NOTICE.md

Dataset migration notice

The canonical Dataset has moved

Please migrate new integrations from this Bucket to the direct-V2 Hugging Face Dataset:

scholarweave/Jobs-Dataset-v2

The new repository is the canonical source for current and future Dataset publishes. This Bucket is retained as legacy migration history and should not be used as the source for new production integrations.

What changes

The direct-V2 Dataset provides:

• A materialized current-active view, so consumers no longer need to replay historical events to obtain open jobs.
• Immutable dated changes/ files for additions and material updates.
• Immutable dated closed/ files for detected closures.
• Separate minimal and full variants, plus published company snapshots where available.
• A versioned data/manifests/current.json that selects the exact active shards for a Dataset revision.

Migration guidance

1. Point your client at scholarweave/Jobs-Dataset-v2.
2. Read data/manifests/current.json at a pinned revision and download only the active shard paths it lists.
3. For incremental processing, consume dated data/changes/ files and then the corresponding data/closed/ files. Upsert changes by posting_id and remove matching active rows for closures.
4. Use the minimal active shards for search and analytics. Use full files only when raw or normalized payload data is required.

See the Dataset repository README for the complete layout, schemas, and download examples. Pin a Hugging Face commit SHA when reproducibility matters.

Existing Bucket objects remain useful for historical migration and lineage, but they do not define the current active-job view.