[dan-newton-sparta]

Attach a jobType facet and an ownership facet to every emitted job.

Summary & Motivation

Emitted jobs carried neither a jobType nor an ownership facet, so a lineage backend could not tell that a pipeline was a Dagster job, nor who owned it.

This adds:

• a jobType facet (integration=DAGSTER, processingType=BATCH) on every job, so backends attribute the pipeline to Dagster rather than a generic fallback; and
• an ownership facet. The asset-emitting methods take an owners argument (the asset's own Dagster owners — team:<team> / email), so different assets can be owned by different teams. Jobs with no per-asset owners fall back to a default team from OPENLINEAGE_TEAM or the team= argument; with neither, no ownership facet is emitted.

This lets a backend such as OpenMetadata route the pipeline to its Dagster service and set the pipeline owner from the team. The sensor wires each asset's native owners into the owners argument in a follow-up (see Related changes). Backward compatible: no owners and no default team reproduces the previous output.

How I Tested These Changes

• Unit tests: jobType present on every job; per-asset owners override the default team; default-team fallback via team= and OPENLINEAGE_TEAM; arg-over-env precedence; no owners → no facet; owners threaded onto a real materialization's job.
• uv run pytest green; ruff check, ruff format --check, and ty check . clean.
• Tested locally end to end: ingested the emitted OpenLineage events into OpenMetadata and used it to visualise the resulting lineage.

Changelog

Added a ### Features entry under ## [Unreleased] in CHANGELOG.md.

Related changes

Part of a set of small improvements to Dagster → OpenLineage. The lineage-fidelity changes are each independent and mergeable on their own; the ownership pair is stacked (#354 builds on #353).

Lineage fidelity:

• feat(dagster-openlineage): decouple job namespace from dataset namespace #344 — decouple job namespace from dataset namespace
• fix(dagster-openlineage): emit asset run-start without a placeholder output #345 — emit asset run-start without a placeholder output
• feat(dagster-openlineage): derive asset inputs from column lineage #346 — derive asset inputs from column lineage
• feat(dagster-openlineage): add exclude_asset_keys to the storage wrapper #347 — add exclude_asset_keys to the storage wrapper

Ownership:

• feat(dagster-openlineage): emit jobType and per-asset ownership facets #353 — emit jobType and per-asset ownership facets
• feat(dagster-openlineage): read native asset owners in the sensor #354 — read native asset owners in the sensor (builds on feat(dagster-openlineage): emit jobType and per-asset ownership facets #353; until feat(dagster-openlineage): emit jobType and per-asset ownership facets #353 merges its diff includes feat(dagster-openlineage): emit jobType and per-asset ownership facets #353's commit)