A compact, reproducible neuron-level Drosophila escape-circuit graph derived from the pinned MaleCNS source used by FlyEye.
Dataset Card
By Tubban, published under cc-by-4.0, revision 61697c11c3e5.
A compact, reproducible neuron-level Drosophila escape-circuit graph derived from the pinned MaleCNS source used by FlyEye.
Dataset summary
| Graph | Neurons | Edges | Bytes |
|---|---|---|---|
| pinned source subset used by builder | 70,000 | 798,715 | 9,864,604 |
| FlyEye escape-neuron-v1 | 3,376 | 78,797 | 959,092 |
The compact graph contains:
- 126 LC4 / looming seed neurons;
- 26 selected direct escape descending-neuron targets;
- directed LC4-to-output pathways within a maximum 3-hop corridor;
- all signed internal edges between retained neurons;
- original source indices;
- MaleCNS body IDs;
- source and derived graph hashes;
- deterministic extraction metadata.
Files
graph.bin— FLYGRAPH v1 binarymanifest.json— IDs, groups, provenance and scientific boundaryreport.json— extraction/compression reportescape-fast-v1.json— lightweight aggregate profileCITATION.cffDATA_LICENSE.md
FLYGRAPH v1 binary format
8 bytes magic "FLYGRAPH"
u32 LE version
u32 LE neuronCount
u32 LE edgeCount
u32[n+1] rowOffsets
u32[e] presynapticIndices
f64[e] signed weights
Rows are postsynaptic neurons.
Scientific boundary
The retained neuron-level edges are real signed edges selected from the pinned MaleCNS-derived graph. The pathway selection itself is task-specific.
This dataset does not claim that FlyEye's camera encoding, thresholds, aggregate runtime dynamics, or behavioral readouts are measured biological physiology. Those components are modeled unless separately validated.
Licensing and attribution
MaleCNS-derived graph assets retain the upstream CC BY 4.0 attribution recorded in the manifest.
Source attribution:
Janelia FlyEM MaleCNS — https://male-cns.janelia.org/download/
FlyEye software and extraction tooling are licensed separately under MIT.
Reproducibility
The source repository pins the upstream graph source and contains the graph builder, CI workflow, graph integrity tests, JavaScript/TypeScript loaders and a Python SDK.
Source:
https://github.com/tubban1/fly-eye
Citation
If you use this dataset or the FlyEye Developer Kit in research, cite the archived software release:
FlyEye Contributors (2026). FlyEye Developer Kit: reusable Drosophila connectome graphs and runtime tooling (v0.5.0-alpha.9). Zenodo. https://doi.org/10.5281/zenodo.22829797
@software{flyeye_developer_kit_2026,
author = {{FlyEye Contributors}},
title = {FlyEye Developer Kit: reusable Drosophila connectome graphs and runtime tooling},
year = {2026},
version = {0.5.0-alpha.9},
doi = {10.5281/zenodo.22829797},
url = {https://doi.org/10.5281/zenodo.22829797}
}
Related resources:
- Live demo: https://fly.fde.fan
- Source: https://github.com/tubban1/fly-eye
- GitHub release: https://github.com/tubban1/fly-eye/releases/tag/v0.5.0-alpha.9
- Zenodo archive: https://doi.org/10.5281/zenodo.22829797
Details
- Repository
- Tubban/flyeye-escape-neuron-v1
- Publisher
- Tubban
- Task category
- Not stated by the source
- Tags
- drosophila, fruit-fly, connectome
- Size category
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 61697c11c3e523a78e6c09fa48948547d409dd6d
- Last updated
- 2026-09-18
Files
8 files, 1.1 MB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| escape-fast-v1.json | Data | 4.1 KB | — |
| manifest.json | Data | 93.4 KB | — |
| report.json | Data | 4.6 KB | — |
| DATA_LICENSE.md | Documentation | 1.2 KB | — |
| README.md | Documentation | 3.5 KB | — |
| CITATION.cff | Other | 1.3 KB | — |
| graph.bin | Other | 959.1 KB | b2378c23f95a |
| .gitattributes | Repository | 2.5 KB | — |
License and Download
- License
- cc-by-4.0
- Access
- No access gate
Released by Tubban through its official repository on Hugging Face. Read the license.