Metadata-Version: 2.4
Name: bpmn-evalkit
Version: 0.2.0
Summary: An open evaluation harness for text-to-BPMN generation
Project-URL: Homepage, https://bpmn-evalkit.org
Project-URL: Source, https://github.com/tillmannschatz/bpmn-evalkit
Author: Tillmann Schatz
License: 
                                         Apache License
                                   Version 2.0, January 2004
                                http://www.apache.org/licenses/
        
           TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
        
           1. Definitions.
        
              "License" shall mean the terms and conditions for use, reproduction,
              and distribution as defined by Sections 1 through 9 of this document.
        
              "Licensor" shall mean the copyright owner or entity authorized by
              the copyright owner that is granting the License.
        
              "Legal Entity" shall mean the union of the acting entity and all
              other entities that control, are controlled by, or are under common
              control with that entity. For the purposes of this definition,
              "control" means (i) the power, direct or indirect, to cause the
              direction or management of such entity, whether by contract or
              otherwise, or (ii) ownership of fifty percent (50%) or more of the
              outstanding shares, or (iii) beneficial ownership of such entity.
        
              "You" (or "Your") shall mean an individual or Legal Entity
              exercising permissions granted by this License.
        
              "Source" form shall mean the preferred form for making modifications,
              including but not limited to software source code, documentation
              source, and configuration files.
        
              "Object" form shall mean any form resulting from mechanical
              transformation or translation of a Source form, including but
              not limited to compiled object code, generated documentation,
              and conversions to other media types.
        
              "Work" shall mean the work of authorship, whether in Source or
              Object form, made available under the License, as indicated by a
              copyright notice that is included in or attached to the work
              (an example is provided in the Appendix below).
        
              "Derivative Works" shall mean any work, whether in Source or Object
              form, that is based on (or derived from) the Work and for which the
              editorial revisions, annotations, elaborations, or other modifications
              represent, as a whole, an original work of authorship. For the purposes
              of this License, Derivative Works shall not include works that remain
              separable from, or merely link (or bind by name) to the interfaces of,
              the Work and Derivative Works thereof.
        
              "Contribution" shall mean any work of authorship, including
              the original version of the Work and any modifications or additions
              to that Work or Derivative Works thereof, that is intentionally
              submitted to Licensor for inclusion in the Work by the copyright owner
              or by an individual or Legal Entity authorized to submit on behalf of
              the copyright owner. For the purposes of this definition, "submitted"
              means any form of electronic, verbal, or written communication sent
              to the Licensor or its representatives, including but not limited to
              communication on electronic mailing lists, source code control systems,
              and issue tracking systems that are managed by, or on behalf of, the
              Licensor for the purpose of discussing and improving the Work, but
              excluding communication that is conspicuously marked or otherwise
              designated in writing by the copyright owner as "Not a Contribution."
        
              "Contributor" shall mean Licensor and any individual or Legal Entity
              on behalf of whom a Contribution has been received by Licensor and
              subsequently incorporated within the Work.
        
           2. Grant of Copyright License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              copyright license to reproduce, prepare Derivative Works of,
              publicly display, publicly perform, sublicense, and distribute the
              Work and such Derivative Works in Source or Object form.
        
           3. Grant of Patent License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              (except as stated in this section) patent license to make, have made,
              use, offer to sell, sell, import, and otherwise transfer the Work,
              where such license applies only to those patent claims licensable
              by such Contributor that are necessarily infringed by their
              Contribution(s) alone or by combination of their Contribution(s)
              with the Work to which such Contribution(s) was submitted. If You
              institute patent litigation against any entity (including a
              cross-claim or counterclaim in a lawsuit) alleging that the Work
              or a Contribution incorporated within the Work constitutes direct
              or contributory patent infringement, then any patent licenses
              granted to You under this License for that Work shall terminate
              as of the date such litigation is filed.
        
           4. Redistribution. You may reproduce and distribute copies of the
              Work or Derivative Works thereof in any medium, with or without
              modifications, and in Source or Object form, provided that You
              meet the following conditions:
        
              (a) You must give any other recipients of the Work or
                  Derivative Works a copy of this License; and
        
              (b) You must cause any modified files to carry prominent notices
                  stating that You changed the files; and
        
              (c) You must retain, in the Source form of any Derivative Works
                  that You distribute, all copyright, patent, trademark, and
                  attribution notices from the Source form of the Work,
                  excluding those notices that do not pertain to any part of
                  the Derivative Works; and
        
              (d) If the Work includes a "NOTICE" text file as part of its
                  distribution, then any Derivative Works that You distribute must
                  include a readable copy of the attribution notices contained
                  within such NOTICE file, excluding those notices that do not
                  pertain to any part of the Derivative Works, in at least one
                  of the following places: within a NOTICE text file distributed
                  as part of the Derivative Works; within the Source form or
                  documentation, if provided along with the Derivative Works; or,
                  within a display generated by the Derivative Works, if and
                  wherever such third-party notices normally appear. The contents
                  of the NOTICE file are for informational purposes only and
                  do not modify the License. You may add Your own attribution
                  notices within Derivative Works that You distribute, alongside
                  or as an addendum to the NOTICE text from the Work, provided
                  that such additional attribution notices cannot be construed
                  as modifying the License.
        
