README.md

---
license: gpl-3.0
language:
- en
- fr
pretty_name: "Ailiance — KiCad 9+ Schematic Corpus (Copyleft)"
task_categories:
- text-generation
tags:
- kicad
- eda
- schematic
- s-expression
- electronics
- hardware
- copyleft
- ailiance
size_categories:
- n<1K
---

# Ailiance — KiCad 9+ Schematic Corpus (Copyleft)

> 🇫🇷 **Ailiance** — curated by Ailiance for production deployment ; co-published with the upstream [`electron-rare/kicad9plus-copyleft`](https://huggingface.co/datasets/Ailiance-fr/kicad9plus-copyleft). 🇪🇺 Compatible EU AI Act (Template AI Office, July 2025).

Corpus de 209 schémas KiCad 9+ (`.kicad_sch`, format S-expression, version ≥ 20240722) collectés sous licences **copyleft / réciproques fortes** (GPL-3.0, CERN-OHL-S-2.0, EUPL-1.2). Compatible GPL-3.0-or-later au niveau aggregé. Pensé pour l'entraînement de modèles open-source soumis aux mêmes obligations de réciprocité.

## Statistics

| Métrique         | Valeur            |
|------------------|-------------------|
| Total samples    | **209** |
| Size             | ~2.1 MB JSONL (raw `.kicad_sch` ≈ 38 MB before chat-format wrapping) |
| Format           | JSONL chat format (`messages` array, user/assistant) |
| Languages        | French + English |
| Aggregate license| **GPL-3.0-or-later** |

## Composition

**Copyleft / strong reciprocal licenses only**: GPL-3.0 (169), CERN-OHL-S-2.0 (36), EUPL-1.2 (4).

Copyleft split of the original `electron-rare/kicad9plus-sch-corpus` (now deprecated). L'aggregé est re-licencié sous **GPL-3.0-or-later**, le plus restrictif des inputs (CERN-OHL-S-2.0 et EUPL-1.2 sont explicitement compatibles avec GPL-3.0+ via FSF / EUPL appendix interoperability).

Pour les samples permissifs (Apache-2.0, MIT, CC0-1.0, CERN-OHL-P-2.0) : voir [`Ailiance-fr/kicad9plus-permissive`](https://huggingface.co/datasets/Ailiance-fr/kicad9plus-permissive).

## EU AI Act compliance (Template AI Office, July 2025)

### General information
- **Name**: kicad9plus-sch-corpus (copyleft subset)
- **Modality**: text (KiCad S-expression source)
- **Languages**: English (technical), French (some title-block descriptions)
- **Intended use**: entraînement / fine-tuning de modèles de génération de schémas KiCad 9 / KiCad 10. **Les modèles entraînés sur ce dataset doivent respecter les obligations GPL-3.0-or-later** (disclosure des poids, des données dérivées d'entraînement, et du code d'inférence si redistribué).

### Data sources

**Publicly available datasets**: None.

**Web scraping**: Yes — public GitHub repositories.
- Discovery method: `gh search code "(kicad_sch (version 202X)" extension:kicad_sch`
- Filtering: SPDX license detection of GPL-3.0, CERN-OHL-S-2.0, EUPL-1.2
- Per-sample provenance: see `metadata.source_url`, `metadata.commit_sha`, `metadata.repo`

**Licensed data**: None (no commercial / proprietary licenses).

### Data processing
- Sparse-clone with `gh repo clone --depth 1`
- Per-file `.meta.json` sidecar with `source_url`, `commit_sha`, `license_spdx`, `kicad_version`
- Deduplication via SHA-256 of file contents
- Truncation at 8 KB to fit training context (marked in metadata when applicable)
- Validation via `kicad-cli sch erc --format json --severity-all` (99.6% pass rate on tested subset)
- All processing scripts available at https://github.com/ailiance/ailiance-bench/tree/main/scripts

### Data characteristics
- **Size**: 209 samples, ~2.1 MB JSONL (raw `.kicad_sch` totals ~38 MB before chat-format wrapping)
- **License mix (input)**:
  - GPL-3.0: 169 samples (80.9%)
  - CERN-OHL-S-2.0: 36 (17.2%)
  - EUPL-1.2: 4 (1.9%)
- **KiCad version mix**: 20240819 / 20240910 / 20241004 / 20241209 (KiCad 9 dev), 20250114 (KiCad 9 stable), 20250227 / 20250318 / 20250610 / 20250829 / 20250901 / 20250922 / 20251012 / 20251028 (later 9.x), 20260101 / 20260306 (KiCad 10)
- **Source repos**: 9 distinct upstream projects (full list in `LICENSE_INVENTORY.md`); largest contributors are `jaguilar/kicad` (133) and `flaviens/kicad` (34) — both KiCad demo / fork repositories under GPL-3.0.

