Test Jupyter Kernel Execution API

[DonJayamanne]

Refs: #14669

• anyOS @rebornix
• anyOS @amunger
• anyOS @Yoyokrazy

Complexity: 5

Create Issue

---

Pre-Requisites

• Ability to run a Jupyter Notebook against a local Python kernel
• Ability to run a Jupyter Notebook against a remote Python kernel

Steps

• Create a sample VS Code extension using yo code or use the sample Sample extension using Jupyter Execution API vscode-extension-samples#923
• Run npm install @vscode/jupyter-extension@proposed to install the typedef of the Jupyter Extension API

Review the Kernel API

• Review the API Jupyter.kernels, Jupyter.kernels.getKernel, <Kernel>.executeCode

Verify Access to Kernel API

• First open a notebook, run code against a Python kernel (local/remote)
• Use the API method Jupyter.kernels.getKernel(< URI of notebook > to get the kernel
• Verify the user is prompted to grant access to this kernel
• Click cancel.
• Verify a kernel is not returned in the API call
• Select the command Manage Access to Jupyter Kernels
  • Verify your extension is displayed in the list
  • Verify your extension is NOT ticked
• Tick your extension from the list
• Go back to your extension code and invoke the getKernel API again
  • Verify a prompt is not displayed
  • Verify a Kernel instance is returned via the API
• Select the command Manage Access to Jupyter Kernels
  • Verify your extension is displayed in the list
  • Verify your extension is ticked
• Un-Tick your extension from the list
• Go back to your extension code and invoke the getKernel API again
  • Verify a prompt is not displayed
  • Verify a Kernel instance is NOT returned via the API
• Reload VS Code
  • Run code against the open notebook
  • Verify calling getKernel from the extension will display the prompt
  • Click OK
  • Verify the Kernel instance is returned
• Reload VS Code
  • Run code against the open notebook
  • Verify calling getKernel from the extension will NOT display the prompt
  • Verify the Kernel instance is returned

Verify Execution of Code to Kernel API

• First open a notebook, run code against a Python kernel (local/remote)
• Access the Kernel via the getKernel API call
• Verify you can execute code such as print("Hello World") via the API and you get the output back
• Verify you can execute code such as for i in range(10):\n import time\n time.sleep(1)\n print(i) via the API and you get 10 outputs back one by one (i.e. verify outputs are streamed as they arrive)
• Verify code is executed against the right Kernel when you have two notebooks with different kernels.

[DonJayamanne]

@amunger @rebornix @Yoyokrazy
If you run into this error, then please edit the api.d.ts file and comment those two lines