              You may add Your own copyright statement to Your modifications and
              may provide additional or different license terms and conditions
              for use, reproduction, or distribution of Your modifications, or
              for any such Derivative Works as a whole, provided Your use,
              reproduction, and distribution of the Work otherwise complies with
              the conditions stated in this License.
        
           5. Submission of Contributions. Unless You explicitly state otherwise,
              any Contribution intentionally submitted for inclusion in the Work
              by You to the Licensor shall be under the terms and conditions of
              this License, without any additional terms or conditions.
              Notwithstanding the above, nothing herein shall supersede or modify
              the terms of any separate license agreement you may have executed
              with Licensor regarding such Contributions.
        
           6. Trademarks. This License does not grant permission to use the trade
              names, trademarks, service marks, or product names of the Licensor,
              except as required for reasonable and customary use in describing the
              origin of the Work and reproducing the content of the NOTICE file.
        
           7. Disclaimer of Warranty. Unless required by applicable law or
              agreed to in writing, Licensor provides the Work (and each
              Contributor provides its Contributions) on an "AS IS" BASIS,
              WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
              implied, including, without limitation, any warranties or conditions
              of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
              PARTICULAR PURPOSE. You are solely responsible for determining the
              appropriateness of using or redistributing the Work and assume any
              risks associated with Your exercise of permissions under this License.
        
           8. Limitation of Liability. In no event and under no legal theory,
              whether in tort (including negligence), contract, or otherwise,
              unless required by applicable law (such as deliberate and grossly
              negligent acts) or agreed to in writing, shall any Contributor be
              liable to You for damages, including any direct, indirect, special,
              incidental, or consequential damages of any character arising as a
              result of this License or out of the use or inability to use the
              Work (including but not limited to damages for loss of goodwill,
              work stoppage, computer failure or malfunction, or any and all
              other commercial damages or losses), even if such Contributor
              has been advised of the possibility of such damages.
        
           9. Accepting Warranty or Additional Liability. While redistributing
              the Work or Derivative Works thereof, You may choose to offer,
              and charge a fee for, acceptance of support, warranty, indemnity,
              or other liability obligations and/or rights consistent with this
              License. However, in accepting such obligations, You may act only
              on Your own behalf and on Your sole responsibility, not on behalf
              of any other Contributor, and only if You agree to indemnify,
              defend, and hold each Contributor harmless for any liability
              incurred by, or claims asserted against, such Contributor by reason
              of your accepting any such warranty or additional liability.
        
           END OF TERMS AND CONDITIONS
        
           APPENDIX: How to apply the Apache License to your work.
        
              To apply the Apache License to your work, attach the following
              boilerplate notice, with the fields enclosed by brackets "[]"
              replaced with your own identifying information. (Don't include
              the brackets!)  The text should be enclosed in the appropriate
              comment syntax for the file format. We also recommend that a
              file or class name and description of purpose be included on the
              same "printed page" as the copyright notice for easier
              identification within third-party archives.
        
           Copyright [yyyy] [name of copyright owner]
        
           Licensed under the Apache License, Version 2.0 (the "License");
           you may not use this file except in compliance with the License.
           You may obtain a copy of the License at
        
               http://www.apache.org/licenses/LICENSE-2.0
        
           Unless required by applicable law or agreed to in writing, software
           distributed under the License is distributed on an "AS IS" BASIS,
           WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
           See the License for the specific language governing permissions and
           limitations under the License.
License-File: LICENSE
Keywords: benchmark,bpmn,evaluation,llm,process-mining
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering
Requires-Python: >=3.11
Requires-Dist: httpx>=0.27
Requires-Dist: lxml>=5.0
Requires-Dist: pm4py>=2.7.15
Requires-Dist: pyyaml>=6.0
Requires-Dist: scipy>=1.11
Requires-Dist: typer>=0.12
Provides-Extra: dev
Requires-Dist: fastapi>=0.115; extra == 'dev'
Requires-Dist: httpx>=0.27; extra == 'dev'
Requires-Dist: jinja2>=3.1; extra == 'dev'
Requires-Dist: mypy>=1.11; extra == 'dev'
Requires-Dist: pytest>=8.0; extra == 'dev'
Requires-Dist: python-multipart>=0.0.9; extra == 'dev'
Requires-Dist: ruff>=0.6; extra == 'dev'
Requires-Dist: types-pyyaml; extra == 'dev'
Requires-Dist: uvicorn>=0.30; extra == 'dev'
Provides-Extra: embedding
Requires-Dist: sentence-transformers>=3.0; extra == 'embedding'
Provides-Extra: ui
Requires-Dist: fastapi>=0.115; extra == 'ui'
Requires-Dist: jinja2>=3.1; extra == 'ui'
Requires-Dist: python-multipart>=0.0.9; extra == 'ui'
Requires-Dist: uvicorn>=0.30; extra == 'ui'
Description-Content-Type: text/markdown

