Skip to content

Visualization verbs

Visualization verbs display cubes inline or generate QA plots without leaving the pipe chain.

v.show_cube_lexcube(**kwargs)

Integrate Lexcube for interactive (time, y, x) exploration. The verb displays the widget as a side effect and returns the original cube so pipelines keep flowing.

from cubedynamics import pipe, verbs as v

pipe(cube) \
    | v.month_filter([6, 7, 8]) \
    | v.show_cube_lexcube(title="PRISM JJA precipitation", cmap="RdBu_r")
  • Parameters: pass any Lexcube keyword arguments (title, cmap, vmin, vmax). Datasets with a single data variable are automatically converted to a DataArray.
  • Notes: Widgets render only in live notebook environments and require a 3D (time, y, x) cube. Reducers such as v.mean/v.variance should use keep_dim=True to preserve this layout.

For QA plots outside Lexcube, call the functional helper cubedynamics.plot_median_over_space(cube, ...).

Update (2025): New v.plot cube viewer

verbs.plot(obj, kind="auto", **kwargs)

Generic plotting verb.

  • kind="auto" (default):
  • If obj is 3D (time,y,x), may dispatch to the cube viewer or a slice-based plot, depending on configuration.

  • For 2D arrays, uses the existing 2D plotting (Matplotlib) behavior.

  • kind="cube":

  • Uses the HTML cube viewer with streaming, non-blocking progress text and Lexcube-style axis labels.

  • Parameters forwarded:

    • out_html, cmap, size_px, thin_time_factor.

  • kind="slice":

  • Old behavior: 2D map with time slider (ipywidgets + Matplotlib), if still kept for backward compatibility.

verbs.map(obj, time_index=None, **kwargs)

Map-style visualization of (y,x) or (time,y,x) DataArrays using a MapGL-style engine (pydeck).

Parameters:

  • obj: piped object or xarray DataArray/Dataset.
  • time_index: optional index along the time dimension for 3D inputs (defaults to the last timestep).
  • **kwargs: forwarded to vis_map.show_map_pydeck, including:
  • cmap, vmin, vmax: color scaling.
  • height, width: canvas size.

Returns a pydeck Deck object displayed inline in Jupyter.

v.quick_map() (planned)

Future work will expose small multiples and static PNG exporters for dashboards. Track development in the Roadmap.