Add orchestration assets and ops for de-identification dataset jobs

[maziyarpanahi]

Closes #845.

Summary

• Adds an optional orchestration adapter with a configurable batch op and daily-partitioned redacted dataset asset.
• Validates canonical policy profiles, redacts configured columns with process_batch, and emits count- and label-only metadata.
• Adds synthetic in-memory tests and a runnable example while keeping imports safe without the optional dependency.

Validation

• .venv/bin/python -m pytest tests/unit/integrations/test_dagster_assets.py -q — 3 passed
• .venv/bin/ruff check openmed/integrations/dagster_assets.py tests/unit/integrations/test_dagster_assets.py examples/dagster/definitions.py
• .venv/bin/ruff format --check openmed/integrations/dagster_assets.py tests/unit/integrations/test_dagster_assets.py examples/dagster/definitions.py
• /Users/maziyar/.local/bin/uv lock --check
• git diff --check

[maziyarpanahi]

Thank you @maziyarpanahi. I reviewed this against #845 and current master, then added maintainer follow-ups cdbedeb4, d097f83f, and 1341bb01.

What changed:

• Made non-text Dagster asset values fail closed before de-identification.
• Completed the optional dependency, lockfile, permissive-license, and stable installation guidance.
• Integrated the current SQLAlchemy dependency state without dropping either optional extra.

Validation:

• .venv/bin/python -m pytest tests/unit/integrations/test_dagster_assets.py -q — 5 passed on the final head with Dagster installed.
• Combined Dagster and SQLAlchemy focused tests — 8 passed.
• .venv/bin/python -m pytest tests/unit -q — 11,593 passed, 89 skipped on the exact aggregate release tree.
• uv lock --check, Ruff lint, Ruff format check, package build, and dependency-license policy — passed.

This now satisfies the issue and is ready to merge.