quicksort solution is incorrect (auxiliary space used)

[wulymammoth]

Awesome repo, but I came across the quicksort solution done here and I don't believe it's the canonical solution when quicksort is typically described to be an "in-place" sort. The implemented solution allocates new arrays in which values to the left and right are placed into their corresponding auxiliary arrays (lists) -- I've denoted the places that I see them in the code w/ a comment. I believe the implementation should utilize swaps w/o allocating auxiliary data structures. Lastly, the list concatenation also incurs a linear time cost as a new list/array is allocated and all elements in each sub-array must be copied over:

def _sort(self, data):
        if len(data) < 2:
            return data
        equal = [] # extra space
        left = [] # extra space
        right = [] # extra space
        pivot_index = len(data) // 2
        pivot_value = data[pivot_index]
        # Build the left and right partitions
        for item in data:
            if item == pivot_value:
                equal.append(item)
            elif item < pivot_value:
                left.append(item)
            else:
                right.append(item)
        # Recursively apply quick_sort
        left_ = self._sort(left)
        right_ = self._sort(right)
        return left_ + equal + right_ # O(n) [where n is the size of the original input]