codeplots

Tiny matplotlib shortcuts for plotting from VS Code, especially over SSH. It handles conversions such as .detach().cpu().numpy() for you and keeps a timestamped plot history when an interactive window is not practical.

import codeplots as cp

cp.plot([1, 3, 2, 5])          # auto line plot
cp.plot(image_array)           # auto image plot
cp.scatter(x, y)
cp.hist(values)
cp.bar({"a": 3, "b": 7})
cp.heatmap(matrix)

All helpers return a matplotlib Figure. Plots are saved automatically to .plots/ with timestamped names like 260428_193012_123456.svg, then opened in VS Code with code <file>.

fig = cp.plot(values)
cp.plot(values, save=False)      # skip autosave
cp.plot(values, prefix="loss")   # saves as .plots/loss_260428_193012_123456.svg

Pass filename="custom.svg" when you want an exact filename; it takes precedence over prefix. If it has no extension, .svg is added.

Shortcuts

plot, show, line, scatter, hist, bar, image, images, heatmap, box, violin, pie, loss, compare

Install

pip install codeplots