Add optional utilities for conversion to a DataFrame and for plotting with Gadfly

REQUIRE

@@ -0,0 +1,2 @@
+julia 0.4-
+Requires

src/AxisArrays.jl

@@ -1,9 +1,12 @@
 module AxisArrays
 
+using Requires
+
 export AxisArray, Axis, Interval, axisnames, axisdim, axes
 
 include("core.jl")
 include("intervals.jl")
 include("indexing.jl")
+include("utils.jl")
 
 end

src/utils.jl

@@ -0,0 +1,67 @@
+@require DataFrames begin
+
+    """
+    Convert an AxisArray into a long-format DataFrame.
+
+    ```julia
+    DataFrame(a::AxisArray)
+    ```
+
+    ### Arguments
+
+    * `a::AxisArray`
+
+    ### Returns
+
+    * `::DataFrame` : a DataFrame view into the AxisArray; columns are
+      added for each axis plus one column for the data (named `:data`).
+
+    """
+    function DataFrames.DataFrame(A::AxisArray)
+        colnames = Symbol[axisnames(A)..., :data]
+        sz = [1; size(A)...; 1]
+        columns = Any[DataFrames.RepeatedVector(a, prod(sz[1:i]), prod(sz[i+2:end])) for (i,a) in enumerate(axes(A))]
+        push!(columns, sub(A.data, 1:prod(sz)))
+        DataFrames.DataFrame(columns, colnames)
+    end
+
+end
+
+
+@require Gadfly begin
+
+    import DataArrays
+    ## Low-level code patching
+    function Gadfly.Scale.discretize_make_pda(values::DataFrames.RepeatedVector, levels=nothing)
+        if levels == nothing
+            return DataArrays.PooledDataArray(values)
+        else
+            return DataArrays.PooledDataArray(values[:], levels)
+        end
+    end
+
+    """
+    Plot an AxisArray using Gadfly.
+
+    ```julia
+    Gadfly.plot(A::AxisArray, args...; kargs...)
+    ```
+
+    ### Arguments
+
+    * `A::AxisArray`
+
+    All other arguments are passed to Gadfly.plot.
+
+    ### Examples
+
+    ```julia
+    using Gadfly
+    A = AxisArray(reshape([1:24], 12,2), (.1:.1:1.2, [:a,:b]))
+    plot(A, x = :row, y = :data, color = :col)
+    ```
+
+    """
+    Gadfly.plot(A::AxisArray, args...; kwargs...) = Gadfly.plot(DataFrames.DataFrame(A), args...; kwargs...)
+
+end