dotfiles

Feel at home and work fast no matter which host you're working on.

Conda

Remember you can set up Conda via something like the following Bash snippet available as a Gist when on Linux:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh &&
    bash miniconda.sh -b -u -p "${HAFH:-${HOME}}/miniconda3" &&
    "${HAFH:-${HOME}}"/miniconda3/bin/conda init bash &&
    export PATH="${HAFH:-${HOME}}/miniconda3/bin:${PATH}" &&
    rm -rf miniconda.sh

Remember that you want to also create a main environment to keep the base Conda environment clean:

conda create --name main --clone base --copy

SSH Keys

When spinning up a new machine, you need to set up SSH keys to authenticate with GitHub

1. Create a new SSH public-private key pair:

ssh-keygen -t ed25519 -C "anilkeshwani@hotmail.com" &&
    cat "${HOME}/.ssh/id_ed25519.pub"

2. Add the public key as a new SSH key on GitHub

Scripts

Common scripts centralised here for use agnostic of machine.

SageMaker Connect (sagemaker_connect.sh)

Note

Source: Set up local Visual Studio Code, specifically Method 3: Connect from the terminal via SSH CLI.

Make the script executable:

chmod +x /path/to/sagemaker_connect.sh

Configure $HOME/.ssh/config to add the following entry:

Host space-name
  HostName 'arn:PARTITION:sagemaker:us-east-1:111122223333:space/domain-id/space-name'
  ProxyCommand '/path/to/sagemaker_connect.sh' '%h' AWS_PROFILE_NAME
  ForwardAgent yes
  AddKeysToAgent yes
  StrictHostKeyChecking accept-new

Breakdown of the HostName:

• 'arn:PARTITION:sagemaker:us-east-1:111122223333:space/domain-id/space-name'

For example, the PARTITION can be aws.

If you need to use a named AWS credential profile, change the proxy command as follows:

ProxyCommand '/path/to/sagemaker_connect.sh' '%h' YOUR_CREDENTIAL_PROFILE_NAME

Connect via SSH or run SCP command:

ssh space-name
scp file_abc space-name:/tmp/