#state-machine #state-charts #fsm #xstate #model-based-testing

rustate

A Rust implementation of statecharts with model-based testing support

8 releases

new 0.2.6 Apr 22, 2025
0.2.5 Apr 22, 2025
0.1.0 Apr 18, 2025

#124 in Testing

Download history 517/week @ 2025-04-17

518 downloads per month

MIT license

200KB
3.5K SLoC

RuState

A Rust implementation of statecharts inspired by XState. RuState provides a type-safe way to model and implement finite state machines and statecharts in Rust, with full support for model-based testing (MBT).

Features

  • ✅ Finite state machines and statecharts
  • ✅ Hierarchical states
  • ✅ Parallel states
  • ✅ Guards/conditions for transitions
  • ✅ Actions/side effects
  • ✅ Context (extended state)
  • ✅ Typesafe API
  • ✅ Serializable machines
  • ✅ Model-based testing support
  • ✅ Cross-crate integration patterns
  • ✅ Code generation (JSON and Protocol Buffers)

Code Generation Features

RuState now supports code generation features to transform state machine definitions into different formats:

  1. JSON Export: Export statecharts to JSON format compatible with XState
  2. Protocol Buffers Export: Generate .proto files for cross-language compatibility
  3. Rust AST Analysis: Parse state machine definitions from Rust source code

Enable Code Generation

[dependencies]
# For JSON export only
rustate = { version = "0.2.5", features = ["codegen"] }

# For Protocol Buffers generation
rustate = { version = "0.2.5", features = ["proto"] }

Export Example

use rustate::{Machine, MachineBuilder, State, Transition, CodegenExt, JsonExportOptions, ProtoExportOptions};

// Create your state machine
let machine = create_my_machine()?;

// Export to JSON
let json_options = JsonExportOptions {
    output_path: "my_machine.json".to_string(),
    pretty: true,
    include_metadata: true,
};
machine.export_to_json(Some(json_options))?;

// Export to Protocol Buffers
let proto_options = ProtoExportOptions {
    output_path: "my_machine.proto".to_string(),
    package_name: "mymachines".to_string(), 
    message_name: "MyStateMachine".to_string(),
};
machine.export_to_proto(Some(proto_options))?;

Advanced Usage: Rust AST Analysis

With the codegen feature, RuState can analyze Rust source code to extract state machine definitions:

use rustate::{Machine, CodegenExt};

// Parse state machine definition from Rust source
let machine = Machine::parse_from_rust_file("src/my_machine.rs")?;
println!("Parsed machine name: {}", machine.name);

This feature enables powerful tooling for documentation generation, visual representation, and integration with other languages and frameworks.

Model-Based Testing Integration

RuState now includes comprehensive model-based testing features:

  1. Automated Test Generation: Generate test cases from your state machine model
  2. Test Execution: Run tests directly against your state machine or export them
  3. Coverage Analysis: Measure state and transition coverage
  4. Model Checking: Verify properties like reachability, safety, and liveness

Key MBT Components:

  • TestGenerator: Creates test cases for states, transitions, and loop coverage
  • TestRunner: Executes test cases against your machine
  • ModelChecker: Verifies model properties and detects deadlocks and unreachable states

Cross-Crate Integration Patterns

RuState provides patterns for integrating state machines across crates with type safety:

  1. Event Forwarding Pattern: Share state machine references and forward events between machines
  2. Context Sharing Pattern: Share context data between multiple state machines
  3. Hierarchical Integration Pattern: Connect parent-child state machines with traits

Enable with the integration feature:

[dependencies]
rustate = { version = "0.2.5", features = ["integration"] }

Integration Example

use rustate::{
    Machine, MachineBuilder, State, Transition,
    integration::{
        SharedMachineRef,
        SharedContext,
        ChildMachine,
    },
};

// Create and share a state machine
let machine = create_machine();
let shared_machine = SharedMachineRef::new(machine);

// Forward events
shared_machine.send_event("EVENT")?;

See the examples/integration directory for complete integration examples.

Integration Patterns in Detail

RuState provides three main integration patterns for connecting state machines across crate boundaries in a type-safe way:

1. Event Forwarding Pattern

This pattern allows state machines to communicate by sending events to each other. It's useful when you need to coordinate multiple state machines with minimal coupling.

use rustate::{Machine, MachineBuilder, State, Transition, Action, ActionType};
use rustate::integration::SharedMachineRef;

// Create a child state machine
let child_machine = MachineBuilder::new("child")
    .state(State::new("idle"))
    .state(State::new("active"))
    .initial("idle")
    .transition(Transition::new("idle", "ACTIVATE", "active"))
    .build()
    .unwrap();

