nb-fabric 0.2.0

Open beta. Commands and flags may change between versions. Pin your version if stability matters.

Install

cargo install nb-fabric

Or download a prebuilt binary from Releases.

Prerequisites

• Azure CLI (az) installed and authenticated (az login)
• Access to a Microsoft Fabric workspace

Quick Start

# Check auth
nb auth status

# List notebooks
nb list "My Workspace"

# Create a Python notebook with a lakehouse attached
nb create "My Workspace/ETL Pipeline" --kernel python --lakehouse MainLH

# Add a cell
nb cell add "My Workspace/ETL Pipeline" --code "print('hello')"

# Execute interactively and see the output
nb exec "My Workspace/ETL Pipeline" --code "print('hello')" --lakehouse MainLH

# Run as a batch job
nb run "My Workspace/ETL Pipeline" --wait

Command Reference

Authentication

nb auth status                     Check Azure CLI authentication

Notebook CRUD

nb list <workspace>                List notebooks in a workspace
nb create <ws/name>                Create a new notebook
  --kernel python|pyspark            Kernel type (default: python)
  --lakehouse <name>                 Attach a lakehouse
  --warehouse <name>                 Attach a warehouse
nb export <ws/name> -o <path>      Export notebook to local .ipynb
nb open <ws/name>                  Open notebook in browser
nb delete <ws/name> --force        Delete a notebook

Cell Operations

nb cells <ws/name>                 List all cells (index, type, preview)
nb cell view <ws/name> <index>     View a single cell's source
nb cell add <ws/name>              Add a new cell
  --code <code>                      Cell content (required)
  --markdown                         Create markdown cell (default: code)
  --at <index>                       Insert at position (default: append)
nb cell edit <ws/name> <index>     Replace a cell's source
  --code <code>                      New content
nb cell rm <ws/name> <index>       Remove a cell

Execution

nb exec <ws/name>                  Execute code interactively via Livy
  --code <code>                      Code to execute
  <cell-index>                       Or execute a specific cell by index
  --lakehouse <name>                 Lakehouse for Livy session (auto-detected)
nb run <ws/name>                   Run notebook as batch job
  --wait                             Wait for completion
  --timeout <secs>                   Timeout in seconds (default: 3600)
nb runs <ws/name>                  List recent runs

Interactive execution auto-detects the kernel type from notebook metadata:

• Python notebooks use Livy kind=python
• PySpark notebooks use Livy kind=pyspark

Both return cell output directly to the terminal.

Sessions

nb session <ws/name>               Show active Livy sessions

Scheduling

nb schedule list <ws/name>         List schedules
nb schedule create <ws/name>       Create a schedule
  --type Cron|Daily|Weekly           Schedule type (default: Cron)
  --interval <n>                     Interval (minutes for Cron)
  --start <datetime>                 Start time (ISO 8601)
  --end <datetime>                   End time (optional)
  --timezone <tz>                    Timezone (default: UTC)
  --enable                           Enable immediately
nb schedule update <ws/name> <id>  Update a schedule
  --enable true|false                Enable or disable
  --interval <n>                     New interval
nb schedule delete <ws/name> <id>  Delete a schedule

How It Works

nb authenticates via Azure CLI (az account get-access-token) using the same credentials as fab and az. It calls the Fabric REST API for notebook management and the Livy API for interactive execution.

Authentication

No service principal required. nb uses your Azure CLI session:

az login        # One-time; opens browser
nb auth status  # Verify

Notebook Formats

nb create generates notebooks with the correct Fabric metadata for either kernel:

Interactive Execution

nb exec creates a Livy session, submits code, returns output, and cleans up the session:

$ nb exec "My Workspace/Analytics" --code "print(2+2)" --lakehouse MainLH
  Creating Livy session (kind=python)...
  Session created; waiting for idle state...
  Submitting code...
4
  Session cleaned up

Building from Source

git clone https://github.com/data-goblin/nb-cli
cd nb-cli
cargo build --release
# Binary at target/release/nb

License

GPL-3.0