from ipylab import JupyterFrontEnd
app = JupyterFrontEnd()
app.version
app.commands.list_commands()
app.commands.execute('console:create', {
'insertMode': 'split-right',
'kernelPreference': {
'shutdownOnClose': True,
}
})
app.commands.execute('apputils:change-theme', { 'theme': 'JupyterLab Dark' })
app.commands.execute('terminal:create-new')
Some functionalities might require the JupyterFrontEnd widget to be ready on the frontend first.
This is for example the case when listing all the available commands, or retrieving the version with app.version.
The on_ready method can be used to register a callback that will be fired when the frontend is ready.
from ipylab import JupyterFrontEnd
from ipywidgets import Output
app = JupyterFrontEnd()
out = Output()
def init():
# show the first 5 commands
cmds = app.commands.list_commands()[:5]
out.append_stdout(cmds)
app.on_ready(init)
out
Or using asyncio:
import asyncio
app = JupyterFrontEnd()
out = Output()
async def init():
await app.ready()
cmds = app.commands.list_commands()[:5]
out.append_stdout(cmds)
asyncio.create_task(init())
out
Let's create a nice plot with bqlot and generate some random data.
See https://github.com/bqplot/bqplot/blob/master/examples/Advanced%20Plotting/Animations.ipynb for more details.
import numpy as np
from bqplot import LinearScale, Lines, Bars, Axis, Figure
from ipywidgets import IntSlider
xs = LinearScale()
ys1 = LinearScale()
ys2 = LinearScale()
x = np.arange(20)
y = np.cumsum(np.random.randn(20))
y1 = np.random.rand(20)
line = Lines(x=x, y=y, scales={'x': xs, 'y': ys1}, colors=['magenta'], marker='square')
bar = Bars(x=x, y=y1, scales={'x': xs, 'y': ys2}, colorpadding=0.2, colors=['steelblue'])
xax = Axis(scale=xs, label='x', grid_lines='solid')
yax1 = Axis(scale=ys1, orientation='vertical', tick_format='0.1f', label='y', grid_lines='solid')
yax2 = Axis(scale=ys2, orientation='vertical', side='right', tick_format='0.0%', label='y1', grid_lines='none')
Figure(marks=[bar, line], axes=[xax, yax1, yax2], animation_duration=1000)
We now define a function to update the data.
def update_data():
line.y = np.cumsum(np.random.randn(20))
bar.y = np.random.rand(20)
update_data()
This function will now be called when the JupyterLab command is executed.
Commands can also custom icons in place of
icon_class.
app.commands.add_command('update_data', execute=update_data, label="Update Data", icon_class="jp-PythonIcon")
Execute it!
app.commands.execute('update_data')
The slider should now be moving and taking random values.
Also the list of commands gets updated with the newly added command:
assert 'random' in app.commands.list_commands()
That's great, but the command doesn't visually show up in the palette yet. So let's add it!
from ipylab.commands import CommandPalette
palette = CommandPalette()
palette.add_item('update_data', 'Python Commands')
Open the command palette on the left side and the command should show now be visible.
To remove a command that was previously added:
app.commands.remove_command('update_data')