dtype consistency (float vs float32)

[taka-rl]

Make sure that the data type shall be consistent np.float32.
Due to the complexity of the project, the project follows these for now.

• Observation(state) and actions are in float32.
• Rendering/Physics calculation are in float64/float.

Rendering requires np.float64 as pygame.draw.polygon(screen, color, pixel_vertex) causes an error if pixel_vertex is np.float32, not np.float64.

Todo

parking_env.py

• dtype consistency (float vs float32)_ParkingEnv #39

Todo later not now

car.py

• each member of car class
• return rotated vertices in np.float32 in rotate_car()
• kinematic bicycle model calculation and update a state
• Add tests to ensure the data type.

parking.py

• Ensure that each method returns in np.float32.
• CarSize
• WheelSize
• ParkingLotSize
• Config
• Add tests to ensure the data type.

references

NumPy type hints

some examples of numpy data type hints:

def example(
    x: np.ndarray[tuple[int], np.dtype[np.bool_]],
    y: np.ndarray[tuple[int, int, int], np.dtype[np.uint8]],
    ) -> np.ndarray[tuple[int], np.dtype[np.float32]]: ...

• A 1D array of Booleans:

np.ndarray[tuple[int], np.dtype[np.bool_]]

• A 3D array of unsigned 8-bit integers:

np.ndarray[tuple[int, int, int], np.dtype[np.uint8]]

• A two-dimensional (2D) array of float32:

np.ndarray[tuple[int, int], np.dtype[np.float32]]

• A 1D array of any numeric type:

np.ndarray[tuple[int], np.dtype[np.number]]

Improving Code Quality with Array and DataFrame Type Hints