Pipeline integration: standalone=False

If you’re integrating GridR into a larger processing pipeline where you already handle preprocessing yourself, you can bypass the standalone preparation step entirely. This page shows when and how to do that.

What you’ll learn

• When to set standalone=False

• Adjusting the input coordinate origin with array_in_origin

• What preprocessing responsibilities you take on

Advanced Usage: Pipeline Integration (standalone=False)

Adjusting the Origin of Input Coordinates

The array_grid_resampling function includes the parameter array_in_origin (default: (0., 0.)) that allows shifting the origin of the input coordinate system.

From an implementation standpoint, this shift is applied directly to the computed target grid coordinates.

This feature is useful in several scenarios:

• Handling different grid conventions without modifying the grid itself. As a reminder, GridR assumes that integer coordinates refer to pixel centers. In other systems, a pixel may be treated as an area, with its center located relatively at (0.5, 0.5).

• Working with subregions of a larger input array. In this case, the shift can represent the position of the subregion within the full array’s coordinate system.

As an example we will take a previous example using an idendity transformation.

# create identity grid
if image.ndim == 2:
    x = np.arange(0, image.shape[0], dtype=grid_dtype)
    y = np.arange(0, image.shape[1], dtype=grid_dtype)
xx, yy = np.meshgrid(x, y)

win_center = np.array((58, 175)) # left eye
win_shape =  np.array((100, 100))
win = np.array((win_center - win_shape//2, win_center - win_shape//2 + win_shape - 1)).T

# Call it without specifying array_in_origin
array_out_origin_0_0, _ = array_grid_resampling(
        interp="cubic",
        array_in=array_in,
        grid_row=yy,
        grid_col=xx,
        grid_resolution=(1,1),
        array_out=None,
        win=win,
        array_in_mask=None,
        grid_mask=None,
        array_out_mask=None,
        nodata_out=None,
        )

# Call it with specifying array_in_origin and passing standalone=False
array_out_origin_30_m10, _ = array_grid_resampling(
        interp="cubic",
        array_in=array_in,
        array_in_origin=(30, -10), # <= shift +30 on rows and -10 on columns
        grid_row=yy,
        grid_col=xx,
        grid_resolution=(1,1),
        array_out=None,
        win=win,
        array_in_mask=None,
        grid_mask=None,
        array_out_mask=None,
        nodata_out=None,
        standalone=False,
        )

/home/docs/checkouts/readthedocs.org/user_builds/gridr/checkouts/latest/notebooks/../python/gridr/core/grid/grid_resampling.py:1784: UserWarning: Standalone mode is enabled : trust_padding will be ignored with `boundary_condition` set as `None`
  ) = standalone_preprocessing(

As you can observe, the coordinate shift is reflected in the bottom image.