fix: dbt microbatch parameter conversion

[eakmanrq]

Fixes two misunderstanding of the dbt microbatch:

• SQLMesh batch_size is always 1 for dbt microbatch. In other words, they never issue a query that spans multiple intervals.
• concurrent_batches really determines, in SQLMesh terms, if a model depends on self which ultimately determine if the snapshot depends on past. Like SQLMesh, dbt determines this automatically by looking for self-referenes but users can force this with the concurrent_batches flag. Therefore if we see is False, meaning it should run sequentially, then we add the model to the depends_on.