coco.py

class CocoSPI()

    def __init__(self, dut):
        super().__init__() 
        self.dut = dut

    # duplicates the behaviour of real spi driver
    async def xfer(self, data):
        await ClockCycles(self.dut.clk, SPI_CLOCK_PERIODS)
        return_data = [0] * len(data)
        for bit in range(8):
            self.dut.spi_copi <= data[bit]
            return_data[bit] = self.dut.spi_cipo

        return return_data

class SPIDriver():
    
    def __init__():
        # real hardware
        self.spi = SpiDev.spi(0)

    def read_register(self, reg)
        self.spi.xfer(reg)
        data = self.spi.xfer(0)

        # do some data processing on data before returning it
        val = struct.unpack('>B', data)

        return val

class CocoSPIDriver(SPIDriver):

    def __init__(dut):
        # simulated hardware
        self.dut = dut
        self.spi = CocoSPI(dut)

    # I want to avoid duplicating all the super class methods
    async def read_register(self, reg)
        self.spi.xfer(reg)
        data = await self.spi.xfer(0)

        # do some data processing on data before returning it
        val = struct.unpack('>B', data)

        return val


# simulated test
@cocotb.test()
async def test_read(dut):
    clock = Clock(dut.clk, CLOCK_PERIOD, units="ns")
    cocotb.fork(clock.start())  # Start the clock
    spi = CocoSPIDriver(dut)
    value = await spi.read_register(0x40)
    assert value == 10

# hardware in the loop test
if __name__ == '__main__':
    spi = SPIDriver()
    value = spi.read_register(0x40)
    assert value == 10