## Sample format

Each line is a chat-format JSON object:

```json
{
  "messages": [
    {"role": "user", "content": "Generate a KiCad 9 schematic (titled '...', by ..., N components, libraries: ...). Use the standard S-expression format starting with `(kicad_sch ...)`."},
    {"role": "assistant", "content": "(kicad_sch\n\t(version 20250114)\n\t..."}
  ],
  "metadata": {
    "repo": "owner/name",
    "rel_path": "path/to/file.kicad_sch",
    "source_url": "https://github.com/owner/name/blob/<sha>/path/to/file.kicad_sch",
    "commit_sha": "<git sha>",
    "license_spdx": "GPL-3.0",
    "kicad_version": "20250114",
    "file_sha256": "<content sha>",
    "file_size_bytes": 12345,
    "downloaded_at": "2026-05-11T...",
    "compliance_notes": "...",
    "ia_act_status": "requires_review"
  }
}
```

## Licenses applied
This dataset (the aggregated work) is released under **GPL-3.0-or-later**.

Per-sample original licenses are **preserved in `metadata.license_spdx`** and listed in `LICENSE_INVENTORY.md`. Downstream users MUST preserve attribution per sample and comply with the strongest applicable copyleft term (GPL-3.0-or-later for the aggregate).

Compatibility notes:
- **CERN-OHL-S-2.0 -> GPL-3.0+**: explicitly compatible (CERN-OHL-S §7 allows redistribution under GPL when combining with GPL works).
- **EUPL-1.2 -> GPL-3.0+**: compatible via the EUPL §5 / Appendix list (GPL-3.0 is a listed compatible licence).
- **GPL-3.0 -> GPL-3.0+**: trivially compatible.

## Copyright considerations
- All sources are public GitHub repositories under copyleft licenses.
- The `.kicad_sch` files are treated as software source under their original licenses.
- **Opt-out mechanism**: contact `c.saillant@gmail.com` (Ailiance) to remove specific samples; nous respectons l'Article 4(3) de la directive DSM (TDM reservations).
- **Reservations of rights**: nous honorons `robots.txt`, les meta tags HTML `noai` / `noimageai`, et le protocole TDM Reservation Protocol (TDMRep) quand discoverable sur les repos sources.

## Pipeline reproducibility
See https://github.com/ailiance/ailiance-bench/tree/main/scripts:
- `kicad9plus_pipeline.sh`, `build_kicad9plus_dataset.py`
- Original audit: https://github.com/ailiance/ailiance-bench/blob/main/docs/audit_kicad9plus.md

## Provenance & upstream attribution

This dataset is co-published with [`electron-rare/kicad9plus-copyleft`](https://huggingface.co/datasets/Ailiance-fr/kicad9plus-copyleft) under the same **GPL-3.0-or-later** license. Original collection, curation, and pipeline tooling: **electron-rare** (upstream contributor). Production maintenance, EU AI Act packaging, and downstream support: **Ailiance** (this org).

Audit log (legal attribution, EU AI Act July 2025 template alignment): see [`docs/audit_kicad9plus.md`](https://github.com/ailiance/ailiance-bench/blob/main/docs/audit_kicad9plus.md) and the companion [`docs/audit_mascarade_se_attribution.md`](https://github.com/ailiance/ailiance-bench/blob/main/docs/audit_mascarade_se_attribution.md) on GitHub.

