Lakehouse_Plumber

Lakehouse Plumber (LHP): the tool that compiles declarative YAML into Databricks Lakeflow/DLT Python. Use whenever the user is working inside an LHP project. Dead-giveaway signals: the word `flowgroup`; the `lhp` CLI (`init`/`validate`/`generate`/`dag`/`diff`); or LHP files (`lhp.yaml`, `substitutions/`, `presets/`, `templates/`, `blueprints/`, `pipelines/*.yaml`). Covers scaffolding a new LHP project; authoring or fixing flowgroup load/transform/write/test actions; regenerating Python after editing YAML and checking the diff; reusing one pattern across regions/tenants/sites via templates or blueprints; setting up substitutions or Asset Bundle integration; and debugging any LHP validate/generate failure (unresolved `${token}`, non-streaming source) including the SQL or YAML behind the generated Python. Fire even when the user only says `flowgroup` or `lhp` with little else. Skip for hand-written DLT/Spark notebooks, generic SQL, Genie, or CI/CD with no LHP project involved.

Lakehouse Plumber (LHP): the tool that compiles declarative YAML into Databricks Lakeflow/DLT Python. Use whenever the user is working inside an LHP project. Dead-giveaway signals: the word `flowgroup`; the `lhp` CLI (`init`/`validate`/`generate`/`dag`/`diff`); or LHP files (`lhp.yaml`, `substitutions/`, `presets/`, `templates/`, `blueprints/`, `pipelines/*.yaml`). Covers scaffolding a new LHP project; authoring or fixing flowgroup load/transform/write/test actions; regenerating Python after editing YAML and checking the diff; reusing one pattern across regions/tenants/sites via templates or blueprints; setting up substitutions or Asset Bundle integration; and debugging any LHP validate/generate failure (unresolved `${token}`, non-streaming source) including the SQL or YAML behind the generated Python. Fire even when the user only says `flowgroup` or `lhp` with little else. Skip for hand-written DLT/Spark notebooks, generic SQL, Genie, or CI/CD with no LHP project involved.