<aside> ⭐ Home | Getting Started with CCR | Learning Resources | Software | Productivity | Hardware | Best Practices

</aside>

Coding

Package Management using Anaconda

In general, to manage Python packages efficiently, one popular tool is Anaconda. Anaconda is a powerful package and environment manager that allows users to create isolated Python environments and easily install, update, and remove packages. With Anaconda, you can create separate environments for different projects, ensuring that package dependencies do not conflict with each other. Whether you are a beginner or an experienced Python user, Anaconda simplifies the process of managing Python packages and ensures a smooth and hassle-free experience.

<aside> 💡 Python Packages Management on CCR

Anaconda will take up a lot of space. This is detrimental to workflow in CCR because CCR only allocates 10 GB for the HOME directory for each individual user. That means your HOME directory will be quickly filled up if you install/configure Anaconda using default settings. Since I used Anaconda for a long time, I’d like to stick to it. One workaround I found is that you can install/configure Anaconda or Miniconda in our Project Space which has 1 TB. Search Google to find out how to install and configure Anaconda in user-specified location. Let me know if you find out a better way to manage your packages in CCR. See here for CCR’s general suggestions.

</aside>

VS Code & GitHub Copilot

Use VS Code on Your Local Machine

I simply recommend every student in the group to write and debug their code in VS Code. There you can use advanced GitHub Copilot to help you write the code.

<aside> 💡 Do you know that teachers and students are eligible for FREE access to GitHub Copilot and more? Check out this post to get your free access today!

</aside>

Use VS Code on CCR

• Access through Remote Tunnels

  <aside> ⚠️ Updated on Feb 2, 2024: Currently, Remote Tunnel function only works for VS Code version **1.85.x** and has some problems with the latest 1.86 version. So please download VS Code version 1.85.x HERE.

  </aside>

  This is the recommended way to connect your local VS Code software to remote CCR. You can check detailed instructions on Microsoft website. Basically:

  1. In your CCR HOME directory, download code CLI:

     # download the latest version of code CLI (currently does not work)
     #curl -Lk '<https://code.visualstudio.com/sha/download?build=stable&os=cli-alpine-x64>' --output vscode_cli.tar.gz
     
     # download code CLI version 1.85.1
     curl -Lk '<https://update.code.visualstudio.com/1.85.1/cli-alpine-x64/stable>' --output vscode_cli.tar.gz
     
     # unzip the downloaded file
     tar -xf vscode_cli.tar.gz
     

  2. Create a secure tunnel with the tunnel command:

     ./code tunnel
     

  3. Follow instructions on the screen to login to your GitHub account and activate the tunnel.

     <aside> 🚨 Note: please DO NOT close the CCR window when you are using the remote tunnel function. If you close the CCR window, the tunnel will close.

     </aside>

  4. If you're already working in VS Code on your local laptop/desktop and would like to connect to this remote CCR tunnel, you can install and use the Remote - Tunnels extension directly. Once you install the extension, open the Command Palette (F1) and run the command Remote Tunnels: Connect to Tunnel. You'll be able to connect to any remote machines with an active tunnel.

  5. Using Remote Tunnel function is the recommended way, because in this case, you have access to all extension functions of VS Code on your local machine, such as GitHub Copilot, GitHub Copilot Chat (requires installation from the extension marketplace).

• Access through Open OnDemand (OOD)

  OOD is a browser based single point of access for all of CCR's clusters, shared storage, and remote visualization servers. See CCR website on instructions to access OnDemand.

  If you wish to use GitHub Copilot, you can not install GitHub Copilot in VS Code using the built-in marketplace on OOD at this time. Instead, you can do so by installing the extension using a VSIX file. Check this Berkeley website for instructions*.*** These instructions will likely also apply to installing other extensions for Visual Studio that are not available in the extension marketplace.

  <aside> 📢

  Update

  As of 10/7/2024, VS Code version on OOD has now been updated to v1.80.2, we need to install more updated version for GitHub Copilot. I have tested v1.88.132 for Copilot and it works fine with the current OOD VS Code version.

  VSIX file for this compatible GitHub Copilot (v1.88.132) can be downloaded from this website (go to “Version History” tab on the website) or directly download using the link below:

  GitHub.copilot-1.88.132.vsix

  The instructions for installing VSIX file can be found in the above Berkeley website.

  </aside>

Cursor