# bpmn-evalkit

An open evaluation harness for text-to-BPMN generation. Bring your own system —
or benchmark a raw LLM — against curated public gold datasets, with
reproducible, clearly separated metrics.

## Why

LLM-based process model generation is advancing fast, but there is no shared,
reusable tooling to evaluate it: metric frameworks exist as papers, gold data
exists in scattered formats, and every group re-implements its own ad-hoc
pipeline. `bpmn-evalkit` is the missing glue — dataset loaders, validity gates,
and a CLI that turns any text-to-BPMN system into a benchmarked one.

## What it is / is not

**It is** a measurement tool. System-agnostic (HTTP, CLI, or raw-LLM adapters),
reproducible (pinned datasets, versioned metrics, k-repetition runs), honest
(rates over all outputs, no composite score, gates that report *skipped* rather
than silently passing).

**It is not** a leaderboard, a vendor comparison, or a judge. This repo ships
tooling; conclusions are yours. Public datasets may be present in LLM training
data — treat absolute scores accordingly and prefer deltas between systems
evaluated under identical conditions.

Four permanent non-goals, stated here because they are the project's
credibility: **no leaderboard or vendor comparisons published by this project ·
no redistribution of third-party datasets · no composite score · no private
held-out suites folded in.**

## Quickstart

```bash
pip install bpmn-evalkit
bek run --adapter fixture --dataset fixture --k 3
bek score runs/<run>
bek report runs/<run>
bek compare runs/<baseline> runs/<candidate>
```

`compare` refuses by default when the two runs disagree on dataset, `k`,
`metrics_version`, generation budget, matcher or threshold — a delta across
those measures the harness as much as the systems, and it looks exactly like a
finding. Settings a run did not record come back as *unverified*, never as
agreement.

That works offline against three built-in fixture cases — no dataset download,
no API key, nothing to configure: the models ship inside the package and
`--adapter fixture` needs no file on disk. Its purpose is to prove the pipeline
runs end to end, not to say anything about quality.

To benchmark a real system, clone the repo (or copy the examples out of it):
`configs/http.example.yaml` points at your endpoint, and
`configs/llm.anthropic.example.yaml` benchmarks a raw model. Both need a real
dataset first:

```bash
bek fetch pmo --accept-licenses
bek run --adapter configs/llm.anthropic.example.yaml --dataset pmo --k 3
```

## The local instance (UI)

```bash
pip install "bpmn-evalkit[ui]"
bek serve            # http://127.0.0.1:8092
```

Start a benchmark, watch it run, browse the history, compare two runs, and see
one system's rates over time. `deploy/bpmn-evalkit.service` runs it as a
systemd user service.

Three properties worth knowing:

- **The UI is not a second code path.** It calls the same runner and the same
  scorer as the CLI, so a run started in the browser is byte-for-byte a run
  started in a shell. Numbers cannot drift between the two.
- **The filesystem is the database.** A run is a directory; there is no index
  to keep in sync and nothing to migrate. Runs stay readable if this tool
  disappears.
- **Your results never leave your machine.** `runs/` and `data/` are
  gitignored, and your real adapter config (endpoint, token) is yours. The
  public repo is tooling; what you measure with it is not.

It binds to localhost and has no authentication, because it is not reachable
from anywhere else. If you bind it publicly, put something that authenticates
in front of it — the UI can start jobs that spend money on API calls.

## Metrics

Stage 1 validity gates: well-formedness, official BPMN 2.0 XSD, bpmnlint,
workflow-net structure, soundness via Petri-net analysis, global reachability,
and diagram-interchange completeness. Normative definitions live in
[`docs/metrics.md`](docs/metrics.md); **any change there is a breaking change**
and `metrics_version` is stamped into every report.

