Lila ModelerExamples and regression fixtures
Lila is a BPMN modelling and simulation tool. Examples demonstrate how to build a model, set assumptions, compare scenarios and interpret results. This policy applies to existing and future cases, including documentation, screenshots, issue/PR text and attachments.
Publication criteria
- Use synthetic workflows and invented input data, or reference material whose publication is permitted and whose source and attribution are documented. Public availability alone is not a licence to copy a source document.
- Keep private case statements, coursework, assessment questions, submitted answers and personal or confidential information outside the public repository. A change of title or names alone does not remove these concerns: review the narrative, input data, labels, IDs, outputs and images.
- For regression tests, reduce the behaviour to a neutral fixture with the smallest useful scope.
Place it under the relevant package's
test/fixtures/directory. Numerical expectations and analytical oracles are appropriate when they validate explicit engine rules or invariants. - Keep user-facing examples under
examples/. Their README must state the source or synthetic origin, purpose, assumptions, reproduction steps and known limitations. Preserve third-party attribution and licensing; do not relabel an attributed reference as an original synthetic case. - Validate BPMN references, scenario resolution and the claimed behaviour. Keep comparisons reproducible, and document any change to element IDs that affects seeded random streams.
Current case inventory
| Location | Purpose and provenance |
|---|---|
examples/pedido/ |
Project restaurant-order demo and regression benchmark; see its README for assumptions and commands. |
examples/mm1/ |
Synthetic queueing models with analytical Erlang C expectations and reproduction scripts. |
examples/outline/ |
Synthetic outline (#97) for lila process create and the MCP tool create_process, and synthetic edit operations (#98) for lila process edit and edit_process. |
examples/bizagi-levels/ |
Attributed reconstructions of public reference examples, with source URLs, expected values and documented differences. |
examples/bizagi-exports/ |
Unmodified BPMN MIWG interoperability fixtures, with upstream commit and CC BY 3.0 attribution. |
packages/engine/test/fixtures/outlines/ |
Synthetic outlines (#97) from the QA of #553: branch order, same-row skips, shared gateway trunks. |
packages/engine/test/fixtures/service-request/ |
Neutral regression inputs for resource, outcome, UI and replay checks. |
packages/engine/test/fixtures/service-shared-denial.bpmn |
Neutral shared-path fixture for conditional-routing invariants. |
packages/engine/test/fixtures/zero-time-driver.ts |
Synthetic zero-time loop and finite-control regression models for #368; executed in timeout-protected child processes. |
This inventory describes the maintained purpose of the cases; provenance and permissions must be reviewed again when source material is added or replaced. Update the inventory when adding a new case family. Other focused test fixtures follow the same publication criteria.
Private work and review
Keep private working files outside the repository, or in the ignored root local/, private/
or exports/ directories. .gitignore does not remove tracked content or protect GitHub text,
attachments, commit history or existing clones. Review the staged diff and PR content together.
Before merging, confirm provenance, publication scope and the relevant acceptance checks in the PR checklist. If provenance is unclear, prepare a new synthetic fixture before publishing.