Merge pull request #2 from databricks-solutions/create-genie-sz

feat: agentic Genie Space creation with conversational agent

Author: hiydavid

.databricksignore

@@ -51,5 +51,20 @@ README.md
 .venv/
 venv/
 
-# MLflow local runs
+# MLflow local runs and database
 mlruns/
+mlflow.db
+
+# Other non-runtime files
+*.md
+!backend/references/schema.md
+CODEOWNERS.txt
+LICENSE.md
+NOTICE.md
+SECURITY.md
+notebooks/
+scripts/
+sql/
+.databricks/
+.gitleaksignore
+.gitignore

.env.example

@@ -42,9 +42,12 @@ GENIE_TARGET_DIRECTORY=
 # =============================================================================
 LAKEBASE_HOST=
 LAKEBASE_PORT=5432
-LAKEBASE_DATABASE=genie_workbench
-LAKEBASE_USER=postgres
-LAKEBASE_PASSWORD=
+LAKEBASE_DATABASE=databricks_postgres
+# Instance name — enables auto credential generation via Databricks SDK
+LAKEBASE_INSTANCE_NAME=
+# Override with static creds if needed (otherwise SDK generates them):
+# LAKEBASE_USER=your-email@company.com
+# LAKEBASE_PASSWORD=
 
 # =============================================================================
 # Local dev auth

.gitignore

@@ -20,7 +20,9 @@ __pycache__/
 .Python
 build/
 develop-eggs/
-# NOTE: dist/ is NOT ignored because frontend/dist/ needs to be deployed
+# frontend/dist/ is ignored in git (build artifact) but NOT in .databricksignore
+# so it still gets synced to the workspace for deployment.
+frontend/dist/
 # The .databricksignore file controls what gets synced to Databricks Apps
 downloads/
 eggs/
@@ -209,13 +211,14 @@ sketch
 
 **/uv.lock
 **/mlruns/
+mlflow.db
 **/.vite/
 **/.databricks
 **/.claude
 **/.env.local
 
 # Frontend build artifacts (Vite)
-# NOTE: frontend/dist/ is NOT ignored - it must be deployed to Databricks Apps
+# frontend/dist/ is deployed via databricks sync (not git)
 **/dist-ssr/
 
 # Frontend debug logs

.gitleaksignore

@@ -0,0 +1,6 @@
+# False positives from Vite-bundled React build artifacts (minified JS)
+dba8a46670f71554d7c6f85b8156c6ff8b927154:frontend/dist/assets/index-DfyblZUj.js:vault-service-token:8
+dba8a46670f71554d7c6f85b8156c6ff8b927154:frontend/dist/assets/index-BEGK0X5-.js:vault-service-token:8
+faff5e63c36691d4cd42c63cdd9c47c7a6e0fc71:frontend/dist/assets/index-BEGK0X5-.js:vault-service-token:8
+faff5e63c36691d4cd42c63cdd9c47c7a6e0fc71:frontend/dist/assets/index-Dkgq7VUt.js:vault-service-token:8
+fa428d65a8d509b4b544de096ccbcff4ef4a3d1a:frontend/dist/assets/index-Dkgq7VUt.js:vault-service-token:8

app.yaml

@@ -29,7 +29,7 @@ env:
   - name: MLFLOW_REGISTRY_URI
     value: "databricks-uc"
   - name: MLFLOW_EXPERIMENT_ID
-    value: ""  # OPTIONAL: Set to your MLflow experiment ID to enable tracing
+    value: "2624606869914616"
 
   # ---------------------------------------------------------------------------
   # LLM Model
@@ -38,17 +38,20 @@ env:
     value: "databricks-claude-sonnet-4-6"
 
   # ---------------------------------------------------------------------------
-  # SQL Warehouse (required for GenieIQ benchmark execution)
+  # SQL Warehouse — pulled from the app resource named "sql-warehouse".
+  # Configure the warehouse resource in the Databricks Apps UI.
+  # Falls back to auto-detect if empty.
   # ---------------------------------------------------------------------------
   - name: SQL_WAREHOUSE_ID
-    value: ""  # Required: your SQL warehouse ID
+    valueFrom: sql-warehouse
 
   # ---------------------------------------------------------------------------
-  # Genie Space target directory (required for creating new Genie Spaces)
-  # e.g. /Workspace/Users/you@company.com/
+  # Genie Space target directory (where new spaces are created)
+  # Defaults to /Shared/ which is writable by most service principals.
+  # Override to a specific folder if needed (e.g. /Workspace/Shared/Analytics/).
   # ---------------------------------------------------------------------------
   - name: GENIE_TARGET_DIRECTORY
-    value: ""
+    value: "/Shared/"
 
   # ---------------------------------------------------------------------------
   # Local dev auth (only used when running outside Databricks Apps)
@@ -57,43 +60,17 @@ env:
     value: ""
 
   # ---------------------------------------------------------------------------
-  # Lakebase PostgreSQL connection
-  # These values are injected automatically by the Lakebase resource binding
-  # defined in the `resources` section below. Do not set them manually here.
+  # Lakebase PostgreSQL — credentials auto-generated via Databricks SDK (SP)
+  # Grant the app's service principal CAN MANAGE on the Lakebase instance.
   # ---------------------------------------------------------------------------
   - name: LAKEBASE_HOST
-    valueFrom:
-      secretRef: lakebase-connection
-      key: host
+    value: ""  # e.g. instance-XXXX.database.cloud.databricks.com
   - name: LAKEBASE_PORT
-    valueFrom:
-      secretRef: lakebase-connection
-      key: port
+    value: "5432"
   - name: LAKEBASE_DATABASE
-    valueFrom:
-      secretRef: lakebase-connection
-      key: database
-  - name: LAKEBASE_USER
-    valueFrom:
-      secretRef: lakebase-connection
-      key: user
-  - name: LAKEBASE_PASSWORD
-    valueFrom:
-      secretRef: lakebase-connection
-      key: password
-
-# ---------------------------------------------------------------------------
-# Lakebase resource binding
-# Provisions a Lakebase PostgreSQL instance and injects connection credentials
-# as the secret reference "lakebase-connection" used by the env vars above.
-# Run sql/setup_lakebase.sql once after first deploy to create the schema.
-# ---------------------------------------------------------------------------
-resources:
-  - name: lakebase-connection
-    description: "PostgreSQL connection for Genie Workbench scan result persistence"
-    resource_type: "database"
-    config:
-      type: postgresql
+    value: "databricks_postgres"
+  - name: LAKEBASE_INSTANCE_NAME
+    value: ""  # e.g. genie-workbench-sz
 
 # ---------------------------------------------------------------------------
 # Service Principal authorization