Python proxy

The structure of the Python proxy is identical for both SCADE Suite and Scade One models.

The Python module defines:

• A class per structure used in the interfaces of the root operators. These classes derive from ctypes.Structure.

• A global variable sensors with one read/write property per sensor in the model.

• A class per root operator which defines:

  • One read/write property per input

  • One read-only property per output

  • Two functions to address the C code:

    • call_reset(self) -> None

    • call_cycle(self, cycles: int = 1, refresh: bool = True, debug: bool = False) -> None

      The refresh and debug parameters are used for SCADE Suite co-simulation.

Example

The python module’s usage is straightforward:

• Create an instance of a root operator.

• Call its reset function.

• Loop:

  • Set the sensors

  • Set the inputs

  • Call the cyclic function

  • Get the outputs

The script hereafter is an example of client for the following model:

# py_box.py/py_box.dll are produced from PyBox.etp
import py_box
# set the sensors
# P::offset: float64
py_box.sensors.offset = 0.5
# create an instance of the root operator P::Root
root = py_box.Root()
# and reset it
root.call_reset()
# set the inputs
# P::Root/c: bool
root.c = True
# P::Root/v: Speed (defined as float64 ^ 3)
root.v = (1.0, 2.0, 3.0)
# P::Root/v: float64
root.dt = 0.1
for cycle in range(4):
    # P::Root/i: int32
    root.i = cycle + 1
    # call the cyclic function
    root.call_cycle()
    # print the results
    # P::Root/pos: Position
    print(root.o, root.pos.x, root.pos.y, root.pos.z)
    # P::Root/c: bool
    root.c = False

Access to values

• Scalar values: Use Python literals

  root.valid = True
  root.array[0] = 3.14
  root.points[2].x = 0
  

• Complex Values: Use Python tuples or ctypes literals

  Consider the following types:

  

  • OK

    root.array = (ctypes.double*3)([3, 1, 4])
    root.array = (3, 1, 4)
    root.points[2] = CPosition_P(x=1, y=2, z=3)
    root.points[2] = (1, 2, 3)
    root.points = ((1, 2, 3), (3, 4, 5), (6, 7, 8))
    

  • NOK

    root.array = [3, 1, 4]
    root.points[2] = {"x": 1, "y": 2, "z": 3}
    

Limitations

The wrapper does not support for now the input_threshold and global_context KCG options.