Tutorial - Local deployment

In this tutorial, you set up a basic development environment and deploy a lakehouse and notebook to a Fabric workspace from your local dev workstation. You can choose between two approaches:

• Python SDK — Use the fabric-cicd Python library to publish items programmatically.
• Fabric CLI — Use the fab deploy command with a YAML configuration file. No Python code required.

The Fabric CLI fab deploy command uses the fabric-cicd library under the hood. Choose the approach that fits your workflow.

Prerequisites

Python SDK

• A test workspace in Fabric
• Admin permissions on the Fabric workspace.
• A GitHub account (required to access demo files)
• VS Code or similar editor: Download VS Code
• Python Install python
• Azure CLI (used only for authentication): Install Azure CLI

Fabric CLI

• A test workspace in Fabric
• Admin permissions on the Fabric workspace.
• A GitHub account (required to access demo files)
• Python Install python

Step 1. Download the source files

1. Fork the fabric-samples repository to your GitHub account.
2. Clone your fork and navigate to the sample folder:

git clone https://github.com/<your-account>/fabric-samples.git
cd fabric-samples/docs-samples/cicd/fabric-cicd-local

Step 2. Install the deployment tool

Python SDK

Using the terminal from within VS Code, install the fabric-cicd python library.

pip install fabric-cicd

Fabric CLI

Using the terminal, install the Fabric CLI.

pip install ms-fabric-cli

Tip

If you already have the Fabric CLI installed, upgrade to the latest version to ensure the deploy command is available:

pip install --upgrade ms-fabric-cli

Verify the installation:

fab --version

Step 3. Configure the deployment

Python SDK

Navigate to the fabric-cicd-local files that you cloned down in step 1. They're located in the samples folder. Edit the deploy.py script, replacing <YOUR_WORKSPACE_ID> with your id. Save the changes.

from pathlib import Path
from fabric_cicd import FabricWorkspace, publish_all_items # 👈 import the function

repo_dir = Path(__file__).resolve().parent # ...\fabric_items

workspace = FabricWorkspace(
 workspace_id="<YOUR_WORKSPACE_ID>",
 repository_directory=str(repo_dir),
 # environment="DEV", # optional, but required if you use parameter replacement via parameter.yml
 # item_type_in_scope=["Notebook", "DataPipeline", "Environment"], # optional scope
)

publish_all_items(workspace) # 👈 call the function

Fabric CLI

Navigate to the fabric-cicd-local/fabric_items folder that you cloned down in step 1. Create a config.yml file in the root of the folder with the following content, replacing <YOUR_WORKSPACE_ID> with your workspace ID.

core:
  workspace_id: "<YOUR_WORKSPACE_ID>"
  repository_directory: "."

For more configuration options, including environment-specific settings, item type scoping, and parameterization, see the fab deploy documentation.

Step 4. Sign in

Python SDK

Using the terminal, sign in using az login.

az login

Note

If you're using a trial or have no Azure subscriptions associated with your account you can use the --allow-no-subscriptions switch.

az login --allow-no-subscriptions

Fabric CLI

Using the terminal, sign in using the Fabric CLI.

fab auth login

Follow the prompts to authenticate with your Microsoft Entra ID credentials.

Step 5. Run the deployment

Python SDK

Run the deploy.py script. From within VS Code, go to Run -> Start Debugging. You should see the following output in the screenshot.

Fabric CLI

Run the fab deploy command from the terminal, pointing to your configuration file.

fab deploy --config config.yml

When prompted to confirm the deployment, type y to proceed.

The CLI publishes all items found in the repository directory to the target workspace and unpublishes any items that no longer exist in the source.

Step 6. Verify the items were created

Once the deployment completes, check your Fabric workspace. You should see the new lakehouse and notebook. Congrats you're done!