Basic information about running this repo

Python

• We use Python as programming languange in this course - https://www.python.org/
• Python is/was popular as a coding language for system administrators
• The language has been adopted by the Data Science/Machine learning community
• The code should be fairly easy to understand
• Packages in Python are installed using pip , pip -q indicates pip to be less verbose

Jupyter Notebooks

• Lessons are in Jupyter notebook form. https://jupyter.org/

• Notebooks are a mixture of markdown and codeblocks and allow for easy experimentation

• When running a Notebook you need to select a "kernel" , this is a virtual python environment

• You can see the button "select kernel" on the right upperhand

• While you can run each lesson in the same kernel, we select different kernels to avoid dependency conflicts

• There should be a kernel for each lesson

• You can reset a kernel to make it start from scratch again

• You can run a notebook step by step, or run all

VScode

• We use VSCode as our code editor
• It has extensions/plugins for python installed
• Under the Go menu there is a command pallete
• Usefull commands:
  • Dev containers : rebuild container
  • Dev containers : open workspace in container
  • Developer : reload windows

Devcontainer and Docker

• We make use of a devcontainer, a way of running our code in a docker container
• As the base container we use a python distribution from Microsoft
• If you are running this repo locally, you'll need to have Docker installed
• The use of a devcontainer makes it easy for us to run it in Github Codespaces
• To achieve compatibility , we run the container in amd64 mode
• For this work on MacOS we need to install the rosetta
  • softwareupdate --install-rosetta --agree-to-license
  • https://www.kenmuse.com/blog/forcing-dev-container-to-use-amd64/
• We have the docker in docker extension installed, so we can run docker commands inside of our container
• Our github credentials are shared inside of the devcontainer so we can edit/commit directly from there
• On Mac , Docker has a default memory limit of 2GB and a small Disk, increase it to 16GB of memory and 60GB of disk if you can

OpenAI and Anthropic

• We'll use OpenAI and for using that , you need an API-key - https://platform.openai.com/api-keys
• We'll optionally use Anthropic , you need an API-key for that too - https://console.anthropic.com/settings/keys

Credentials

• Credentials can be configured in the .env file . Copy over the .env_sample and fill in the right credentials

• Use OPENAI_API_KEY and ANTHROPIC_API_KEY as keys

• In vscode: By default, the Python extension looks for and loads a file named .env in the current workspace folder, then applies those definitions. The file is identified by the default entry "python.envFile": "${workspaceFolder}/.env" in your user settings (see General Python settings). You can change the python.envFile setting at any time to use a different definitions file. https://code.visualstudio.com/docs/python/environments

• To configure the credentials in codespaces , go to the settings of this repo and configure the secrets there

  • Repo -> Settings -> Security -> Secrets and Variables -> Codespaces ->
  • https://docs.github.com/en/codespaces/managing-codespaces-for-your-organization/managing-development-environment-secrets-for-your-repository-or-organization#adding-secrets-for-a-repository

Github and Github Codespaces

• To run Github codespaces, you need a github account

• You should get 60 hours of running a container for free each month

• Make sure to register this ahead of the course as the first day it has some limits applied

• Make sure to stop/delete your codespace if you're not using it

  • Page to manage your codespaces https://github.com/codespaces

• To run the codespaces :

  •