I have just get my hands dirty with VSCode's extension Google Colab (by Google).
While I can connect to google servers for running Jupyter Notebooks (.ipynb files), Is there a way to use VSCode directly on .py files of my project and execute them on Google servers.
Note that I dont want to use the command !python <filename.py> from my notebooks to run the .py files of my project, I want to run them directly with VSCode just like I am able to with Jupyter Notebook.
Yes, edit .py files in VSCode and run them on Google servers via the Google Colab extension, like Jupyter notebooks.
Setup Steps:
Install Jupyter and Google Colab extensions in VSCode.
Open .py or .ipynb file, select Colab kernel, and sign in with Google.
Connect to Colab runtime (CPU/GPU/TPU) for remote execution.
Limitations: No direct Google Drive mount or userdata access in VSCode.
Yes, based on what this GitHub repo says: https://github.com/googlecolab/colab-vscode
Copied from the repo's README.md:
Install VS Code.
Install the Colab extension from either the Visual Studio Marketplace or Open VSX.
Open or create a notebook file.
When prompted, sign in.
Click Select Kernel > Colab > Auto Connect.
😎 Enjoy!
!python <filename.py>from my notebooks to run the .py files of my project": while there are the rare occasions that can be useful, you primarily should be using%run <filename.py>to run Python script files from inside a running notebook. This magic is more full-featured and offers better handling of the stdout and stderr than sending it off to a temporary shell instance like!pythonwould. The best feature for development is%run -iwhich allows you to run the script attached to the current namespace. See...%runfor more information.