>"high")
- plotMAP will respect the defined label properties and use the labels for its legend
- if a labeled cube is used in the by argument when fitting an OnlineStat, the output axis is automatically created
- accessing a static variable (currently only water_mask or country_mask) through getCubeData will by default only return the first time step