workflow-patterns

Workflow Patterns

A workflow engine needs to demonstrate fundamental control-flow patterns: sequence, parallel split, synchronization, exclusive choice, and merge. Each pattern is a building block that more complex workflows combine. Understanding them individually clarifies how the engine executes branching and joining logic.

Pipeline

[wp_chain_step]
     |
     v
     +─────────────────────────────+
     | [wp_split_a] | [wp_split_b] |
     +─────────────────────────────+
     [join]
     |
     v
[wp_merge_results]
     |
     v
     +── loop ──────────────+
     |  [wp_loop_iteration]
     +───────────────────────+

Workflow inputs: inputData, iterations

Workers

WpChainStepWorker (task: wp_chain_step)

Chain step: sequential processing in a chain pattern.

• Sets result = "chain_processed"
• Reads data. Writes result

WpLoopIterationWorker (task: wp_loop_iteration)

Loop iteration: processes one iteration in a DO_WHILE loop.

• Reads iteration. Writes iteration, processed

WpMergeResultsWorker (task: wp_merge_results)

Merge results from fork branches A and B.

• Reads resultA, resultB. Writes combined, total

WpSplitAWorker (task: wp_split_a)

Split branch A: parallel fork processing.

• Sets result = "branch_a_done"
• Reads chainOutput. Writes result, value

WpSplitBWorker (task: wp_split_b)

Split branch B: parallel fork processing.

• Sets result = "branch_b_done"
• Reads chainOutput. Writes result, value

---

38 tests | Workflow: workflow_patterns_demo | Timeout: 60s

See RUNNING.md for setup and usage.