MatPlotLibNet 1.3.0

MatPlotLibNet

A .NET 10 / .NET 8 charting library inspired by matplotlib. Fluent API, dependency injection, parallel SVG rendering, polymorphic export (SVG / PNG / PDF / GIF), and multi-platform output to Blazor, MAUI, Avalonia, Uno Platform, ASP.NET Core, Angular, React, and Vue.

v1.3.0 — Cross-platform native UI controls + MathText completion + 3-D round 2. Three headline features:

1. Native controls — MplChartControl (Avalonia 12) and MplChartElement (Uno Platform) render charts natively via SkiaSharp — no browser, no WebView, no SignalR. IsInteractive="True" enables local pan / zoom / reset / brush-select with rubber-band overlay, legend-toggle hit-testing, and hover tooltips via NearestPointFinder. .WithServerInteraction(hubConnection) bridges native controls to a SignalR hub. Two new NuGet packages: MatPlotLibNet.Avalonia and MatPlotLibNet.Uno.
2. MathText completion — \frac{a}{b}, \sqrt{x}, \sqrt[n]{x}, accents (\hat, \bar, \vec, \tilde, \dot), font variants (\mathrm, \mathbf, \mathit, \mathcal, \mathbb), \text{}, spacing (\,, \:, \;, \quad), scaling delimiters (\left(...\right)), and 45+ new symbol mappings (blackboard bold, arrows, relations, set operators).
3. 3-D round 2 — six new series: Line3D, Trisurf3D (Delaunay), Contour3D (marching squares), Quiver3D (vector field), Voxels (face-culled cubes), Text3D (3D annotations). Series count: 61 → 67.

3 842 tests green across 9 test projects. The managed interaction layer in core (six IInteractionModifier implementations, InteractionController, ChartLayout) is shared between desktop controls and SignalR — one vocabulary, two transports.

v1.2.2 — Brush-select + hover round-trip (deferred v1.2.0 items). v1.2.0 shipped four mutation events (Zoom, Pan, Reset, LegendToggle) that rewrite the authoritative Figure and broadcast the updated SVG. v1.2.2 introduces the first two notification events — BrushSelectEvent and HoverEvent — that observe the user's gesture, route it to a per-chart handler in .NET code, and (for hover) return a caller-only response. The architectural move: a new tier-2 abstract record FigureNotificationEvent with sealed override ApplyTo makes "a notification event accidentally mutates the figure" structurally impossible. Brush-select is fire-and-forget (ChartSessionOptions.OnBrushSelect); hover is request-response returning HTML, delivered to the originating client only via IChartHubClient.ReceiveTooltipContent — the first per-caller mechanism in the library (v1.2.0 only had group broadcast). Backward compatible: v1.2.0/v1.2.1 code paths untouched. 3 519 passing across 7 test projects (+23 core, +10 AspNetCore including 4 real-SignalR round-trip tests with two connected clients verifying caller-only responses).

v1.2.1 — Font-factory subsystem fix + zero-warning CI sweep. The outside-legend clipping bug that v1.1.4's changelog claimed to have fixed turned out to have a deeper root cause: eight duplicate themed-font factories across AxesRenderer, ChartRenderer, ConstrainedLayoutEngine, and LegendMeasurer had silently drifted (engine used DefaultFont.Size − 2, renderer used DefaultFont.Size). v1.2.1 consolidates them into one ThemedFontProvider — a single source of truth that makes the drift bug class structurally impossible. Seven new drift-regression tests cover every measurer/renderer pair. Every non-MAUI library now builds at 0 warnings / 0 errors (21 warnings cleared, 5 stale Color.Blue/Orange references fixed in WebApi + GraphQL samples, 16 xUnit1051 warnings refactored to TestContext.Current.CancellationToken). 3 641 tests green across 7 test projects.

v1.2.0 — Bidirectional SignalR interactive charts. Browser wheel-zoom, drag-pan, reset, and legend-toggle round-trip through ChartHub to a server-authoritative Figure that is mutated on a per-chart channel-drained background task and re-published through the existing SignalR fan-out. Pure .NET, no JavaScript charting library, server stays the source of truth. See MatPlotLibNet.AspNetCore README or the Samples/MatPlotLibNet.Samples.AspNetCore + Interactive.razor demos.

Packages

Quick start

using MatPlotLibNet;
using MatPlotLibNet.Styling;

double[] x = [1, 2, 3, 4, 5];
double[] y = [2, 4, 3, 5, 1];

Plt.Create()
    .WithTitle("My First Chart")
    .WithTheme(Theme.Dark)
    .Plot(x, y, s => { s.Color = Color.Blue; s.Label = "Data"; })
    .WithLegend()
    .Save("chart.svg");

67 series types — line, scatter, bar, histogram, pie, box, violin, heatmap, contour, candlestick, OHLC, treemap, sunburst, Sankey, polar, polar heatmap, 3D surface, Bar3D, PlanarBar3D, Line3D, Trisurf3D, Contour3D, Quiver3D, Voxels, Text3D, radar, waterfall, funnel, gauge, and more.

Native UI controls (v1.3.0) — MplChartControl for Avalonia 12 and MplChartElement for Uno Platform render charts natively via SkiaSharp — no browser, no WebView, no SignalR required. Set IsInteractive="True" for local pan / zoom / reset / brush-select using the managed interaction layer (InteractionController, six IInteractionModifier implementations, ChartLayout coordinate transform). Connect to a SignalR hub with .WithServerInteraction(hubConnection) for server-authoritative mode.

