Spectra Feature Guide

Core Rendering

• Vulkan-powered rendering — Explicit GPU control, multi-buffered frames, async buffer uploads, depth buffer, MSAA up to 4x
• Anti-aliased lines — Screen-space quad expansion with SDF smoothing, resolution-independent
• 18 marker types — SDF-based markers (circle, square, diamond, triangles, pentagon, hexagon, star, plus, cross, filled variants)
• Dash patterns — GPU-accelerated dash rendering with 8 customizable dash values
• GPU text rendering — Vulkan text pipeline with stb_truetype atlas (Inter font), depth-tested 3D labels, ImGui fallback
• MATLAB-style plot API — Format strings like "r--o", "b:*" for quick styling, plus PlotStyle struct for full control

3D Visualization

Easy API (MATLAB-style)

• Progressive complexity — From one-liners to full control in 7 levels
• 2D functions — plot(), scatter(), subplot(), title(), xlabel(), ylabel(), xlim(), ylim(), grid(), legend()
• Statistical plots — box_plot(), violin(), histogram(), bar() with filled shapes, semi-transparent fills, auto-computed statistics
• 3D functions — plot3(), scatter3(), surf(), mesh(), subplot3d(), zlabel(), zlim()
• Multi-window — figure() creates new OS windows, tab() adds tabs
• Real-time — on_update(callback) with configurable FPS
• State accessors — gcf(), gca(), gca3d(), cla() for MATLAB-style workflows

Animation & Recording

UI & Productivity

• Command palette — Fuzzy search (Ctrl+K), 30+ registered commands, recent tracking, category headers
• Configurable shortcuts — Rebindable keybindings with JSON persistence (~/.config/spectra/keybindings.json)
• Undo/redo system — Full property change history with grouped operations, undoable property helpers
• Multi-figure tabs — Tab switching (Ctrl+Tab), duplication, context menus (rename, close others, close to right), per-figure state
• Docking & split view — Horizontal / vertical splits (Ctrl+\\), drag-to-dock with edge detection, splitter handles with grip dots, max 8 panes
• Inspector panel — Series statistics (min / max / mean / median / std / percentiles / count), sparkline preview, style editing, axes aggregate stats
• CSV data loading — File→Data→Load from CSV with column picker, delimiter auto-detection, data preview
• Workspace management — Save / load full state (v4 format with 3D support), autosave, backward compat (v1–v3)
• Plugin architecture — C ABI for binary-compatible extensions, PluginManager with dlopen / LoadLibrary, auto-discovery

Data Interaction

• Hover tooltips — Nearest-point query with series name, coordinates, color swatch
• Crosshair overlay — Shared across linked subplots with Y-value interpolation from series data
• Data markers — Pin / remove persistent markers that survive zoom and pan
• Region selection — Shift-drag rectangular selection with statistics mini-toolbar
• Legend interaction — Click-to-toggle visibility, drag-to-reposition, animated opacity
• Multi-axis linking — Link X / Y / Both axes across subplots, synchronized zoom / pan / box-zoom / auto-fit
• Shared cursor — Linked subplots show interpolated crosshair values from source axes
• Data transforms — 14 built-in types: Identity, Log10, Ln, Abs, Negate, Normalize, Standardize, Derivative, CumulativeSum, Diff, Scale, Offset, Clamp, Custom — with pipeline chaining and transform registry

Themes & Accessibility

• Dark / light themes — Smooth animated transitions (non-mutating display colors)
• 8 colorblind-safe palettes — Okabe-Ito, Tol Bright / Muted, IBM, Wong, Viridis, Monochrome
• CVD simulation — Test designs for Protanopia, Deuteranopia, Tritanopia, Achromatopsia (Vienot / Brettel matrices)
• Theme export / import — JSON-based theme customization with save / load
• Design tokens — Consistent spacing, typography, color system
• WCAG contrast checking — Color utility methods (luminance, contrast ratio, sRGB / linear / HSL conversions)

