bshada/open-schematics-v4

Open Schematics Dataset

A comprehensive dataset of electronic schematics and PCB layouts from open-source hardware projects. Designed for training AI models on circuit design, component recognition, PCB layout understanding, and hardware engineering tasks.

Dataset Description

Each record corresponds to one schematic file and includes the raw source, rendered images, structured metadata, and all associated PCB files from the same project.

Dataset Structure

Field	Type	Description
schematic	string	Raw schematic file content (.kicad_sch, .sch, or .SchDoc)
schematic_image	image	Rendered PNG of the schematic (None if unavailable)
pcb_files	list[string]	Raw content of PCB files (.kicad_pcb / .pcb) from the same project
pcb_images	list[image]	Rendered PNGs of each PCB file, parallel to pcb_files (None per entry if unavailable)
components_used	list[string]	Component names extracted from library symbols (.kicad_sch only; None for other formats)
schematic_json	string	kiutils JSON representation of the schematic (.kicad_sch only; None otherwise)
schematic_yaml	string	kiutils YAML representation of the schematic (.kicad_sch only; None otherwise)
name	string	Source repository (owner/repo)
description	string	Repository description
extensions_used	list[string]	File extensions present: schematic extension first, then unique PCB extensions (e.g. [".kicad_sch", ".kicad_pcb"])

Schematic formats included

• .kicad_sch — KiCad v5+ (primary format; includes JSON/YAML/components)
• .sch — KiCad v4 legacy
• .SchDoc — Altium Designer

PCB filters applied

• PCB files larger than 10 MB are excluded (extreme outliers; affects <1% of files)
• Projects with more than 20 PCB files retain the 20 with the largest rendered image (highest detail), ranked by pixel area

Image generation

PNG images are sourced directly from pre-rendered files. When a PNG is absent but an SVG is available, the PNG is generated on the fly from the SVG using cairosvg at 150 DPI.

Dataset Creation

Source Data

Schematics and PCB layouts collected from publicly available hardware repositories on GitHub.

Data Collection

Files were fetched and processed to extract:

• Raw schematic and PCB source files
• Rendered images (PNG/SVG → PNG)
• Structured metadata via kiutils (JSON/YAML)
• Component lists from library symbol references

Data Processing

• Empty .kicad_sch files (no library symbols or schematic symbols) are filtered out
• .sch and .SchDoc rows are included with raw text only; JSON/YAML/components are None when kiutils conversion is unavailable
• Component names are extracted from libSymbols fields and deduplicated per library prefix
• extensions_used is inferred from actual file paths on disk, not from YAML metadata fields

Use Cases

• Circuit Design AI: Training models to generate or understand electronic schematics
• PCB Layout Understanding: Learning the relationship between schematics and their physical PCB layouts
• Component Recognition: Identifying electronic components from schematic and PCB diagrams
• Design Assistance: Building tools to help engineers with circuit and PCB design
• Multimodal Learning: Paired schematic text + image + PCB text + image for cross-modal tasks
• Documentation Generation: Automatically generating documentation from schematics
• Design Validation: Training models to detect errors or suggest improvements
• Educational Tools: Creating interactive learning systems for electronics education

Limitations

• JSON/YAML/components are only populated for .kicad_sch format (~85% of rows)
• PCB images are unavailable for ~55% of PCB entries (no PNG or SVG rendered)
• Schematic images are unavailable for ~10% of rows
• Quality and complexity of designs vary significantly across projects
• Component naming conventions vary between projects and library sources

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC-BY-4.0) license. Please ensure you comply with the original licenses of the source projects when using this data.

Citation

If you use this dataset in your research or projects, please cite:

@dataset{open_schematics_2025,
  title={Open Schematics: A Dataset of Electronic Schematics and PCB Layouts from Hardware Projects},
  author={bshada},
  year={2025},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/datasets/bshada/open-schematics}}
}

Contact

For questions, issues, or contributions, please open an issue on the dataset repository.