new-intelligent-embedded-dashboard ✨

An AI-powered Databricks App that generates AI/BI (Lakeview) dashboards from a natural-language prompt. A ReAct agent probes your data (cardinality, ranges, null rates) before choosing widgets, streams its reasoning live, validates every generated query, and stages a dashboard for review before creating it. Packaged as a Databricks Asset Bundle.

🏢 Deploy it in your workspace

Deploying for your business users? See DEPLOYMENT.md for the full guide. It's bring-your-own-infra (you point it at an existing SQL warehouse + Lakebase instance), runs as its own service principal (no personal access token), and is per-user aware (each user gets their own profile, sessions, dashboards, and preferences).

The short version:

# 1. In app.yml, set DATABRICKS_DATABASE_INSTANCE_NAME to your Lakebase instance name.
# 2. Deploy the bundle pointing at your existing warehouse + Lakebase:
databricks bundle deploy -t standalone \
  --var warehouse_id=<id> --var database_instance_name=<name> -p <profile>

# 3. Start + deploy the app code:
APP=ddil-app-standalone
ME=$(databricks current-user me | sed -n 's/.*"userName": *"\([^"]*\)".*/\1/p')
databricks apps start "$APP" -p <profile>
databricks apps deploy "$APP" \
  --source-code-path "/Workspace/Users/$ME/.bundle/ddil-app-standalone/standalone/files/.build" -p <profile>

# 4. Grant the app's service principal access to the catalogs your users will use:
scripts/grant_app_access.sh "$APP" <warehouse-id> <catalog> --apply -p <profile>

# 5. Give business users CAN_USE on the app (workspace UI), then share its URL.

🛠️ Tech Stack

This application leverages a powerful, modern tech stack:

• Backend 🐍 Python + FastAPI
• Frontend ⚛️ React + shadcn/ui
• API Client 🔄 Auto-generated with orval from OpenAPI schema

🚀 Quick Start

Development Mode

Start all development servers (backend, frontend, and OpenAPI watcher) in detached mode:

uv run apx dev start

This will start an apx development server, which in it's turn runs backend, frontend and OpenAPI watcher. All servers run in the background, with logs kept in-memory of the apx dev server.

📊 Monitoring & Logs

# View all logs
uv run apx dev logs

# Stream logs in real-time
uv run apx dev logs -f

# Check server status
uv run apx dev status

# Stop all servers
uv run apx dev stop

✅ Code Quality

Run type checking and linting for both TypeScript and Python:

uv run apx dev check

📦 Build

Create a production-ready build:

uv run apx build

🚢 Deployment

To deploy the app into a Databricks workspace (for yourself or for business users), follow DEPLOYMENT.md — it covers prerequisites, the --var infra inputs, the service-principal grant (via scripts/grant_app_access.sh), and giving users access. A quick summary is in the Deploy it in your workspace section above.

---

Built with ❤️ using apx