Multi-Window & Multi-Process

• Multi-window support — Independent OS windows with per-window Vulkan swapchain, no "primary window" concept
• Tab tear-off — Detach tabs into new windows, in-process window creation
• Dual-mode architecture — Runtime selection between inproc (single-process) and multiproc (daemon + agents) modes
• Window manager — Uniform window lifecycle, figure-to-window assignment, per-window resize handling
• FigureRegistry — Stable FigureId-based figure ownership (no positional indexing)

Python API

Python Examples

# Session API
import spectra as sp
import numpy as np

s = sp.Session()
fig = s.figure("My Plot")
ax = fig.subplot(1, 1, 1)

x = np.linspace(0, 10, 1000)
line = ax.line(x, np.sin(x), label="sin(x)")
ax.set_title("Trigonometry")
s.show()

# Statistical Plots
import spectra as sp
import numpy as np

data = [np.random.normal(mu, 1, 200) for mu in [0, 2, 5]]

sp.subplot(1, 2, 1)
sp.boxplot(data, positions=[1, 2, 3])
sp.title("Box Plot")

sp.subplot(1, 2, 2)
sp.violin_plot(data, positions=[1, 2, 3])
sp.title("Violin Plot")
sp.show()

# Live Streaming
import spectra as sp
import numpy as np

fig = sp.figure("Live Dashboard")
ax = fig.subplot(1, 1, 1)
line = ax.line([], [], label="sensor")

def update(dt, t):
    line.append(t, np.sin(t) + np.random.normal(0, 0.1))
    ax.set_xlim(max(0, t - 10), t)

sp.live(update, fps=30)
sp.show()

Export & Integration

• Headless mode — Offscreen rendering without a window (CI, servers, batch export)
• PNG export — Render to image via stb_image_write
• GIF export — Custom GIF89a writer with LZW encoding and median-cut quantization
• MP4 export — Frame piping to ffmpeg (behind SPECTRA_USE_FFMPEG flag)
• Subplot layouts — Grid-based multi-axes figures with automatic margin / tick layout
• Zero-copy data — std::span<const float> interfaces avoid unnecessary copies
• CMake find_package — find_package(spectra) + target_link_libraries
• Optional Eigen support — Adapters for Eigen::VectorXf behind a feature flag

Roadmap

UI Redesign (Complete)

Phase	Status	Highlights
Phase 1 — Modern Foundation	✅	Theme system, layout, inspector, animated interactions, tooltips, crosshair, transition engine
Phase 2 — Power User Features	✅	Command palette, undo/redo, workspace v2, timeline, recording, colorblind palettes, tabs
Phase 3 — Elite Differentiators	✅	Docking/split, MATLAB API (18 markers, dashes), axis linking, shared cursor, transforms, plugins

3D Visualization (Complete)

Phase	Status	Highlights
3D Phase 1	✅	Math library, FrameUBO, depth buffer, 3D pipelines, Camera, Axes3D
3D Phase 2	✅	Scatter3D, Line3D, Surface, Mesh, colormaps, camera animation
3D Phase 3	✅	Blinn-Phong, transparency, MSAA, wireframe, materials, 2D↔3D transition, workspace v4

Multi-Window & IPC (Complete)

Milestone	Status	Highlights
WindowContext extraction	✅	Per-window Vulkan resources, no primary window assumption
Multi-window rendering	✅	Tab tear-off, per-window swapchain, figure-to-window assignment
FigureRegistry	✅	Stable FigureId ownership, replaces positional indexing
Python IPC	✅	20+ message types, Session/Figure/Axes/Series proxies, easy API, live streaming
Multiproc daemon	✅	spectra-backend + spectra-window agent architecture

Deployment (Complete)

Milestone	Status	Highlights
Cross-platform CI	✅	Linux (GCC+Clang), macOS (ARM), Windows (MSVC), sanitizers
Release pipeline	✅	.deb, .rpm, .tar.gz, AppImage, .zip, Python wheels, PyPI
Docker	✅	Multi-stage build, headless + X11 forwarding
Packaging	✅	CPack, AppImage, Homebrew, AUR, shell completions

Recommended Paths