I have just added some intersting functions to create arrays from Elixir. Now you can create a 1D-array from an elixir list (I’m still thinking on what the best way to implement the creating of multidimensional arrays) and a 2D-array from a black and white PNG image:
iex> iex(2)> SciEx.Float64.Array1.from_list([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0])
#SciEx.Float64.Array1<[1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], shape=[7], strides=[1], layout=CFcf (0xf), const ndim=1>
iex> SciEx.Float64.Array2.from_image("test/data/visualization/heightmap.png")
#SciEx.Float64.Array2<[[0.27450980392156865, 0.2549019607843137, 0.21176470588235294, 0.1843137254901961, 0.1568627450980392, ..., 0.4470588235294118, 0.44313725490196076, 0.43137254901960786, 0.3843137254901961, 0.3215686274509804],
[0.2823529411764706, 0.25098039215686274, 0.21176470588235294, 0.1803921568627451, 0.14901960784313725, ..., 0.3843137254901961, 0.3764705882352941, 0.38823529411764707, 0.34509803921568627, 0.2980392156862745],
[0.25882352941176473, 0.21568627450980393, 0.18823529411764706, 0.1607843137254902, 0.13333333333333333, ..., 0.3176470588235294, 0.3176470588235294, 0.3333333333333333, 0.3176470588235294, 0.2901960784313726],
[0.22745098039215686, 0.2, 0.1803921568627451, 0.15294117647058825, 0.12941176470588237, ..., 0.3058823529411765, 0.29411764705882354, 0.27450980392156865, 0.24705882352941178, 0.23921568627450981],
[0.21176470588235294, 0.1803921568627451, 0.16470588235294117, 0.1411764705882353, 0.12549019607843137, ..., 0.27450980392156865, 0.2549019607843137, 0.2, 0.17647058823529413, 0.1803921568627451],
...,
[0.2980392156862745, 0.27450980392156865, 0.2627450980392157, 0.27058823529411763, 0.2823529411764706, ..., 0.23137254901960785, 0.22745098039215686, 0.2235294117647059, 0.22745098039215686, 0.21568627450980393],
[0.30980392156862746, 0.29411764705882354, 0.2980392156862745, 0.2980392156862745, 0.2823529411764706, ..., 0.21568627450980393, 0.2235294117647059, 0.2196078431372549, 0.21176470588235294, 0.21176470588235294],
[0.3333333333333333, 0.3254901960784314, 0.3254901960784314, 0.3215686274509804, 0.30196078431372547, ..., 0.21176470588235294, 0.21568627450980393, 0.2235294117647059, 0.2196078431372549, 0.21176470588235294],
[0.34901960784313724, 0.34509803921568627, 0.35294117647058826, 0.34901960784313724, 0.32941176470588235, ..., 0.20392156862745098, 0.21568627450980393, 0.21176470588235294, 0.21568627450980393, 0.2196078431372549],
[0.35294117647058826, 0.3411764705882353, 0.35294117647058826, 0.35294117647058826, 0.3411764705882353, ..., 0.17647058823529413, 0.2, 0.20784313725490197, 0.21176470588235294, 0.2235294117647059]], shape=[257, 257], strides=[-257, 1], layout=c (0x4), const ndim=2>
These are small things but very important in practice when exploring the library. Although you’ll probably want to load your data from big files stored on your hard drive somehow (I still need to add support for that), having simple ways of loading basic data goes a long way when trying things out.