Four design decisions in the registry are worth knowing before you read a
number:

| | |
|---|---|
| **Soundness is three-valued** | Models that are not workflow nets — more than one start or end event, which is legitimate BPMN — are reported separately and **never counted as unsound**. A naive single rate is 20–30 points too pessimistic on real corpora. |
| **Outputs are normalised first** | `<incoming>`/`<outgoing>` are rebuilt from `sourceRef`/`targetRef` before any gate runs. Without it bpmnlint calls every node disconnected and pm4py imports zero sequence flows — the gates would measure the parser, not the model. |
| **Lint has two rates** | Error-free and clean-including-warnings are different questions and differ a lot. Both are reported, with a mandatory per-rule breakdown. |
| **Skipped ≠ passed** | A gate whose checker is unavailable (no Node, no network for the XSDs) is excluded from its denominator and counted, never treated as a pass. |

Stage 2 compares an output against the gold reference element by element —
precision, recall and F1 for activities, events, gateways, sequence flows and
lanes (`struct.f1.*`). Two properties decide whether such a number means
anything:

| | |
|---|---|
| **Matching is two-pass, not label-based** | Elements labelled on both sides are matched by label similarity as a bipartite assignment; everything else — unlabelled gateways, unlabelled events, every flow — is matched by its position relative to those anchors. Pure label matching measures labelling style: the PMo gold models trip `label-required` 102 times, so a system that labels well would be *penalised* against them. |
| **The matcher and threshold are part of the score** | Both are recorded and printed in every report. A score computed at another threshold is a different score, and comparing the two is a category error. The default matcher is deterministic and offline — no model download, nothing to drift. |

And a stage-2 number needs a scale. Against a multi-reference dataset
(`bek fetch mangler`) the harness also scores the **gold models against each
other** and reports that agreement ceiling next to the system's own figure:
measured on that corpus, two expert models of the same description agree at F1
0.38 on activities and 0.58 on gateways — not 1.0. A system read against an
implied ceiling of 1.0 is being read against a standard no human meets. The
ceiling is a second number to put beside the first, never a divisor.

Stage 3 asks whether the process *does* the same thing — two models can be
built differently and admit exactly the same runs, which stage 2 penalises
without noticing. It compares the output's Petri net with the reference's
directly, so it needs **no event log and no play-out sampling**: a gold BPMN
model is already a behavioural specification. Fitness and precision are always
reported together, because fitness alone rewards a model that permits
everything and precision alone one that permits nothing.

Its ceiling is the number to read first: two human gold models of the same
description agree on **7%** of behavioural relations. A relation needs both
endpoints matched, so the label-matching bottleneck compounds — on a corpus
with divergent vocabularies this metric says more about activity matching than
about behaviour, and the report says so next to every figure.

## Requirements

Python ≥ 3.11. `bpmnlint` runs through a small Node bridge (Node ≥ 20,
`cd node && npm install`); without it the lint gate reports **skipped**, never
passed. The BPMN 2.0 XSDs are downloaded once and cached on first use.

From a checkout: `pip install -e ".[dev]"`, then `pytest && ruff check . &&
mypy`.

## Roadmap

Released versions and what each one changed are in
[`CHANGELOG.md`](CHANGELOG.md); `metrics_version` moves on its own schedule and
is governed by the registry.

- **v0.1** — stage-1 gates, PMo loader, http/command/llm adapters, k-repetition
  runner, `metrics.json`, `stability.k`
- **v0.2** — element similarity and the agreement ceiling, multi-reference
  loader, threshold sensitivity, subtype confusion, model-to-model behaviour,
  `bek compare` with the R4 check enforced. *(Graph distance — `struct.graph.*`
  — was on this list and left it without being built: it would rest on the same
  matching as `struct.f1.*`, and the matching is what limits both. The ids stay
  reserved, now without a date.)*
- **v0.3** — log-based behavioural replay against event-log gold. *(A better
  matcher was the other half of the plan: sentence embeddings help modestly,
  stemming and 1:n grouping do not. All three are measured in the registry.)*
- **v0.4** — guideline checks, MIWG round-trip module, per-construct analysis

## Standing on shoulders

This harness operationalizes prior work — please cite the originals rather than
this tool: the BEF4LLM quality-dimension framework; Kourani et al. (SoSyM 2025)
for the behavioural evaluation approach; the PMo Dataset (Polytechnique
Montréal) and the Mangler/Klievtsova multi-reference dataset for gold data; and
`pm4py` and `bpmnlint` for the machinery underneath. See
[`CITATIONS.bib`](CITATIONS.bib).

## Contributing

Issues and dataset-loader PRs welcome. A new metric requires an entry in
`docs/metrics.md` and a version bump — the registry is the contract, not the
code.

---

Created and maintained by Tillmann Schatz. Apache-2.0.
