Crate ringkernel_procint

Crate ringkernel_procint 

Source
Expand description

§RingKernel Process Intelligence

GPU-accelerated process mining and intelligence with real-time visualization.

This crate provides tools for analyzing business processes through:

  • DFG Construction: Build Directly-Follows Graphs from event streams
  • Partial Order Mining: Discover concurrent activity patterns
  • Pattern Detection: Identify bottlenecks, loops, rework, and anomalies
  • Conformance Checking: Validate traces against reference models

§Architecture

┌─────────────────┐     ┌──────────────────┐     ┌────────────────┐
│   Data Fabric   │────▶│   GPU Kernels    │────▶│  Visualization │
│ (Event Stream)  │     │ (CUDA/WGSL)      │     │  (egui Canvas) │
└─────────────────┘     └──────────────────┘     └────────────────┘
        │                        │                       │
        ▼                        ▼                       ▼
┌─────────────────┐     ┌──────────────────┐     ┌────────────────┐
│ Sector Templates│     │ DFG Construction │     │ Force Layout   │
│ Event Generator │     │ Pattern Detection│     │ Token Animation│
│ Anomaly Inject  │     │ Conformance Check│     │ Timeline View  │
└─────────────────┘     └──────────────────┘     └────────────────┘

§Quick Start

ⓘ
use ringkernel_procint::prelude::*;

// Create event generator for healthcare sector
let config = GeneratorConfig::default()
    .with_sector(SectorTemplate::Healthcare)
    .with_events_per_second(10_000);
let mut generator = ProcessEventGenerator::new(config);

// Generate events and build DFG
let events = generator.generate_batch(1000);
let dfg = DFGBuilder::new().build_from_events(&events);

// Detect patterns
let patterns = PatternDetector::new().detect(&dfg);

§GPU Kernel Types

  1. DFG Construction - Batch kernel for edge frequency counting
  2. Partial Order Derivation - Stencil kernel for precedence matrix
  3. Pattern Detection - Batch kernel for bottleneck/loop/rework detection
  4. Conformance Checking - Batch kernel for fitness scoring

Modules§

actors
Actor runtime for GPU kernel coordination.
analytics
Analytics engines for process intelligence.
cuda
CUDA kernel code generation and execution.
fabric
Data fabric for synthetic process event generation.
gui
GUI components for process intelligence visualization.
kernels
GPU kernel implementations for process intelligence.
models
Core data structures for process intelligence.
prelude
Prelude for convenient imports.

Constants§

NAME
Crate name.
VERSION
Version information.