Common Functions of JupyterLab

Introduction

To access JupyterLab from a running notebook instance, perform the following operations:

1. Log in to the ModelArts management console. In the navigation pane on the left, choose Development Workspace > Notebook.
2. Click Open in the Operation column of a running notebook instance to access JupyterLab.
3. The Launcher page is automatically displayed. Perform required operations. For details, see JupyterLab Documentation.
   Figure 1 JupyterLab homepage
   
   

   The notebook and console kernels and versions displayed on the Launcher page vary depending on the AI engine based on which a notebook instance is created. Figure 1 shows an example only. Obtain the notebook and console kernels and versions on the management console.

   • Notebook: Select a kernel for running notebook, for example, TensorFlow or Python.
   • Console: Call the terminal for command control.
   • Other: Edit other files.

Creating a Terminal in JupyterLab

You can run Python commands on the terminal to operate the terminal. The following describes how to open the terminal of JupyterLab.

1. Create a notebook instance. When the instance is running, click Open in the Operation column. The JupyterLab page is displayed.
2. Choose File > New > Terminal. The Terminal page is displayed.
   Figure 2 Terminal
   
3. You can use pip to install external libraries in the TensorFlow-1.8 environment on the Terminal page. For example, to install Shapely:

   Enter the following commands in the code input box to obtain the kernel of the current environment and activate the Python environment on which the installation depends:

   cat /home/ma-user/README

   source /home/ma-user/anaconda3/bin/activate TensorFlow-1.8

   

   To install TensorFlow in another Python environment, replace TensorFlow-1.8 in the command with the target engine.

   Figure 3 Activating the environment
   

   Run the following command in the code input box to install Shapely:

   pip install Shapely

Creating an IPYNB File in JupyterLab

On the JupyterLab homepage, click a proper AI engine in the Notebook area to create an IPYNB file.

The AI engines supported by each notebook instance vary depending on the runtime environment. The following figure is only an example. Select an AI engine based on site requirements.

Figure 4 Selecting an AI engine and creating IPYNB file

The created IPYNB file is displayed in the navigation pane on the left.

Figure 5 Created IPYNB file

Creating a Notebook File and Accessing the Console

A console is a Python terminal, which is similar to the native IDE of Python, displaying the output after a statement is entered.

On the JupyterLab homepage, click a proper AI engine in the Console area to create a notebook file.

The AI engines supported by each notebook instance vary depending on the runtime environment. The following figure is only an example. Select an AI engine based on site requirements.

Figure 6 Selecting an AI engine and creating a console

After the file is created, the console page is displayed.

Figure 7 Creating a notebook file (console)

Editing a File in JupyterLab

JupyterLab allows you to open multiple notebook instances or files (such as HTML, TXT, and Markdown files) in one window and displays them on different tab pages.

In JupyterLab, you can customize the display of multiple files. In the file display area on the right, you can drag a file to adjust its position. Multiple files can be concurrently displayed.

Figure 8 Customized display of multiple files

When writing code in a notebook instance, you can create multiple views of a file to synchronously edit the file and view execution results in real time.

To open multiple views, open an IPYNB file and choose File > New View for Notebook.

Figure 9 Multiple views of a file

Renewing or Automatically Stopping a Notebook Instance

If you enable auto stop when you created or started a notebook instance, the remaining duration for stopping the instance is displayed in the upper right corner of JupyterLab. You can click the time for renewal.

Figure 11 Remaining duration

Figure 12 Renewing an instance

Common JupyterLab Buttons and Plug-ins

Figure 13 Common JupyterLab buttons and plug-ins

Figure 14 Buttons in the navigation bar

Figure 15 Buttons in the menu bar of an IPYNB file

Using Code-based Plug-ins

The code parametrization plug-in simplifies notebook cases. You can quickly adjust parameters and train models based on notebook cases without complex code. This plug-in can be used to customize notebook cases for competitions and learning.

• The Add Form and Edit Form buttons are available only to the shortcut menu of code cells, as shown below.
  Figure 16 Viewing a code cell
  

• After opening new code, add a form before editing it.
  Figure 17 Shortcut menu of code cells
  
• If you click Add Form, a code cell will be split into the code and form edit area. Click Edit on the right of the form to change the default title.
  Figure 18 Two edit areas
  
• If you click Edit Form, four sub-options will be displayed: Add new form field, Hide code, Hide form, and Show All.

  Table 5 Edit Form sub-options

  Edit-Form Sub-option	Description
  Add new form field	The form field types include dropdown, input, and slider, as shown in Figure 19. Each time a field is added, the corresponding variable is added to the code and form areas. If a value in the form area is changed, the corresponding variable in the code area is also changed. NOTE: When creating a dropdown form, click ADD Item and add at least two items, as shown in Figure 20. If the form field type is set to dropdown, the supported variable types are raw and string. If the form field type is set to input, the supported variable types are boolean, date, integer, number, raw, and string. If the form field type is set to slider, the minimum value, maximum value, and step can be set.
  Hide code	Hide the code.
  Hide form	Hide the forms.
  Show all	Display both code and forms.

  Figure 19 Form field types
  

  Figure 20 Creating a dropdown form
  

  Figure 21 Deleting a form
  

Monitoring Resources

To obtain resource usage, select Resource Monitor in the right pane. The CPU usage and memory usage can be viewed.

Figure 22 Resource usage