PeytonT/repo_graph

---
pretty_name: Repository Graph Universe
viewer: true
tags:
- datasets
- code
- graph
- software-engineering
- repository-graphs
size_categories:
- 10M<n<100M
configs:
- config_name: repos
default: false
data_files:
- split: train
path: "repos/*.parquet"
- config_name: entities
default: false
data_files:
- split: train
path: "entities/*.parquet"
- config_name: edges
default: false
data_files:
- split: train
path: "edges/*.parquet"
- config_name: artifacts
default: false
data_files:
- split: train
path: "artifacts/*.parquet"
- config_name: universe_nodes
default: true
data_files:
- split: train
path: "universe_nodes/*.parquet"
- config_name: universe_edges
default: false
data_files:
- split: train
path: "universe_edges/*.parquet"
- config_name: repo_knn
default: false
data_files:
- split: train
path: "repo_knn/*.parquet"
---

## Created with Research Library:

https://github.com/peytontolbert/Research_Library

# Repository Graph Universe Dataset

Parquet-first export of the repository graph artifacts already built under the Repository Library.

This dataset preserves:

- per-repo raw graph exports (`entities`, `edges`, `artifacts`)
- repo manifest metadata (`repos`)
- the aggregated universe graph when available (`universe_nodes`, `universe_edges`, `repo_knn`)

## Coverage

- repos in library manifest: `178`
- repos represented in universe: `135`
- node embeddings split included: `false`

## Configs

- default viewer/config: `universe_nodes`
- `repos`: `178` rows
- `entities`: `906024` rows
- `edges`: `8402561` rows
- `artifacts`: `453278` rows
- `universe_nodes`: `216987` rows
- `universe_edges`: `5518898` rows
- `repo_knn`: `1350` rows
TASET", "repo_knn")

## Notes

- `repos` joins universe repo coordinates into the repo metadata rows when available.
- `universe_nodes` includes `x`, `y`, `z` coordinates from the existing 3D projection.
- `artifacts` only contains program URIs and hashes, not file contents.

## Universe Assets

The local export also includes these visualization assets when present:

- `universe_3d.png`
- `universe_3d_detailed.png`
- `nodes_3d_sample.html`
- `universe_3d_hover.html`

### Universe 3D Detailed View


- [Open the sampled 3D HTML view](./nodes_3d_sample.html)

- [Open the interactive hover view](./universe_3d_hover.html)