MathText — LaTeX-like inline math in any label or title: $\alpha^{2}$, $\frac{a}{b}$, $\sqrt{x}$, $\hat{x}$, $\mathbf{F}$, $\mathbb{R}$. Supports Greek letters, super/subscript, fractions, square roots (with optional index), accents (\hat, \bar, \vec, \tilde, \dot), font variants (\mathrm, \mathbf, \mathit, \mathcal, \mathbb), \text{} for roman text inside math, spacing (\,, \:, \;, \quad), scaling delimiters (\left(...\right)), and 90+ symbol mappings.

3-D charts — six series types added in v1.3.0: Line3D (projected polyline), Trisurf3D (Delaunay triangulated surface), Contour3D (marching-squares contour lines projected onto 3D planes), Quiver3D (3D vector field with arrow mesh), Voxels (volumetric cube rendering with face culling), Text3D (3D annotations). All share the existing Projection3D pipeline, DepthQueue3D painter's algorithm, and Svg3DRotationScript client-side rotation.

Bidirectional SignalR (v1.2.0, extended in v1.2.2) — FigureBuilder.WithServerInteraction("chart-id", i => i.All()) opts a figure into server-authoritative interaction. The embedded dispatcher script invokes ChartHub.OnZoom / OnPan / OnReset / OnLegendToggle (mutation events, v1.2.0) and OnBrushSelect / OnHover (notification events, v1.2.2). Mutation events flow through a stacked-record hierarchy (ZoomEvent : AxisRangeEvent : FigureInteractionEvent) to a FigureRegistry-owned background reader task that rewrites the figure and re-publishes the SVG; notification events stack under FigureNotificationEvent (tier-2, sealed override ApplyTo no-op) and route to per-chart handlers via ChartSessionOptions.OnBrushSelect / OnHover. Hover returns HTML to the originating client only via IChartHubClient.ReceiveTooltipContent. Natural coalescing — bursts of wheel or mousemove events produce exactly one re-render / one response per batch. First-class Blazor + raw-HTML samples ship in Samples/.

Accessibility — SVG exports carry role="img", <title>/<desc>, and ARIA labels on all structural groups; all 5 interactive JS features are keyboard-navigable; Okabe-Ito color-blind safe palette (Theme.ColorBlindSafe); WCAG AAA high-contrast theme (Theme.HighContrast).

Matplotlib look-alike themes — Theme.MatplotlibClassic mimics matplotlib's pre-2.0 default (white background, the iconic bgrcmyk cycle, DejaVu Sans 12pt). Theme.MatplotlibV2 mimics the modern matplotlib default since 2017 (white background, soft-black text, the tab10 10-color cycle, DejaVu Sans 10pt). Drop-in matplotlib look in pure .NET — no Python runtime required.

Series capability interfaces — IHasColor, IHasAlpha, IHasEdgeColor, ILabelable allow polymorphic access to common series properties without casting; enables generic theming and rendering utilities.

NumPy-style numerics — Mat matrix type with SIMD operators, Linalg (Solve/Inv/Det/Eigh/Svd), NpStats (Diff/Median/Histogram/Argsort/Unique/Cov/Corrcoef), NpRandom (Normal/Uniform/Lognormal/Integers), and Fft.Inverse/Frequencies/Shift — all zero new dependencies, pure C# + TensorPrimitives.

Broken / discontinuous axis — AxisBreak + BreakStyle (Zigzag, Straight, None); WithXBreak / WithYBreak on the fluent builder; visual markers drawn at break boundaries.

DataFrame indicator + regression bridges — MatPlotLibNet.DataFrame now includes DataFrameIndicatorExtensions (16 methods: SMA, EMA, RSI, Bollinger, OBV, MACD, DrawDown, ADX, ATR, CCI, WilliamsR, Stochastic, ParabolicSar, KeltnerChannels, VWAP) and DataFrameNumericsExtensions (PolyFit, PolyEval, ConfidenceBand) — compute indicators and polynomial regressions directly from named DataFrame columns.

Documentation

Full documentation is on the GitHub Wiki:

• Getting Started — installation, output formats, subplots
• Fluent Cheatsheet — one-page reference for Plt / FigureBuilder / AxesBuilder
• Package Map — all 12 NuGet + 3 npm packages in detail
• DataFrame — charting, 16 indicators, and polynomial regression from Microsoft.Data.Analysis.DataFrame
• Notebooks — Polyglot Notebooks + Jupyter inline rendering
• Chart Types — all 67 series with code examples
• Styling — themes, colormaps, PropCycler
• Matplotlib Themes — Theme.MatplotlibClassic and Theme.MatplotlibV2 look-alikes
• Accessibility — SVG semantics, keyboard navigation, color-blind palette, high-contrast theme
• Interactive Controls — Avalonia + Uno native controls, managed interaction layer, server mode adapter
• Bidirectional SignalR — server-authoritative interactive charts, event hierarchy, FigureRegistry, hub wiring
• Advanced — date axes, math text, animations, GIF, real-time
• Benchmarks — SVG rendering, SIMD transforms, indicators
• Roadmap — version history and planned phases
• Contributing — build, test, coding conventions

License

MIT — free for any use, open-source or commercial, with no copyleft conditions.