Set multiscale resolution level

Demonstrates the locked_data_level feature, which lets you lock a multiscale image to a specific resolution level instead of letting napari choose automatically.

Each level has its index number drawn into every slice so you can immediately tell which resolution is being rendered.

Tags: visualization-advanced

import argparse

import numpy as np

import napari

# -- 5x3 bitmap font for digits 0-9 ----------------------------------------
_DIGITS = {
    0: [0b111, 0b101, 0b101, 0b101, 0b111],
    1: [0b010, 0b110, 0b010, 0b010, 0b111],
    2: [0b111, 0b001, 0b111, 0b100, 0b111],
    3: [0b111, 0b001, 0b111, 0b001, 0b111],
    4: [0b101, 0b101, 0b111, 0b001, 0b001],
    5: [0b111, 0b100, 0b111, 0b001, 0b111],
    6: [0b111, 0b100, 0b111, 0b101, 0b111],
    7: [0b111, 0b001, 0b010, 0b010, 0b010],
    8: [0b111, 0b101, 0b111, 0b101, 0b111],
    9: [0b111, 0b101, 0b111, 0b001, 0b111],
}


def _stamp_digit(arr, digit, scale=1):
    """Draw *digit* centred on every YX slice of a 3D array.

    Parameters
    ----------
    arr : np.ndarray
        3D array (Z, Y, X) to draw into (modified in-place).
    digit : int
        Single digit 0-9.
    scale : int
        Pixel multiplier for the 5x3 glyph.
    """
    rows = _DIGITS[digit]
    glyph_h, glyph_w = 5 * scale, 3 * scale
    y0 = (arr.shape[1] - glyph_h) // 2
    x0 = (arr.shape[2] - glyph_w) // 2
    for r, bits in enumerate(rows):
        for c in range(3):
            if bits & (1 << (2 - c)):
                y_start = y0 + r * scale
                x_start = x0 + c * scale
                arr[:, y_start : y_start + scale, x_start : x_start + scale] = 255


# -- Parse arguments --------------------------------------------------------
parser = argparse.ArgumentParser(description='Multiscale volume demo')
parser.add_argument(
    '--labels',
    action='store_true',
    help='Add as a Labels layer instead of Image',
)
args = parser.parse_args()

# -- Build a 4-level 3D multiscale pyramid ----------------------------------
shapes = [(128, 256, 256), (64, 128, 128), (32, 64, 64), (16, 32, 32)]
scales = [40, 20, 10, 5]
multiscale_data = []
for i, (shape, sc) in enumerate(zip(shapes, scales, strict=True)):
    if args.labels:
        arr = np.zeros(shape, dtype=np.uint8)
    else:
        rng = np.random.default_rng(i)
        arr = rng.integers(0, 50, size=shape, dtype=np.uint8)
    _stamp_digit(arr, i, scale=sc)
    multiscale_data.append(arr)

# -- Open viewer in 3D and add the layer -----------------------------------
viewer = napari.Viewer(ndisplay=3)
if args.labels:
    layer = viewer.add_labels(
        multiscale_data, name='multiscale', multiscale=True
    )
else:
    layer = viewer.add_image(
        multiscale_data,
        name='multiscale',
        multiscale=True,
        rendering='attenuated_mip',
        interpolation3d='nearest',
    )

# Tilt the camera for a better 3D view
viewer.camera.angles = (-10, 10, -10)

# -- Lock to the highest resolution level so you can see the "0" digit.
# Without locking, 3D mode defaults to the coarsest level (digit "3").
layer.locked_data_level = 0

# To restore automatic level selection:
# layer.locked_data_level = None

if __name__ == '__main__':
    napari.run()