Easier Genie Code Install Options

[Zachary-Higgins-Altera]

Hi - I found the Genie Code install for Genie Code to be unnecessarily complicated and not compatible with Genie Code. It doesn't seem like you can run it from a notebook in the workspace. install_skills.sh is a pre-req for install_skills_to_genie_code.sh, but install_skills.sh doesn't run unless we're in a project directory which is defined in the following code block:

# Check if we're in a git repo or project directory
if [ ! -d ".git" ] && [ ! -f "pyproject.toml" ] && [ ! -f "package.json" ] && [ ! -f "databricks.yml" ]; then
    echo -e "${YELLOW}Warning: This doesn't look like a project root directory.${NC}"
    echo -e "Current directory: $(pwd)"
    read -p "Continue anyway? (y/N): " confirm
    if [ "$confirm" != "y" ] && [ "$confirm" != "Y" ]; then
        echo "Aborted."
        exit 1
    fi
fi

This conflicts with the official documentation that states skills in Genie Code live in the workspace level user home directory (not project): https://learn.microsoft.com/en-us/azure/databricks/genie-code/skills

Skills live in the /Users/{username}/.assistant/skills/ directory. Each skill must have its own folder and a SKILL.md file within that folder.

... so there probably won't be a pyproject.toml, package.json, or databricks.yml in the user's home folder. I don't know, maybe the assumption is that users would try to run this from a project, and then the installer would do its job and eventually we would migrate to the workspace level skills dir? But what if the workspace is fresh? There are too many assumptions with this pattern in my opinion.

Please consider providing something like this to give users an easy way to install these to their workspace (just run a notebook).
https://gist.github.com/Zachary-Higgins-Altera/943a0e162bd6ffe6ab56e32d07652a46

[scottDBX1886]

The problem with this approach is the databricks-skills does not have all the skills bundled in the ai-dev-kit. inside the install_skills.sh, we are installing skills from other repos as well (app apx and ml). So just coping the skills, you will miss some.

I downloaded and tested the nb and only 26 skills of the 35 available were uploaded. This is where calling the install_skills.sh is key to having everything.

I don't disagree that having a notebook would be helpful. I will look at making the genie_install_script extendable to the workspace, where you could just clone the repo in a workspace folder and use the web terminal to execute. Then no matter what env you are in, the same install script works.

[Zachary-Higgins-Altera]

The problem with this approach is the databricks-skills does not have all the skills bundled in the ai-dev-kit. inside the install_skills.sh, we are installing skills from other repos as well (app apx and ml). So just coping the skills, you will miss some.

I downloaded and tested the nb and only 26 skills of the 35 available were uploaded. This is where calling the install_skills.sh is key to having everything.

I don't disagree that having a notebook would be helpful. I will look at making the genie_install_script extendable to the workspace, where you could just clone the repo in a workspace folder and use the web terminal to execute. Then no matter what env you are in, the same install script works.

Thanks Scott - That works for me. The notebook was more or less a conversation starter, so I came with something besides a wish list :)

I think this would be very beneficial for anyone using Genie Code.