// Create a shared reference
let shared_child = SharedMachineRef::new(child_machine);
let shared_child_clone = shared_child.clone();

// Parent machine action that forwards events to child
let forward_action = Action::new(
    "forwardToChild",
    ActionType::Transition,
    move |_ctx, evt| {
        if evt.event_type == "PARENT_EVENT" {
            let _ = shared_child_clone.send_event("ACTIVATE");
        }
    }
);

2. Context Sharing Pattern

This pattern allows multiple state machines to share data through a common context. It's ideal for scenarios where state machines need to access and modify shared state.

use rustate::{Machine, MachineBuilder, State, Action, ActionType};
use rustate::integration::SharedContext;

// Create shared context
let shared_context = SharedContext::new();
let context_for_a = shared_context.clone();
let context_for_b = shared_context.clone();

// Action that writes to shared context
let write_action = Action::new(
    "writeData",
    ActionType::Transition,
    move |_ctx, _evt| {
        let _ = context_for_a.set("status", "active");
    }
);

// Action that reads from shared context
let read_action = Action::new(
    "readData",
    ActionType::Transition,
    move |ctx, _evt| {
        if let Ok(Some(status)) = context_for_b.get::<String>("status") {
            let _ = ctx.set("localStatus", status);
        }
    }
);

3. Hierarchical Integration Pattern

This pattern establishes parent-child relationships between state machines using traits. It's powerful for complex systems where you need to model hierarchical relationships with high cohesion but low coupling.

use std::sync::{Arc, Mutex};
use rustate::{Machine, MachineBuilder, State, Transition};
use rustate::integration::hierarchical::{ChildMachine, DefaultChildMachine, coordination};

// Create a child state machine
let child_machine = MachineBuilder::new("child")
    .state(State::new("initial"))
    .state(State::new("running"))
    .state(State::new_final("complete"))
    .initial("initial")
    .transition(Transition::new("initial", "START", "running"))
    .transition(Transition::new("running", "COMPLETE", "complete"))
    .build()
    .unwrap();

// Wrap with trait implementation
let child = DefaultChildMachine::new(child_machine, "complete");
let child = Arc::new(Mutex::new(child));

// Create action that monitors child machine state
let monitor_action = coordination::create_child_monitor_action(
    "monitorChild",
    child.clone()
);

// Create action that forwards events to child machine
let forward_action = coordination::create_event_forwarder_action(
    "forwardToChild",
    child,
    "PARENT_START",  // Event received by parent
    "START"          // Event forwarded to child
);

Combining Integration Patterns

For complex systems, these patterns can be combined to create powerful integration strategies. See the examples/integration/combined_demo.rs for a complete example that demonstrates all three patterns working together.

For asynchronous integration capabilities, use the integration_async feature:

[dependencies]
rustate = { version = "0.2.4", features = ["integration_async"] }

Usage Example

Simple State Machine

use rustate::{Action, ActionType, Machine, MachineBuilder, State, Transition};

// Create states
let green = State::new("green");
let yellow = State::new("yellow");
let red = State::new("red");

// Create transitions
let green_to_yellow = Transition::new("green", "TIMER", "yellow");
let yellow_to_red = Transition::new("yellow", "TIMER", "red");
let red_to_green = Transition::new("red", "TIMER", "green");

// Define actions
let log_green = Action::new(
    "logGreen",
    ActionType::Entry,
    |_ctx, _evt| println!("Entering GREEN state - Go!"),
);

// Build the machine
let mut machine = MachineBuilder::new("trafficLight")
    .state(green)
    .state(yellow)
    .state(red)
    .initial("green")
    .transition(green_to_yellow)
    .transition(yellow_to_red)
    .transition(red_to_green)
    .on_entry("green", log_green)
    .build()
    .unwrap();

// Send an event to the machine
machine.send("TIMER").unwrap();

Model-Based Testing Example

use rustate::{Machine, TestGenerator, TestRunner, ModelChecker, Property, PropertyType};

// Assuming you have a state machine defined as above...
let machine = /* ... */;

// Generate test cases
let mut generator = TestGenerator::new(&machine);
let test_cases = generator.generate_all_transitions();

// Run tests
let mut runner = TestRunner::new(&machine);
let results = runner.run_tests(test_cases);
println!("Test success rate: {}%", results.success_rate());

// Coverage analysis
let coverage = results.get_coverage();
println!("State coverage: {}%", coverage.state_coverage());
println!("Transition coverage: {}%", coverage.transition_coverage());

// Model checking
let mut checker = ModelChecker::new(&machine);

// Define property to check
let property = Property {
    name: "Can reach red state".to_string(),
    property_type: PropertyType::Reachability,
    target_states: vec!["red".to_string()],
    description: None,
};

// Verify the property
let verification = checker.verify_property(&property);
if verification.satisfied {
    println!("Property satisfied: {}", property.name);
} else {
    println!("Property not satisfied: {}", property.name);
    if let Some(counterexample) = verification.counterexample {
        println!("Counterexample found with {} events", counterexample.len());
    }
}

// Detect deadlocks
let deadlocks = checker.detect_deadlocks();
println!("Deadlock states found: {}", deadlocks.len());

See the examples directory for complete examples.

