How to use the numcodecs.compat.ndarray_copy function in numcodecs

# normalise inputs
        buf = ensure_contiguous_ndarray(buf)
        if out is not None:
            out = ensure_contiguous_ndarray(out)

        # N.B., bz2 cannot handle ndarray directly because of truth testing issues
        buf = memoryview(buf)

        # do decompression
        dec = _bz2.decompress(buf)

        # handle destination - Python standard library bz2 module does not
        # support direct decompression into buffer, so we have to copy into
        # out if given
        return ndarray_copy(dec, out)

# find out how many bits were padded
        n_bits_padded = int(enc[0])

        # apply decoding
        dec = np.unpackbits(enc[1:])

        # remove padded bits
        if n_bits_padded:
            dec = dec[:-n_bits_padded]

        # view as boolean array
        dec = dec.view(bool)

        # handle destination
        return ndarray_copy(dec, out)

def decode(self, buf, out=None):
        # filter is lossy, decoding is no-op
        dec = ensure_ndarray(buf).view(self.astype)
        dec = dec.astype(self.dtype, copy=False)
        return ndarray_copy(dec, out)

def decode(self, buf, out=None):

        # normalise input
        enc = ensure_ndarray(buf).view(self.astype)

        # flatten to simplify implementation
        enc = enc.reshape(-1, order='A')

        # setup decoded output
        dec = np.empty_like(enc, dtype=self.dtype)

        # decode differences
        np.cumsum(enc, out=dec)

        # handle output
        out = ndarray_copy(dec, out)

        return out

def decode(self, buf, out=None):

            # normalise inputs
            buf = ensure_contiguous_ndarray(buf)
            if out is not None:
                out = ensure_contiguous_ndarray(out)

            # do decompression
            dec = _lzma.decompress(buf, format=self.format, filters=self.filters)

            # handle destination
            return ndarray_copy(dec, out)

def decode(self, buf, out=None):

        # normalise inputs
        buf = ensure_contiguous_ndarray(buf)
        if out is not None:
            out = ensure_contiguous_ndarray(out)

        # do decompression
        dec = _zlib.decompress(buf)

        # handle destination - Python standard library zlib module does not
        # support direct decompression into buffer, so we have to copy into
        # out if given
        return ndarray_copy(dec, out)

def encode(self, buf):
        arr = ensure_contiguous_ndarray(buf).view('u1')
        checksum = self.checksum(arr) & 0xffffffff
        enc = np.empty(arr.nbytes + 4, dtype='u1')
        enc[:4].view('

# normalise input
        enc = ensure_ndarray(buf).view(self.astype)

        # flatten to simplify implementation
        enc = enc.reshape(-1, order='A')

        # setup output
        dec = np.full_like(enc, fill_value='', dtype=self.dtype)

        # apply decoding
        for i, l in enumerate(self.labels):
            dec[enc == (i + 1)] = l

        # handle output
        dec = ndarray_copy(dec, out)

        return dec

def decode(self, buf, out=None):

        # normalise input
        enc = ensure_ndarray(buf).view(self.encode_dtype)

        # convert and copy
        dec = enc.astype(self.decode_dtype)

        # handle output
        out = ndarray_copy(dec, out)

        return out

def decode(self, buf, out=None):
        arr = ensure_contiguous_ndarray(buf).view('u1')
        expect = arr[:4].view('