Tracking: trace types not yet exposed in plotly.rs (plotly.js 3.7.0 parity)

Tracking: trace types not yet exposed in plotly.rs (plotly.js 3.7.0 parity)

Summary

Measured against the actively-supported trace types in plotly.js 3.7.0, plotly.rs exposes 27 of 46, leaving 19 unimplemented. This issue tracks the remaining trace types so contributions can be coordinated and duplicate work avoided.

plotly.js 3.7.0 defines 49 trace types, but 3 are deprecated (the Mapbox family — see below) and are excluded from the parity target. The list was generated by diffing the trace type tags plotly.rs can serialize (the PlotType enum) against the authoritative plot-schema.json shipped with plotly.js v3.7.0.

Note: this is trace-type coverage. Attribute-level parity for the already-implemented traces is being addressed separately in the plotly.js 3.7 attribute backfill work.

Deprecated in plotly.js 3.7.0 — excluded from parity (3)

These are deprecated upstream in favour of the MapLibre map subplot traces (migration guide):

Deprecated trace	Recommended replacement	plotly.rs status
scattermapbox	scattermap	both already exposed
densitymapbox	densitymap	both already exposed
choroplethmapbox	choroplethmap	replacement already exposed; deprecated trace not implemented (and won't be)

plotly.rs already ships every recommended map replacement, so there is no coverage gap here — no action needed. The two deprecated traces that are implemented (scattermapbox, densitymapbox) can be retained for back-compat and eventually deprecated in lockstep with upstream.

Currently exposed — actively-supported (27)

bar · box · candlestick · choropleth · choroplethmap · contour · densitymap · heatmap · histogram · histogram2dcontour · image · mesh3d · ohlc · pie · sankey · scatter · scatter3d · scattergeo · scattergl · scattermap · scatterpolar · scatterpolargl · sunburst · surface · table · treemap · violin

(plus the deprecated scattermapbox and densitymapbox, listed above)

Missing — actively-supported (19)

3D scientific / volumetric

• cone — 3D vector / quiver fields
• streamtube — 3D flow streamlines
• isosurface — 3D isosurfaces from volume data
• volume — 3D volume rendering

Carpet family (needs a shared carpet-axis abstraction)

• carpet — base carpet axis system
• contourcarpet — contours on a carpet axis
• scattercarpet — scatter on a carpet axis

Cartesian statistical / financial

• histogram2d — 2D histogram heatmap
• funnel — funnel charts
• waterfall — waterfall / bridge charts

Parallel-coordinate

• parcoords — parallel coordinates (continuous)
• parcats — parallel categories (categorical)

Specialized subplots

• barpolar — bar charts on a polar axis (wind-rose)
• scatterternary — scatter on a ternary (3-component) plot
• scattersmith — scatter on a Smith chart (RF / impedance)

Hierarchical

• icicle — icicle charts (sibling of sunburst / treemap)

Other

• indicator — KPI gauges / number + delta displays
• splom — scatter-plot matrix
• funnelarea — funnel-area (pie-like funnel)

Suggested prioritization

• Cheap wins — close siblings of existing traces that can reuse most of their building blocks: histogram2d (cf. histogram2dcontour / heatmap) and icicle (cf. sunburst / treemap).
• High user demand — common business-chart types: indicator, funnel, waterfall, barpolar.
• Larger effort — the carpet family (carpet + contourcarpet + scattercarpet) depends on first modelling a shared carpet-axis abstraction.

Notes for implementers

Adding a trace follows the established pattern (see CONTRIBUTING.md / CLAUDE.md):

1. Create plotly/src/traces/<name>.rs, derive FieldSetter with #[field_setter(box_self, kind = "trace")].
2. Add the corresponding PlotType variant (with the correct serde rename).
3. Implement Trace::to_json, then re-export from traces/mod.rs and lib.rs.
4. Add a serialize round-trip test and a CHANGELOG.md entry.

Commencez par CONTRIBUTING.md et CLAUDE.md, puis examinez un trace existant dans plotly/src/traces/ ainsi que l’enum PlotType. Choisissez un type de trace non coché et ajoutez son module de trace, la variante de l’enum, les exports, un test de round-trip de sérialisation et une entrée dans CHANGELOG.md en suivant le modèle établi. C’est terminé lorsque le trace sélectionné est exposé et se sérialise correctement.