Jupyter notebooks

With the jupyter extra, charts are drawn in notebooks:

pip install "geneviewer-py[jupyter]"

The charts are anywidget widgets, so they work in JupyterLab, Jupyter Notebook, VS Code, Google Colab and marimo.

A chart in a cell

A cell that ends with a chart draws it below the cell:

from geneviewer import GC, datasets

clusters = datasets.load("ophA_clusters")
ophA = clusters[clusters["cluster"] == "ophA"]

chart = GC(ophA, group="class", height="150px")

The chart is the live chart of the other pages of this site: hover a gene for its tooltip, and the chart for the button that saves it as SVG or PNG.

Events

To read what happens in a chart, keep it as a widget with .widget(). The events switched on with .events() set attributes of the widget:

widget = (
    GC(clusters, cluster="cluster", group="class", height="300px")
    .labels("name")
    .zoom()
    .events(select=True, drag=True, zoom=True)
    .widget()
)
widget

After a click on a gene, a later cell can read it:

widget.select["data"]["name"]

Attribute	Value	Switched on by
widget.click	The element clicked	click=True (the default)
widget.dblclick	The element double-clicked	dblclick=True
widget.select	The element selected last	select=True
widget.selection	Every selected element, as a list	select=True
widget.drag	An element moved by dragging, and its offset	drag=True
widget.moves	The elements a reset, undo or redo moved	drag=True
widget.zoom	The window of the zoom slider	zoom=True
widget.rendered	The chart was drawn	rendered=True

Each is None until its event happens, and then a dict (a list for selection); Events describes what each holds. These are the values a Shiny app gets as input.<id>_<event>().

widget.on() calls a function at each event, also when the same event happens twice:

selected = []
widget.on("select", selected.append)

Moving the zoom slider

widget.zoom_to() sets the window of a chart with a zoom slider (.zoom()), without drawing the chart again. None shows everything:

widget.zoom_to([0, 0.5])
widget.zoom_to(None)

With .events(zoom=True), widget.zoom["range"] is the window the reader picked, which can go back into zoom_to().

Drawing another chart

Setting widget.chart draws another chart in the same place, for example the same data with another theme:

widget.chart = widget.chart.theme("dark")

Toolbar icons

The toolbar's own icons need nothing. Custom icons, set with .toolbar(), are Font Awesome icons by name and need Font Awesome on the page. JupyterLab does not load it, so load it in a cell:

from IPython.display import HTML

HTML(
    '<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/'
    '@fortawesome/fontawesome-free@6/css/all.min.css">'
)