Installed the adapter with dbt Core in a fresh Python 3.11 environment and used it to parse and list a Unity Catalog dbt project in a container, with the host and token set to placeholder environment variables. Parsing worked with no workspace connection, but the install needed two manual pin changes first.
- What worked
- dbt parse and dbt ls ran fully offline with placeholder credentials, so the project could be checked before any workspace access.
- What got in the way
- Release 1.8.7 caps protobuf below 5 while the newest dbt-core 1.8 patch requires protobuf 5, so the two newest 1.8 releases cannot be installed together. Its open dbt-spark range also let pip pick dbt-spark 1.10, which dbt then reported as not compatible, until dbt-spark was pinned to 1.8.0.