Network and Remote Integration

For network-based state machine control and monitoring, check out rustate-grpc, which provides:

  • Remote state machine creation and control via gRPC
  • Real-time state change monitoring via streaming
  • Type-safe client/server communication
  • Cross-language support through protocol buffers
[dependencies]
rustate-grpc = { version = "0.1.0", features = ["full"] }

See the rustate-grpc documentation for detailed usage examples.

Installation

Add this to your Cargo.toml:

[dependencies]
rustate = "0.2.5"

Documentation

Core Concepts

  • State: Represents a node in the state chart
  • Transition: Defines how the machine moves between states in response to events
  • Guard: Conditional logic that determines if a transition should occur
  • Action: Side effects that execute during state transitions
  • Context: Stores extended state for the machine
  • TestGenerator: Creates test cases from your state machine model
  • TestRunner: Executes test cases against your machine
  • ModelChecker: Verifies properties and analyzes your state machine model

API Overview

  • State: Create simple, compound, parallel, or history states
  • Transition: Define transitions between states, including guards and actions
  • Guard: Create guard conditions for transitions
  • Action: Define actions/side effects for state transitions
  • Context: Store and retrieve extended state data
  • Machine: The runtime state machine instance
  • MachineBuilder: Fluent API for creating state machines
  • TestGenerator: Generate test cases from a state machine model
  • TestRunner: Run tests against your state machine
  • ModelChecker: Verify properties and analyze your state machine model

Future Directions

  • Advanced model checking algorithms
  • Property-based testing integration
  • Test visualization tools
  • Fuzzing-based MBT
  • Temporal logic (LTL/CTL) property specification and verification
  • Performance optimizations for large state machines
  • ✅ Distributed system state machine coordination
  • Enhanced WebAssembly (WASM) support
  • Integration with visual state machine editors
  • Automatic state machine model generation from existing systems
  • Advanced concurrency model support
  • Domain-specific language (DSL) for state machine definition
  • Microcontroller-optimized version

License

MIT

テスト機能

RuStateには複数のテスト支援機能があります:

モデルベーステスト (MBT)

状態マシンのモデルに基づいて自動的にテストケースを生成します。

let machine = create_test_machine();
let mut generator = TestGenerator::new(&machine);

// 状態カバレッジのテストケースを生成
let state_tests = generator.generate_all_states();

// 遷移カバレッジのテストケースを生成
let transition_tests = generator.generate_all_transitions();

プロパティベーステスト

proptest との統合により、ランダムなイベントシーケンスでプロパティを検証します。

let property = Machine::property("state property")
    .given(|m| m.is_in("idle"))
    .when(|m| {
        let _ = m.send("START");
        Ok(m.current_state().clone())
    })
    .then(|m| m.is_in("running"));

let runner = PropertyTestRunner::new(machine);
let result = runner.verify_property(property, Config::default());

XState互換モデルベーステスト (新機能)

XState v5互換のモデルベーステストインターフェースを提供します。

// ステートマシンからテストモデルを作成
let mut model = create_test_model(machine);

// アサーションを追加
model.assert("conditionName", |m| {
    // 条件をチェック
    true
});

// アクターの実装を提供
model.provide("actorName", |ctx, evt| {
    // アクターの実装
    Ok(())
});

// テストプランを作成(手動または自動生成)
let plan = XStateTestPlan {
    name: "Test Plan",
    paths: vec![
        XStateTestPath {
            name: "Path 1",
            segments: vec![
                XStatePathSegment {
                    state: "idle",
                    event: Some("START"),
                    assertions: None,
                },
                // ...
            ],
            description: Some("Test path description"),
        },
    ],
    // ...
};

// または自動生成
let generated_plan = model.generate_paths(max_depth);

// プランを実行
let results = execute_test_plan(&mut model, &plan)?;

モデル検査

状態マシンの性質(到達可能性、安全性など)を検証します。

let mut checker = ModelChecker::new(&machine);

// 到達可能性プロパティの検証
let reachability = Property {
    name: "Can reach completed".to_string(),
    property_type: PropertyType::Reachability,
    target_states: vec!["completed".to_string()],
    description: None,
};

let result = checker.verify_property(&reachability);

Dependencies

~0.9–16MB
~212K SLoC