Textual is an excellent Python framework for building applications in the terminal, or on the web. This library provides a plot widget which your app can use to plot all kinds of quantitative data. So, no pie charts, sorry. The widget support scatter plots and line plots, and can also draw using high-resolution characters like unicode half blocks, quadrants and 8-dot Braille characters. It may still be apparent that these are drawn using characters that take up a full block in the terminal, especially when plot series overlap. However, the use of these characters can reduce the line thickness and improve the resolution tremendously.
The daytime spectrum dataset shows the visible-light spectrum recorded by an Ocean Optics USB2000+ spectrometer using the DeadSea Optics software. It was taken in the morning while the detector was facing my office window.
- Line plots
- Scatter plots
- Automatic scaling and tick placement at nice intervals (1, 2, 5, etc.)
- Axes labels
- High-resolution modes using unicode half blocks (1x2), quadrants (2x2) and braille (2x8) characters
- Mouse support for zooming (mouse scrolling) and panning (mouse dragging)
- Horizontal- or vertical-only zooming and panning when the mouse cursor is in the plot margins
Using uv:
uvx textual-plotUsing pipx:
pipx run textual-plotInstall the package with either
uv tool install textual-plotor
pipx install textual-plotAlternatively, install the package with pip (please, use virtual environments) and run the demo:
pip install textual-plotIn all cases, you can run the demo with
textual-plotA minimal example is shown below:
from textual.app import App, ComposeResult
from textual_plot import PlotWidget
class MinimalApp(App[None]):
def compose(self) -> ComposeResult:
yield PlotWidget()
def on_mount(self) -> None:
plot = self.query_one(PlotWidget)
plot.plot(x=[0, 1, 2, 3, 4], y=[0, 1, 4, 9, 16])
MinimalApp().run()You include a PlotWidget in your compose method and after your UI has finished composing, you can start plotting data. The plot() method takes x and y data which should be array-like. It can be lists, or NumPy arrays, or really anything that can be turned into a NumPy array which is what's used internally. The plot() method further accepts a line_style argument which accepts Textual styles like "white", "red on blue3", etc. For standard low-resolution plots, it does not make much sense to specify a background color since the text character used for plotting is a full block filling an entire cell.
The plot widget supports high-resolution plotting where the character does not take up the full cell:
from textual.app import App, ComposeResult
from textual_plot import HiResMode, PlotWidget
class MinimalApp(App[None]):
def compose(self) -> ComposeResult:
yield PlotWidget()
def on_mount(self) -> None:
plot = self.query_one(PlotWidget)
plot.plot(
x=[0, 1, 2, 3, 4],
y=[0, 1, 4, 9, 16],
hires_mode=HiResMode.BRAILLE,
line_style="bright_yellow on blue3",
)
MinimalApp().run()Admittedly, you'll be mostly plotting with foreground colors only. The plot widget supports four high-resolution modes: Hires.BRAILLE (2x8), HiRes.HALFBLOCK (1x2) and HiRes.QUADRANT (2x2) where the size between brackets is the number of 'pixels' inside a single cell.
