v0.5.0

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog,
and this project adheres to Semantic Versioning.

[0.5.0] - 2026-10-11

Added

• Pluggable Graph Subsystem & Multi-Format Renderers:
  • WorkflowGraph pure domain model with cycle detection, topological sorting, reachability analysis (ancestors, descendants, blast_radius), and critical path latency bottleneck calculation.
  • Multi-format diagram adapters conforming to GraphRendererPort:
    • MermaidGraphRenderer: Stage-clustered flowcharts with execution pool styling, rule badges, and mapped indicators.
    • DotGraphRenderer: Graphviz DOT format with customizable styles and node highlighting.
    • AsciiGraphRenderer: Unicode box-drawing visual DAG representation for terminal output.
    • JsonGraphRenderer: JSON-serialized structural representation for programmatic integration.
    • RendererRegistry: Extensible renderer catalog supporting third-party format plugins.
  • Optional visual artifact export (render_image(..., image_format="svg" | "png")) with CLI fallbacks (mmdc, dot).
• Dry-Run Simulation Mode & Chaos Fault Injection:
  • dry_run=True simulation execution across scalar and mapped fan-out steps without mutating databases or external APIs.
  • Fail-closed safety guards: steps declaring side_effects=True require explicit mock fallbacks (dry_run=...) or allow_unsafe=True, preventing unintended production mutations.
  • Chaos fault injection: fault_injection={step_name: Exception} parameter allows simulating component failures and testing resilient compensating workflows.
• Workflow Memento Pattern & Time-Travel Rewind:
  • Immutable state snapshots (WorkflowMemento) capturing complete execution state, checkpoints, and inputs.
  • Time-travel execution rewind (rewind_to(step_name)): resets execution to before any historical step, clearing downstream checkpoints for selective re-execution.
• Onion-Layered Step Middleware Pipeline:
  • Abstract StepMiddlewarePort protocol and InOutStepMiddleware base class for pre/post step interception across sync and async callables.
  • Built-in middleware implementations:
    • LoggingStepMiddleware: Structured logging of step start, completion, and failure events.
    • TimingStepMiddleware: High-precision latency measurement in milliseconds stored on step context metadata.
    • MockFallbackMiddleware: Automatic synthetic fallback generation for dry-run simulation mode.
  • Fluent registration via engine.use(middleware) and Workflow(..., middleware=[...]).
• Enhanced CLI Subcommands (hf graph render & hf graph info):
  • hf graph render <target> --format <mermaid|dot|ascii|json>: Visualizes workflow DAGs directly from Python files, functions, or GitHub Actions YAML files.
  • Direct GitHub Actions YAML pipeline parsing (GitHubActionsYamlLoader): parses job dependencies (needs) into a WorkflowGraph without custom scripts.
  • --dry-run simulation flag on hf graph render to expand and visualize dynamic mapped steps before execution.
  • --summary: writes rendered Mermaid diagrams directly to GitHub Actions $GITHUB_STEP_SUMMARY.
  • hf graph info <target>: Displays topological metadata, estimated latency, bottleneck step, and critical path nodes.

Changed

• Refactored AsyncioWorkflowEngine execution pipeline into an onion-layered middleware dispatch chain.
• Constructor worker limits (max_process_workers, max_thread_workers) on Workflow retain original positional parameter ordering.

[0.4.0] - 2026-09-25

Added

• Multi-Core Worker Pools & Execution Pools (ExecutionPool):
  • ExecutionPool enum supporting ASYNC (asyncio event loop coroutine), THREAD (ThreadPoolExecutor), and PROCESS (ProcessPoolExecutor).
  • pool parameter on StepDefinition, @wf.step(..., pool=...), and @wf.map_step(..., pool=...) for true multi-core CPU scaling.
  • Safe cross-process parameter binding and trampoline execution (mp_context="spawn") for CPU-bound step and mapped sub-step execution.
  • Granular process and thread worker limit tuning via max_process_workers and max_thread_workers on AsyncioWorkflowEngine and Workflow.
  • Resource lifecycle management with close(), aclose(), sync and async context managers (with Workflow(...), async with Workflow(...)).
  • LocalAsyncWorkflowEngine alias for AsyncioWorkflowEngine.

[0.3.0] - 2026-09-16

Added

• Dynamic Step Mapping (@wf.map_step):
  • Runtime fan-out decorator allowing a step to dynamically map over a collection produced by an upstream step or context input.
  • Automatic dependency inference linking mapped steps to their upstream targets.
  • Granular concurrency throttling via concurrency_limit parameter.
  • Flexible callable signature binding (supporting (item), (item, ctx), (ctx, item), and context keyword arguments).
  • Sub-step checkpoint isolation (<step>[<i>]) supporting individual retries, timeouts, and seamless resumption skipping previously completed items.
• DAG Exporters:
  • WorkflowDefinition.to_mermaid(direction="TD") and Workflow.to_mermaid() for rendering valid Mermaid flowcharts clustered by stage with rule and mapped node annotations.
  • WorkflowDefinition.to_ascii() and Workflow.to_ascii() for rendering clean Unicode hierarchical trees for terminal and dashboard output.
• Governance:
  • Upgraded dev dependencies to hexaqual[all]>=0.3.0 and unified pre-commit hook configuration.

