Docs · 01
Quickstart
From pip install to your first verdict in a few minutes.
1. Install
pip install "bricksentry[databricks,mcp]"
# local demo only, no Databricks needed
pip install "bricksentry[duckdb]"2. Initialise
bricksentry init writes two files in the current directory: bricksentry.yml (backend, tables, lineage settings) and policy.yml (the rules that decide verdicts).
bricksentry init3. Configure the backend
Set warehouse_id under backend in bricksentry.yml, or export DATABRICKS_WAREHOUSE_ID. Authentication uses the standard Databricks SDK chain: environment variables (DATABRICKS_HOST, DATABRICKS_TOKEN or OAuth client credentials), a ~/.databrickscfg profile, or the CLI login.
backend: {type: databricks, warehouse_id: abc123}4. First commands
bricksentry impact main.sales.orders # who depends on this table
bricksentry diff main.sales.orders main.sales_pr_42.orders
bricksentry preview main.sales.orders --sql "DELETE FROM {{target}} WHERE status <> 'paid'"
bricksentry assess --from-git origin/main --pr 42
bricksentry rollback main.sales.orders # restore command from Delta historyExit codes
| Code | Meaning |
|---|---|
0 | Verdict is below the --fail-on level. |
1 | Verdict is at or above the --fail-on level. |
2 | Configuration error. |