v0.7.0 dbt_lever

PR #21 includes the following updates:

Features

• For Fivetran Lever connectors created on or after July 27, 2024, the USER and INTERVIEWER_USER source tables have been renamed to USERS and INTERVIEW_USER, respectively. This package now prioritizes the USERS and INTERVIEW_USER tables if available, falling back to USER and INTERVIEWER_USER if not.
  • If you have both tables in your schema and would like to specify this package to leverage the USER and/or INTERVIEWER_USER tables, you can set the variables lever__using_users and/or lever__using_interview_user to false in your dbt_project.yml.
  • For more information, refer to the July 2024 connector release notes and the related README section.
• Introduced the ability to union source data from multiple Lever connectors. For more details, see the related README section.
• Streamlined testing by removing tests from intermediate models and consolidating them in the end models, optimizing resource usage and prioritizing the final outputs.

Bug fixes

• Fixed an issue where the dbt package would error due to a missing CONTACT_LINK source table for users without source data, even though it was enabled in the Fivetran Connector. A null-filled table will now be generated in such cases.

Documentation updates

• Add missing field definitions to dbt docs.

Under the hood

• In the source package, updated temporary models to union source data using the fivetran_utils.union_data macro.
• Added the source_relation column in each staging model to identify the origin of each field, utilizing the fivetran_utils.source_relation macro.
• Updated tests to include the new source_relation column.
• Included the source_relation column in all joins and window function partition clauses in the transform package.
• Added consistency tests for end models.
• Turned off freshness tests for the USER/USERS and INTERVIEW_USER/INTERVIEWER_USER sources to avoid possible conflicts.

Full Changelog: v0.6.0...v0.7.0