microsoft / vscode-tools-for-ai

Azure Machine Learning for Visual Studio Code, previously called Visual Studio Code Tools for AI, is an extension to easily build, train, and deploy machine learning models to the cloud or the edge with Azure Machine Learning service.
Other
325 stars 95 forks source link

Launching VS Code from AML puts users in an unexpected directory/path #1462

Open rukiafarraj opened 2 years ago

rukiafarraj commented 2 years ago

Expected Behavior

Launching edit in VS Code from AzureML, I expect to find myself in the same directory/path that I'm in when in AzureML.

Actual Behavior

Currently, launching in VS Code puts you two directories above where you are in AzureML. This is confusing for users. It also doesn't appear to be consistent with launching Jupyter/Jupyterlab from AML.

Steps to Reproduce the Problem

  1. In an Azure Machine Learning workspace, with compute instance running

  2. Navigate to Notebooks

  3. Launch a terminal in AML, and note current working directory Snag_abfb58f

  4. In the global toolbars, click "Edit in VS Code"

  5. Once in VS Code, launch terminal and notice working directory (2 directories above where you expect to be) Snag_abff3e0

  6. For comparison, launch Jupyter and Jupyterlab, and notice working directory Snag_ac03faf Snag_ac08f55 Snag_ac0d4da

Specifications

tbombach commented 2 years ago

Thanks for filing Rukia! I've added this to our current iteration plan to investigate and determine a fix.