To create scatter plots, use the scatter() method, which accepts a marker argument which can be any unicode character (as long as it is one cell wide, which excludes many emoji characters and non-Western scripts):
import numpy as np
from textual.app import App, ComposeResult
from textual_plot import PlotWidget
class MinimalApp(App[None]):
def compose(self) -> ComposeResult:
yield PlotWidget()
def on_mount(self) -> None:
rng = np.random.default_rng(seed=4)
plot = self.query_one(PlotWidget)
x = np.linspace(0, 10, 21)
y = 0.2 * x - 1 + rng.normal(loc=0.0, scale=0.2, size=len(x))
plot.scatter(x, y, marker="⦿")
MinimalApp().run()Finally, the code of the demo is given below, showing how you can handle multiple plots and updating 'live' data:
import importlib.resources
import itertools
import numpy as np
from textual.app import App, ComposeResult
from textual.containers import Container
from textual.widgets import Footer, Header, TabbedContent, TabPane
from textual_hires_canvas import HiResMode
from textual_plot import PlotWidget
class SpectrumPlot(Container):
BINDINGS = [("m", "cycle_modes", "Cycle Modes")]
_modes = itertools.cycle(
[HiResMode.QUADRANT, HiResMode.BRAILLE, None, HiResMode.HALFBLOCK]
)
mode = next(_modes)
def compose(self) -> ComposeResult:
yield PlotWidget()
def on_mount(self) -> None:
# Read CSV data included with this package
self.spectrum_csv = importlib.resources.read_text(
"textual_plot.resources", "morning-spectrum.csv"
).splitlines()
# plot the spectrum and set ymin limit once
self.plot_spectrum()
self.query_one(PlotWidget).set_ylimits(ymin=0)
def plot_spectrum(self) -> None:
x, y = np.genfromtxt(
self.spectrum_csv,
delimiter=",",
names=True,
unpack=True,
)
plot = self.query_one(PlotWidget)
plot.clear()
plot.plot(x, y, hires_mode=self.mode)
plot.set_xlabel("Wavelength (nm)")
plot.set_ylabel("Intensity")
def action_cycle_modes(self) -> None:
self.mode = next(self._modes)
self.plot_spectrum()
class SinePlot(Container):
_phi: float = 0.0
def compose(self) -> ComposeResult:
yield PlotWidget()
def on_mount(self) -> None:
self._timer = self.set_interval(1 / 24, self.plot_moving_sines, pause=True)
def on_show(self) -> None:
self._timer.resume()
def on_hide(self) -> None:
self._timer.pause()
def plot_moving_sines(self) -> None:
plot = self.query_one(PlotWidget)
plot.clear()
x = np.linspace(0, 10, 41)
y = x**2 / 3.5
plot.scatter(
x,
y,
marker_style="blue",
# marker="*",
hires_mode=HiResMode.QUADRANT,
)
x = np.linspace(0, 10, 200)
plot.plot(
x=x,
y=10 + 10 * np.sin(x + self._phi),
line_style="blue",
hires_mode=None,
)
plot.plot(
x=x,
y=10 + 10 * np.sin(x + self._phi + 1),
line_style="red3",
hires_mode=HiResMode.HALFBLOCK,
)
plot.plot(
x=x,
y=10 + 10 * np.sin(x + self._phi + 2),
line_style="green",
hires_mode=HiResMode.QUADRANT,
)
plot.plot(
x=x,
y=10 + 10 * np.sin(x + self._phi + 3),
line_style="yellow",
hires_mode=HiResMode.BRAILLE,
)
self._phi += 0.1
class DemoApp(App[None]):
AUTO_FOCUS = "PlotWidget"
def compose(self) -> ComposeResult:
yield Header()
yield Footer()
with TabbedContent():
with TabPane("Daytime spectrum"):
yield SpectrumPlot()
with TabPane("Moving sines"):
yield SinePlot()
def main():
app = DemoApp()
app.run()
if __name__ == "__main__":
main()clear(): clear the plot.plot(x, y, line_style, hires_mode, label): plot a dataset with a line using the specified linestyle and high-resolution mode.scatter(x, y, marker, marker_style, hires_mode, label): plot a dataset with markers using the specified marker, marker style and high-resolution mode.set_xlimits(xmin, xmax): set the x-axis limits.Nonemeans autoscale.set_ylimits(xmin, xmax): set the y-axis limits.Nonemeans autoscale.set_xticks(ticks): manually specify x-axis tick locations.set_yticks(ticks): manually specify y-axis tick locations.set_xlabel(label): set the x-axis label.set_ylabel(label): set the y-axis label.show_legend(location, is_visible): show or hide the plot legend.
Various other methods exist, mostly for coordinate transformations and handling UI events to zoom and pan the plot.
Textual-plotext uses the plotext library which has more features than this library. However, it does not support interactive zooming or panning and the tick placement isn't as nice since it simply divides up the axes range into a fixed number of intervals giving values like 0, 123.4, 246.8, etc.
The performance can be much improved, but we're working on that. Next, we'll work on adding some features like date axes. This will (probably) not turn into a general do-it-all plotting library. We focus first on handling quantitative data in the context of physics experiments. If you'd like to see features added, do let us know. And if a PR is of good quality and is a good fit for the API, we'd love to handle more use cases beyond physics. And who knows, maybe this will turn into a general plotting library!