D. Python Environment Setup

• On GitHub, new a project and choose Python.gitignore

• Git clone the new project with a token, which gives you permission to push

• In vs code, go to File > Open Folder to open the project

Manual

• python -m venv .venv in the project root to create a virtual environment called .venv

  • A virtual environment is a folder for an independent python.exe and its own site-packages which this python.exe can import from

  • venv is part of the python standard library

  • -m is to run a module, venv in this case

• .venv\Scripts\activate to activate the virtual environment to start use the independent python.exe

• Create a requirements.txt file in the project root to keep a list of all packages (and possibly versions) you want to use

• pip install -r requirements.txt to pip install packages from the list

  • pip downloads packages from pypi.org

uv

• uv init in the project root to initialize the project

  • It creates a pyproject.toml file amoung others

  • pyproject.toml is the recommended replacement of requirements.txt. Other package managers (e.g. poetry) will also give you a pyproject.toml file

• uv sync to create and sync the venv as defined in pyproject.toml

  • You still need to manually run .venv\Scripts\activate activate the virtual environment

  • Or you can choose the project venv in vs code at the bottom right corner. Once done, vs code will activate the venv automatically every time you start a new command prompt

• uv add numpy to add numpy to pyproject.toml and to the venv

• uv remove numpy to remove numpy from pyproject.toml and from the venv

• Try not to edit pyproject.tmol manually; Do it through uv

• Install python version

  • uv venv --python 3.12.4

  • uv python install 3.12.4

Git

• Download Git and install

• git init in the project root to version control the project by Git

• Create a .gitignore file in the project root

  • Copy the contents of Python.gitignore and paste in yours

  • The same repo has recomended .gitignore in other languages/type of projects

• git add * to add current code changes to the staging area

  • The staging area is an intermediate space to keep the changes before you commit

• git commit -m "YOUR_COMMIT_MESSAGE"

  • Commit message is required

• Configure Git username and email

  • git config --global user.name "YOUR_NAME"

  • git config --global user.email "your_email@example.com"

  • Please use the same username and email as your GitHub account

VS Code

• Install the Python extension

• Open a .py file and look at the bottom right, VS Code should be aware of the venv. Kill the terminal and open a new one and the venv will be activated automatically

• Keyboard shortcut

  • Ctrl + Shift + G: Focus version control tab on the left

  • Ctrl + Shift + E: Focus file explorer on the left

  • Ctrl + 1: Focus editor

    • When you are in editor, Ctrl + Page Up/Dn to switch between files

  • Ctrl + Shift + ~: Focus terminal

  • Ctrl + Shift + Y: Focus debug console

  • Ctrl + Shift + I: Focus github copilot chat

Side Project Ideas

GARCH

• Fit a GARCH(1, 1) model to the SPX daily return using the arch package

  • There is an open source library yfinance that scrapes yahoo finance, but it breaks from time to time

  • One reliable data source is Alpha Vantage

    • Free API key has a limit of 25 requests per day

• Follow the example on this Plotly Dash documentation page to build and run your first app locally

• Dockerize your minimal Dash app

• Deploy your docker image on Render free teir

Numerical Math

• If we know the point values of a \(C^{\infty}\) function \(f\) at \(a\) and \(a+h\), we can use a linear function

  \[y = f(a) + (x-a)\frac{f(a+h) - f(a)}{h}\]

  to approximate \(f\) in the interval \([a, a+h]\). This is linear interpolation. This approximation has an error bond \(Ch^2\). Can you find me the error bond and its proof?