[WIP] Add session mode for dbt-databricks adapter for 1.10.x version

[alexeyegorov]

Execute DBT in session mode (e.g. on a job cluster)

This PR was created with help by Claude/Cursor to implement session mode for dbt-databricks adapter.

It should allow:

• execution of sql and python models using a session like in dbt-spark
• run the whole dbt pipeline on a job cluster to save costs
• ...

Asset Bundle DBT task

The native dbt_task on Databricks does not provide a Spark session. It is also not possible to retrieve it.
It is about Databricks to allow retrieving a Spark Session within dbt_task in order to keep the Asset Bundle deployment as simple as it is right now.

The workaround to still use the session mode is to define the dbt tasks as python scripts or notebooks.
This approach could be added as a template via Databricks Asset Bundles.

In our job example, this looks as below:

With the job cluster selected, it now executes as expected:

• prepare:
  • clone the repository with DBT code
  • run dbt deps and dbt seed
• run: execute dbt run with optional --full-refresh and a passed --select (e.g. state, specific model, or empty for full selection)
• test: execute dbt test
• docs: generate dbt docs
• cleanup: remove cloned repository files

Example of dbt cli for run execution

Pros/Cons

Pros:

• keep profiles.yml and requirements.txt in the asset bundles repository
  • currently, profiles.yml need to be uploaded manually to an available path on e.g. Databricks; changes require to update and reupload the file;
  • this strategy allows keeping this file in asset bundles repository or in the original dbt repository and link the job to use directly after it was cloned
• execute the complete DBT pipeline with SQL and Python models on both, all-purpose and job clusters as well as SQL warehouse
• use a cluster with init script to install 3rd party libraries (e.g. Apache Sedona for geospatial functions)

Cons:

• not able to use the native dbt_task provided by Databricks -> could be adjusted by Databricks
• slightly more setup for the definition of the tasks (Python notebooks using dbt CLI) -> could be added as a template to Asset Bundles repository

Description

Checklist

• I have run this code in development and it appears to resolve the stated issue
• This PR includes tests, or tests are not required/relevant for this PR
• I have updated the CHANGELOG.md and added information about my change to the "dbt-databricks next" section.