[nabinchha]

📋 Summary

Ensure all four published packages include the complete Apache license text and NVIDIA copyright notice. Complete contributor sign-off instructions, source headers, and external-material notices while preserving tutorial notebook metadata.

🔗 Related Issue

N/A — package licensing and contribution-documentation corrections.

🔄 Changes

• Synchronize the root and four package license files, explicitly declare packaged license files, and explain the independently distributed Slurm copy in README.
• Add per-commit DCO sign-off instructions, command and trailer examples, the existing first-time-contributor bot signature step, and the complete unchanged certification to both contribution guides.
• Add missing notices to Makefile, six tutorial sources, their tracked Colab outputs, and the local-preview stylesheet. Preserve Jupytext front matter when maintaining headers, including in empty notebooks.
• Add the full external-materials notice to README and the published installation page, and clarify the scope of license checks for separately distributed bundles and containers.
• Extend the package-install CI check to build all wheels and source archives and verify complete license text plus metadata. Add regression tests for license delivery and notebook-safe header updates.

🔍 Attention Areas

The bundled ZIP-to-area-code asset is unchanged. Confirming its original source and applicable redistribution terms remains a separate follow-up; this PR does not resolve that provenance item.

🧪 Testing

• Focused documentation and package-license suites: 84 passed, 2 skipped.
• All four wheels and four source archives contain the canonical license and correct license metadata; strict Twine validation passed for all eight artifacts.
• Full package-install CI script: base install, Slurm extra, direct Slurm install, CLI discovery, and startup-time checks passed.
• All six tutorials preserve existing cells and metadata and retain notices through Jupytext, Colab, and Fern conversion. Tracked Colab outputs match fresh generation, ignoring random cell IDs.
• Header check: 869 files, no updates needed. Changed-source lint/format and whitespace checks passed.
• Fern validation: zero errors and one authentication warning for redirects; internal links passed across 101 pages.
• Full application suite: not rerun; changes are limited to packaging, maintenance scripts, and documentation.

✅ Checklist

• Follows commit message conventions
• Commits are signed off (DCO)
• Published contribution and installation documentation updated
• External-contributor triage requirement: N/A — repository collaborator.
• Architecture docs: N/A — package layering and runtime architecture are unchanged.

[github-actions]

Fern preview: https://nvidia-preview-pr-1133.docs.buildwithfern.com/nemo/datadesigner

Fern previews include the docs-website version archive with PR changes synced into latest. Notebook tutorials are rendered without execution outputs in previews.

[andreatnvidia]

Thanks @nabinchha for covering the package licenses, notices, and DCO updates. Requesting changes to sync the generated Colab notebooks before this reaches main. There’s also a smaller, non-blocking correction to the DCO instructions.

[andreatnvidia]

[P2] Sync the generated Colab notebooks

These headers create an extra Jupytext cell in all six tutorials, but the tracked Colab notebooks are unchanged and omit the notices. Please run make generate-colab-notebooks and commit the updated outputs. The notebook-sync check only runs against main, so it hasn’t caught the mismatch on this PR’s Slurm base branch.

[nabinchha]

Fixed in 1f2802f. Ran make generate-colab-notebooks and committed all six regenerated Colab notebooks. Verified that each tracked output matches fresh generation when ignoring random cell IDs and contains the license notice; unchanged cell IDs were retained to reduce diff noise. The focused documentation/package-license suites passed: 84 tests, 2 skips.

[andreatnvidia]

[P3] Keep the DCO bot-signing step documented

Could we retain the bot-signing instructions alongside git commit -s in both contribution guides? The existing workflow still requires a first-time contributor’s PR-comment signature, so following these instructions alone won’t satisfy the DCO check. This is non-blocking, but documenting both steps would avoid confusion.

[nabinchha]

Fixed in 1f2802f. Both contribution guides now document the first-time DCO Assistant signature alongside the per-commit git commit -s requirement, including the exact PR-comment phrase configured in the workflow and the recheck command. The full DCO certification remains unchanged. Fern validation and internal-link checks passed.

[greptile-apps]

Confidence Score: 5/5

[Medium impact] This PR appears safe to merge.

Summary

This PR adds complete license copies to the published packages and checks their contents in wheels and source archives.

• Each published package carries the full Apache license with its release.
• The contribution guides show how to sign off every commit.
• The published docs explain the terms around external materials and separate bundles.
• Tutorial notebook settings stay in place when license headers are added.

nabinchha explicitly deferred the unchanged ZIP-to-area-code asset’s source and redistribution terms. This review does not treat that item as resolved.

_{Reviews (1) · Last reviewed commit: "docs: sync notebook notices and DCO sign..." · Reviewed by Greptile}