NDWidget Spike Raster

Browse a spike raster with an NDWidget. The array is [raster, spike, xy], where the value dim holds the spike time in seconds and the raster index of each spike, so each of the 20 rasters is a separate scatter in the collection.

The datapoints dim is time, so the one slider scrubs along it and display_window sets how many seconds of spikes are rendered at a time. That window is what keeps a long recording viewable, only the spikes in view are read and uploaded.

x_range_mode="auto" couples the camera to that window in both directions: the x-range follows the slider, and panning or zooming sets the display window to the new width and the index to the new center. The linear selector marks the current time.

# test_example = false

import numpy as np
import fastplotlib as fpl

n_rasters = 10
n_spikes = 2_000  # spikes per raster
duration = 120.0  # seconds

rng = np.random.default_rng(0)

# a shared sequence of population events, each raster fires near every event with its own jitter
events = np.sort(rng.uniform(0, duration, n_spikes))

# [raster, spike, xy], the value dim holds the spike time in seconds and the raster index
data = np.empty((n_rasters, n_spikes, 2), dtype=np.float32)
data[:, :, 0] = events + rng.normal(0, 0.05, (n_rasters, n_spikes))
data[:, :, 1] = np.arange(n_rasters)[:, None]

data2 = np.empty((n_rasters, n_spikes, 2), dtype=np.float32)
data2[:, :, 0] = events + rng.normal(0, 0.05, (n_rasters, n_spikes))
data2[:, :, 1] = np.arange(n_rasters)[:, None]


# reference space is seconds, one step per 60 Hz frame
ranges = {"time": (0, duration, 1 / 60)}

ndw = fpl.NDWidget(ranges=ranges, size=(700, 700), shape=(2, 1))

ndw[0, 0].add_nd_timeseries(
    data,
    dims=("raster", "time", "xy"),
    display_dims=("raster", "time", "xy"),
    graphic_type=fpl.ScatterCollection,
    slider_maps={"time": events},  # seconds -> index into the spike list
    display_window=5.0,  # seconds of spikes to render
    max_display_datapoints=1_000_000,  # spikes are cheap points, never decimate them
    x_range_mode="auto",
    cmap="viridis",  # one color per raster
    sizes=4,
    name="raster",
)

ndw[1, 0].add_nd_timeseries(
    data2,
    dims=("raster", "time", "xy"),
    display_dims=("raster", "time", "xy"),
    graphic_type=fpl.ScatterCollection,
    slider_maps={"time": events},  # seconds -> index into the spike list
    display_window=5.0,  # seconds of spikes to render
    max_display_datapoints=1_000_000,  # spikes are cheap points, never decimate them
    x_range_mode="auto",
    cmap="viridis",  # one color per raster
    sizes=4,
    name="raster",
)

subplot = ndw.figure[0, 0]

ndw.show(maintain_aspect=False)

figure = ndw.figure


# NOTE: fpl.loop.run() should not be used for interactive sessions
# See the "JupyterLab and IPython" section in the user guide
if __name__ == "__main__":
    print(__doc__)
    fpl.loop.run()