[0.2.0] - 2026-09-14

Added

• TriggerRule enum supporting ALL_SUCCESS, ALL_FAILED, ALL_DONE, ONE_SUCCESS, ONE_FAILED, NONE_FAILED, and ALL_SUCCESS_OR_SKIPPED.
• Pure function evaluate_trigger_rule(rule, parent_statuses) in domain models.
• Support for trigger_rule on StepDefinition and fluent builder Workflow.step(..., trigger_rule=...).
• Execution-time step skipping via skip_steps: set[str] | list[str] | None on WorkflowEnginePort and AsyncioWorkflowEngine (run, run_async, resume, resume_async).
• Cascading skip propagation: skipped upstream dependencies cascade downstream according to each child step's trigger_rule.
• WorkflowCliBinder in hexaflow.cli.binder for introspecting workflow DAGs and dynamically generating CLI option signatures (--skip-<step>) for Typer commands.
• Workflow.create_cli_binder() DSL convenience helper.
• Comprehensive unit test coverage across trigger rules, skip cascades, and CLI binder parameter translation.

[0.1.1] - 2026-09-13

Added

• Quality gate integration with Hexaqual (hexaqual[all]>=0.2.0).
• Test symmetry verification and __all__ integrity enforcement.
• Automated release workflow with provenance attestations and PyPI publication.

[0.1.0] - 2026-09-13

Added

• Lightweight, embeddable Python workflow engine with stages, steps, splits (fan-out), and joins (barriers).
• Pure domain models for workflows, stages, steps, retry policies, and execution state.
• In-memory and SQLite-backed state stores (InMemoryStateStore, SqliteStateStore) with checkpointed resumption.
• Asynchronous local execution engine (AsyncioWorkflowEngine).
• Fluent workflow definition DSL (Workflow).
• Command-line interface (hf) with commands for inspection, execution, resumption, and status dashboards.
• End-to-end resumable pipeline integration tests and Hypothesis property-based tests.

What's Changed

• fix(security): resolve CodeQL alerts, Scorecard findings, and modernize governance by @TheTrueSCU in #17
• feat(graph): pluggable graph subsystem, multi-format renderers, simulation mode, and workflow memento by @TheTrueSCU in #21
• feat(middleware): implement onion-layered StepMiddlewarePort and adapters by @TheTrueSCU in #23
• test(ports): use new invocation to test abstract instantiation across ty versions by @TheTrueSCU in #25
• chore(deps): Bump the workspace-deps group across 1 directory with 6 updates by @dependabot[bot] in #24
• chore(release): release v0.5.0 by @TheTrueSCU in #26

Full Changelog: v0.4.1...v0.5.0

v0.4.1

What's Changed

• chore(deps): Bump astral-sh/setup-uv from 10.1.0 to 10.2.0 by @dependabot[bot] in #5
• chore(deps): Bump github/codeql-action/analyze from 4.38.0 to 4.38.1 by @dependabot[bot] in #6
• chore(deps): Bump github/codeql-action/upload-sarif from 4.38.0 to 4.38.1 by @dependabot[bot] in #7
• chore(deps): Bump github/codeql-action/init from 4.38.0 to 4.38.1 by @dependabot[bot] in #8
• chore(deps): Bump virtualenv from 21.7.9 to 21.7.13 in the uv group across 1 directory by @dependabot[bot] in #9
• chore(deps): Bump github/codeql-action/init from 4.38.1 to 4.38.2 by @dependabot[bot] in #11
• chore(deps): Bump astral-sh/setup-uv from 10.1.0 to 10.2.0 by @dependabot[bot] in #14
• chore(deps): Bump github/codeql-action/upload-sarif from 4.38.1 to 4.38.2 by @dependabot[bot] in #13
• fix(audit): remediate full codebase audit findings by @TheTrueSCU in #16

New Contributors

• @TheTrueSCU made their first contribution in #16

Full Changelog: v0.4.0...v0.4.1

v0.4.0

What's Changed

• chore(deps): Bump github/codeql-action/init from 3.37.9 to 4.38.0 by @dependabot[bot] in #2
• chore(deps): Bump codecov/codecov-action from 7.0.0 to 7.1.1 by @dependabot[bot] in #3
• chore(deps): Bump github/codeql-action/analyze from 3.37.9 to 4.38.0 by @dependabot[bot] in #4

Full Changelog: v0.3.0...v0.4.0

v0.3.0

Full Changelog: v0.2.0...v0.3.0

v0.2.0

Full Changelog: v0.1.1...v0.2.0

v0.1.1

Full Changelog: v0.1.0...v0.1.1

v0.1.0 - First Standalone Release

🌊 hexaflow v0.1.0

First official release of hexaflow: lightweight, embeddable Python workflow engine with stages, steps, splits, joins, and checkpointed resumption.

Highlights

• Hierarchical DAG Workflow Engine: Workflows structured into Stages and Steps with -based topological resolution.
• Splits & Joins: Concurrently fan-out steps within stages and synchronize at multi-step join barriers.
• Fault Partitioning: Transient retries with exponential backoff and full jitter vs. permanent failure suspension (SUSPENDED).
• Resumption & Restart Semantics: Checkpointed step inputs/outputs in embedded SQLite (with automatic $>64\text{ KB}$ artifact disk spillover). Resume suspended runs bypassing previously completed steps.
• Dual DSL: Decorator-based syntax (@wf.stage, @wf.step) alongside declarative classes.
• Developer CLI: Full Typer/Rich CLI (hexaflow and hf) for run, status, inspect, resume, restart, and abort.
• Ecosystem Ready: Zero dependencies on heavy orchestrators, ready to serve as the local execution core for hexastack and the cluster job definition for hexaqueue.