Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions data/synthetic/.gitignore
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
__pycache__/
*.py[cod]
.venv/
dist/
build/
*.egg-info/
.pytest_cache/
.ruff_cache/
225 changes: 225 additions & 0 deletions data/synthetic/DATACARD.md

Large diffs are not rendered by default.

190 changes: 190 additions & 0 deletions data/synthetic/LICENSE
Original file line number Diff line number Diff line change
@@ -0,0 +1,190 @@
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 the 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 the 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 any 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

Copyright 2026 Unpatterned Labs

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.
105 changes: 105 additions & 0 deletions data/synthetic/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,105 @@
# arche-synthetic

Known entity worlds, imperfect observations, exact ground truth — and every disagreement labelled with **why it happened**.

```bash
pip install arche-synthetic # pyarrow and pyyaml. Nothing else.
python -m arche_synthetic --scale 400 --out ./worlds/ng
```

```python
from arche_synthetic import Benchmark, Predictions, build

world, manifest = build("ng_supplier_v0", seed=42, out=Path("worlds/ng"))

bench = Benchmark.load(Path("worlds/ng"))
report = bench.score(Predictions(arm="mine", pairs=my_predicted_pairs))
report["recall_by_stratum"]["change/ORG_RELOCATED"] # the column nobody else has
```

From a checkout of this repository, run the scripts **from the repo root**. `data/synthetic` has its own `pyproject.toml`, so `uv run` with the working directory inside it resolves to the standalone project and will not find arche:

```bash
python data/synthetic/build_ng_supplier_v0.py # ~7s, writes worlds/ng_supplier_v0/
python data/synthetic/run_benchmark.py --world data/synthetic/worlds/ng_supplier_v0_s150
```

## A second world: artists, for the variant-list question

`artists_v0` is 9,944 records of 2,941 real African musicians under their real Wikidata aliases, built to answer one question put to us by Splink's author: what is a name-variant list worth to a matcher? The world, the independent MusicBrainz-derived list, a ninety-line Splink example and the full six-arm result are in [`datasets/artists_dataops/`](../../datasets/artists_dataops/README.md).

```bash
python -m arche_synthetic --world-pack artists_v0 --scale 10000 --out data/synthetic/worlds/artists_v0
```

## Why this exists

arche keeps being blocked by data licensing. ai4privacy's 300k is academic-only and forbids derivatives; i2b2 needs a data-use agreement; OpenSanctions Pairs is a purchase. Every one of those is somebody else's permission. This is the path that needs nobody's.

And one thing arche can do had never been measured: `ledger.observe()` — add evidence about a record, re-decide its open pairs, supersede rather than overwrite — had no benchmark, because measuring it needs data where an entity *legitimately changed* and the truth says so. No public entity-resolution benchmark has that column. This generates it, and [RESULTS.md](RESULTS.md) is what it found.

## The idea in one table

| kind | example | are both records true? |
|---|---|---|
| `change` | the supplier moved; the director resigned; the account was replaced | **yes**, at their respective times |
| `representation` | `Établissements Koné & Fils SARL` → `ETS Kone & Fils` | yes — same fact, different encoding |
| `error` | typo, OCR misread, truncation, mutated digit | no — one is a distortion |

Existing synthetic ER benchmarks produce only the third row: take a record, corrupt it, call the result a duplicate. A matcher tuned on that learns that all disagreement means "different entity", and then splits entities in production the first time a supplier moves.

The labels are computed **by construction**. Every observed value carries a note of which state it came from and which transforms were applied, so a difference is explained by comparing provenance rather than by guessing from the strings.

## The modules

```
arche_synthetic/
├── ids.py opaque record ids — the invariant that keeps truth out of the input
├── world.py entities, states, relationships; the frequency-weighted name draw
├── lifecycle.py events: relocation, rename, director change, account change
├── observe.py sources, representation rules, corruption rules, difference labelling
├── export.py Parquet + JSON Schema + manifest, and the write-time leak check
├── evaluate.py the stratified scorer — recall stratifies, precision does not
└── data/ the vendored name lexicon (CC-BY-4.0)
```

**Nothing imports `arche` at all.** The name lexicon is vendored (from arche's `african_names_v1`, itself built from Wikidata/ParaNames), and the RC/TIN format checks are local — with `test_synthetic_world.py::test_the_local_validators_agree_with_arches` asserting they still agree with arche's, which is stronger than importing either, because a one-sided import would hide a drift.

Verified rather than asserted: installed into an empty virtualenv with only `pyarrow` and `pyyaml` beside it, the package generates a world whose content fingerprints are **identical** to one built inside this repo.

## Is it realistic enough to measure on?

`examples/notebooks/24_is_the_synthetic_world_realistic.ipynb` holds this world up against **51,023 real Nigerian facility names** (GRID3) and tests three hypotheses stated in advance. Summary: the collision *rate* is realistic (13.5% against 12.0%), the vocabulary is 2.8x thinner at matched sample size, and **0.0% of generated name tokens carry regional signal against 90.4% of real ones** — the correlation defect, measured rather than asserted. Which published numbers that does and does not affect is the notebook's last cell.

## Output

See [DATACARD.md](DATACARD.md) for what the shipped world contains, every measured number, and — as importantly — what it is not.

**The contract is the files, not a Python class.** Parquet with documented columns, a JSON Schema beside each one, a manifest with provenance. A Splink, Zingg, R or SQL user consumes it without installing anything of ours. The moment the truth is only readable through our own types, matcher-neutrality is gone.

## Running the benchmark

```bash
python data/synthetic/run_benchmark.py --world data/synthetic/worlds/ng_supplier_v0_s150
python data/synthetic/run_benchmark.py --arms exact,jaro_winkler # the fast ones
```

Seven arms: `exact` (normalised name + city join), `jaro_winkler`, `splink` (Fellegi-Sunter, its own decision boundary), `arche/pack`, `arche/custom`, and the two ledger arms — `arche/ledger` (records ingested in arrival order) and `arche/ledger+observe` (plus the operator's registry lookup). **Measured results and their caveats: [RESULTS.md](RESULTS.md).** The short version — every engine that reads the address fails on relocated suppliers, and the ledger path is the only thing that moves the number.

The arms are extras, so the base install stays two dependencies: `pip install "arche-synthetic[bench]"` for the baselines, `[splink]` for the Splink arm, `[arche]` for the arche and ledger arms, `[all]` for everything.

### Cutting the repository

The trigger the plan set was *a stratified result **and** arche not being the only arm*. Both hold: Splink runs on this data and lands within 0.007 recall of `arche/pack`.

This directory is now a complete standalone project — `pyproject.toml`, `LICENSE`, `.gitignore`, a vendored lexicon, no arche import anywhere, installable and verified in a clean environment. What remains is one command to extract its history and one to publish it:

```bash
git subtree split -P data/synthetic -b arche-synthetic
```

then push that branch to a newly created `unpatterned-labs/arche-synthetic` and set `main` to it.

**Publishing is the owner's to do.** Nothing in this branch creates a remote, pushes, or uploads anything; the repository's guards block an agent from doing so, and that is the correct arrangement.

Afterwards, in this repository: drop the directory, add `arche-synthetic` as a dev dependency, and keep a thin consumer that runs the arche arms against the published package. The three test files in `packages/arche-core/tests/` move last — they become the new repo's `tests/`, and until the cut they are what keeps the generator honest inside arche's CI.
Loading
Loading