"Routes OGB dataset-export workflows for DatasetSaver and
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill dataset-contribution --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Dataset Contribution?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-dataset-contribution)More formats (shields.io, HTML) on the badges page.
---
name: dataset-contribution
description: "Routes OGB dataset-export workflows for DatasetSaver and
OGB-compatible releases."
metadata:
disco-role: operating
disable-model-invocation: true
license: MIT
---
# Dataset Contribution
Use this subskill when the task is about packaging a new OGB-compatible dataset
with `DatasetSaver`, generating a `meta_dict.pt`, or validating the layout of a
submission archive.
## Use this subskill when
- The task names `DatasetSaver` or `save_graph_list` / `save_target_labels` /
`save_split` / `copy_mapping_dir` / `get_meta_dict`.
- The task asks how to prepare `meta_dict.pt`, `mapping/`, or a release zip for
an external dataset contribution.
- The task needs a tiny end-to-end smoke of the export pipeline.
## First decisions
1. Read [`references/workflows.md`](references/workflows.md) for the step order
and family constraints.
2. Read [`references/data-formats.md`](references/data-formats.md) for graph,
split, label, and mapping directory shapes.
3. Read [`references/api-reference.md`](references/api-reference.md) for the
method order and parameter expectations.
4. Read [`references/troubleshooting.md`](references/troubleshooting.md) when a
shape, mapping, or zip step fails.
5. If you want a safe end-to-end smoke, run
[`scripts/datasetsaver-tiny-smoke.py`](scripts/datasetsaver-tiny-smoke.py).
## Main workflow
- Decide whether the dataset is graph-, node-, or link-oriented.
- Build the graph list in the format OGB expects.
- Save labels and splits in the same order that the export helper expects.
- Copy the mapping directory only after the raw and split files are present.
- Generate the `meta_dict.pt` and zip the prepared directory.
- Reload the tiny smoke dataset to confirm the archive structure still works.
## Common routing choices
- Use this subskill for new OGB-compatible releases.
- Use the graph/node/link subskills when the user only needs to consume an
existing OGB dataset, not package a new one.
## What not to do here
- Do not tell the user to hand-edit the packaged release after the zip step.
- Do not assume heterogeneous `ogbg` export is implemented.
- Do not depend on the original checkout for final runtime guidance.
## Related references
- [`../../references/api-overview.md`](../../references/api-overview.md)
- [`../../references/troubleshooting.md`](../../references/troubleshooting.md)
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!