Convert vector in mojo to image and save in python

[ThuyyTran]

struct Matrix3D:
    var height: Int
    var width: Int
    var depth: Int
    var total_items: Int
    var data: Pointer[Float64]
    
    fn __init__(inout self, default_value: Float64, height: Int, width: Int, depth: Int) -> None:
        self.height = height
        self.width = width
        self.depth = depth
        self.total_items = height * width * depth
        self.data = Pointer[Float64].alloc(self.total_items)
        for i in range(self.total_items):
            self.data.store(i, default_value)
    
    fn __getitem__(borrowed self, row: Int, column: Int, depth: Int) -> Float64:
        let loc: Int = (row * self.width * self.depth) + (column * self.depth) + depth
        if loc >= self.total_items:
            print("index out of range")
            return self.data.load(0)
        return self.data.load(loc)
    
    fn __setitem__(inout self, row: Int, column: Int, depth: Int, item: Float64) -> None:
        let loc: Int = (row * self.width * self.depth) + (column * self.depth) + depth
        if loc >= self.total_items:
            print("index out of range")
        else:
            self.data.store(loc, item)
    
    fn __del__(owned self) -> None:
        self.data.free()
    
    fn __len__(borrowed self) -> Int:
        return self.total_items
    
    fn __copyinit__(inout self, other: Self) -> None:
        self.height = other.height
        self.width = other.width
        self.depth = other.depth
        self.total_items = other.total_items
        self.data = Pointer[Float64].alloc(self.total_items)
        memcpy[Float64](self.data, other.data, self.total_items)
    
    fn __printall__(borrowed self) -> None:
        print('[')
        for i in range(self.height):
            print("  [")
            for j in range(self.width):
                print_no_newline("    [")
                for d in range(self.depth):
                    print_no_newline(self[i, j, d])
                    if d != self.depth - 1:
                        print_no_newline(", ")
                print("]," if j != self.width - 1 else "]")
            print("  ]," if i != self.height - 1 else "  ]")
        print("]")

I work on 3-dimensional matrices (specifically images). During processing, if I process with PythonObject(numpy array), it takes a very long time, but when I use SIMD or build a 3-dimensional matrix structure myself, my processing time is very fast. But when it comes to saving my processed image, I don't know how to change from SIMD format or from my structural format to the image format to save. Do you have any ideas for this problem?

[Vettinari]

Hey, they need to provide jpeg encoder/decoder for Io operations in Mojo.