extending-flowcraft

name extending-flowcraft description Customize Flowcraft runtime behavior with pluggable components. Covers serializers, evaluators, loggers, and orchestrators. Use when the user needs custom serialization, complex data types, edge condition expressions, custom logging, or alternative execution strategies. Extending Flowcraft The FlowRuntime can be configured with pluggable components to tailor behavior to specific needs. Extensibility Points Component Interface Purpose See Serializer ISerializer Serialize/deserialize workflow context serializers.md Evaluator IEvaluator Evaluate edge conditions and transforms evaluators.md Logger ILogger Custom logging infrastructure loggers.md Orchestrator IOrchestrator Alternative execution strategies orchestrators.md Quick Examples Custom Serializer import { ISerializer , FlowRuntime } from 'flowcraft' import superjson from 'superjson' class SuperJsonSerializer implements ISerializer { serialize ( data : Record < string , any >): string { return superjson. stringify (data) } deserialize ( text : string ): Record < string , any > { return superjson. parse (text) as Record < string , any > } } const runtime = new FlowRuntime ({ serializer : new SuperJsonSerializer () }) Custom Evaluator import { IEvaluator , FlowRuntime } from 'flowcraft' import jsep from 'jsep' class JsepEvaluator implements IEvaluator { evaluate ( expression : string , context : Record < string , any >): any { const ast = jsep (expression) return interpretAst (ast, context) // Your AST interpreter } } const runtime = new FlowRuntime ({ evaluator : new JsepEvaluator () }) Custom Logger import { ILogger , FlowRuntime } from 'flowcraft' class FileLogger implements ILogger { debug ( msg : string , meta ?: Record < string , any > ) { /* ... */ } info ( msg : string , meta ?: Record < string , any > ) { /* ... */ } warn ( msg : string , meta ?: Record < string , any > ) { /* ... */ } error ( msg : string , meta ?: Record < string , any > ) { /* ... */ } } const runtime = new FlowRuntime ({ logger : new FileLogger ( 'workflow.log' ) }) Custom Orchestrator import { IOrchestrator , FlowRuntime , StepByStepOrchestrator } from 'flowcraft' // Use step-by-step for debugging const runtime = new FlowRuntime ({ orchestrator : new StepByStepOrchestrator (), }) Advanced topics Serializers : Handle complex types like Date, Map, Set — see serializers.md Evaluators : Property access vs full expressions — see evaluators.md Loggers : Integrate with Winston, Pino, etc. — see loggers.md Orchestrators : Default, step-by-step, event-driven, resumption — see orchestrators.md