Canvas Widget

import napari
import numpy as np
from biaplotter.plotter import CanvasWidget, ArtistType
from napari.utils import nbscreenshot

Adding Canvas Widget to napari

viewer = napari.Viewer()
canvas_widget = CanvasWidget(viewer)
viewer.window.add_dock_widget(canvas_widget)
nbscreenshot(viewer)

WARNING: QWindowsWindow::setGeometry: Unable to set geometry 2172x1310+728+260 (frame: 2198x1381+715+202) on QWidgetWindow/"_QtMainWindowClassWindow" on "\\.\DISPLAY1". Resulting geometry: 4350x2119+733+287 (frame: 4376x2190+720+229) margins: 13, 58, 13, 13 minimum size: 385x500 MINMAXINFO maxSize=0,0 maxpos=0,0 mintrack=796,1071 maxtrack=0,0)

Check canvas_widget default artists and selectors

canvas_widget.artists

{<ArtistType.SCATTER: 2>: <biaplotter.artists.Scatter at 0x2bc78b5eac0>,
 <ArtistType.HISTOGRAM2D: 1>: <biaplotter.artists.Histogram2D at 0x2bc78b5eaf0>}

canvas_widget.selectors

{<SelectorType.LASSO: 1>: <biaplotter.selectors.InteractiveLassoSelector at 0x2bc78b5eb80>,
 <SelectorType.ELLIPSE: 2>: <biaplotter.selectors.InteractiveEllipseSelector at 0x2bc78b5eca0>,
 <SelectorType.RECTANGLE: 3>: <biaplotter.selectors.InteractiveRectangleSelector at 0x2bc78b5ecd0>}

Generate some data

def generate_gaussian_data(n_samples):
    """Generate a 2D dataset with two Gaussian clusters."""
    # Gaussian 1
    x1 = np.random.normal(loc=2, scale=1, size=n_samples//2)
    y1 = np.random.normal(loc=2, scale=1, size=n_samples//2)
    # Gaussian 2
    x2 = np.random.normal(loc=-2, scale=0.5, size=n_samples//2)
    y2 = np.random.normal(loc=-2, scale=0.5, size=n_samples//2)
    x_data = np.concatenate([x1, x2])
    y_data = np.concatenate([y1, y2])
    return np.vstack([x_data, y_data]).T

n_samples = 1000
data = generate_gaussian_data(n_samples)

Histogram2D artist is the default active artist.

canvas_widget.active_artist

<biaplotter.artists.Histogram2D at 0x2bc78b5eaf0>

Adding data to the canvas

Adding data to artist will update the histogram.

canvas_widget.active_artist.data = data
nbscreenshot(viewer)

Similarly, changing the active artist to SCATTER and adding data to it will update the scatter plot.

canvas_widget.active_artist = canvas_widget.artists[ArtistType.SCATTER]
canvas_widget.active_artist.data = data
nbscreenshot(viewer)

To go back to the histogram, change the active artist to HISTOGRAM2D.

canvas_widget.active_artist = canvas_widget.artists[ArtistType.HISTOGRAM2D]

Manually selecting data

Lasso Selector

Selecting data using lasso selector will update the active artist plot.

Ellipse Selector

Ellipse selector produces an interactive ellipse that can be dragged and expanded before being applied.

Right-clicking applies the current class to the selected data.

Rectangle Selector

Rectangle selector produces an interactive rectangle that can be dragged and expanded as well.

One way to clear previuos selections is to apply class 0 to the selected data.