## About Ailiance

🇫🇷 **Ailiance** is a French AI organisation building EU-compliant resources for embedded systems and electronics design. Ailiance curates open datasets and fine-tuned models targeting:

- 🇪🇺 EU AI Act compliance (aligned with the GPAI Code of Practice signatories: Anthropic, Mistral, Google)
- ⚡ Electronics, embedded systems, hardware design
- 🔬 SPICE simulation, KiCad PCB / schematic, EDA workflows
- 🇫🇷 French + English technical content

Maintainer contact: `c.saillant@gmail.com` — see also the public bench/audit repo: [electron-bench](https://github.com/ailiance/ailiance-bench).

## License & EU AI Act

**GPL-3.0-or-later**. Données collectées et curées dans le cadre du projet **electron-rare**, packagées et maintenues par **Ailiance** pour déploiement production aligné EU AI Act.

Compatible **EU AI Act** : voir les signataires du *GPAI Code of Practice* (Anthropic, Mistral, Google) et la documentation transparence : [electron-bench](https://github.com/ailiance/ailiance-bench).

Audit log: [`docs/audit_kicad9plus.md`](https://github.com/ailiance/ailiance-bench/blob/main/docs/audit_kicad9plus.md).

## Citation

```bibtex
@dataset{ailiance_kicad9plus_copyleft_2026,
  author       = {Ailiance},
  title        = {{Ailiance — KiCad 9+ Schematic Corpus (Copyleft)}},
  year         = {2026},
  publisher    = {Hugging Face},
  license      = {GPL-3.0-or-later},
  url          = {https://huggingface.co/datasets/Ailiance-fr/kicad9plus-copyleft},
  note         = {Co-published with upstream electron-rare/kicad9plus-copyleft}
}

@dataset{electron_rare_kicad9plus_copyleft_2026,
  author       = {electron-rare},
  title        = {{Upstream: kicad9plus-copyleft}},
  year         = {2026},
  publisher    = {Hugging Face},
  url          = {https://huggingface.co/datasets/Ailiance-fr/kicad9plus-copyleft}
}
```

## Related datasets

Ailiance dataset family (co-published with `electron-rare/*`):

- [Ailiance-fr/kicad9plus-permissive](https://huggingface.co/datasets/Ailiance-fr/kicad9plus-permissive) — KiCad 9+ schematics, permissive subset
- [Ailiance-fr/kicad9plus-copyleft](https://huggingface.co/datasets/Ailiance-fr/kicad9plus-copyleft) — KiCad 9+ schematics, copyleft subset
- [Ailiance-fr/kill-life-embedded-qa](https://huggingface.co/datasets/Ailiance-fr/kill-life-embedded-qa) — Kill_LIFE embedded knowledge base
- [Ailiance-fr/mascarade-stm32-dataset](https://huggingface.co/datasets/Ailiance-fr/mascarade-stm32-dataset) — STM32 & ARM Cortex-M Q&A
- [Ailiance-fr/mascarade-spice-dataset](https://huggingface.co/datasets/Ailiance-fr/mascarade-spice-dataset) — SPICE & analog simulation Q&A
- [Ailiance-fr/mascarade-iot-dataset](https://huggingface.co/datasets/Ailiance-fr/mascarade-iot-dataset) — IoT & connected devices Q&A
- [Ailiance-fr/mascarade-embedded-dataset](https://huggingface.co/datasets/Ailiance-fr/mascarade-embedded-dataset) — embedded systems generic Q&A

## Used to train models evaluated in `ailiance/ailiance-bench` v0.2

This dataset contributes to training data for hardware-domain LoRA
adapters benchmarked in the [Ailiance bench suite](https://github.com/ailiance/ailiance-bench).

Phase 6 scoreboard verdicts (7-task KiCad/SPICE evaluation):
- 🥇 `eu-kiki`: champion 4/7 tasks (DSL/PCB/SPICE/extract)
- 🥇 `mascarade-embedded`: champion P3 extraction (+48 pts)
- ⚠️ `mascarade-kicad`: catastrophic forgetting on SPICE/P2/P3

See full scoreboard:
[ailiance-bench README#scoreboard-lora-phase-6](https://github.com/ailiance/ailiance-bench#scoreboard-lora-phase-6--2026-05-11).

> **Note**: this corpus is GPL-3.0 (upstream KiCad library symbols).
> Use of derived models trained exclusively on this dataset must comply
> with GPL distribution requirements.