Starting a new Jupyter project should be quick and easy. Instead of starting your project first and then realize the terrors of version controlling and sharing your notebook results you can use this repository as a template for quickly getting started.
No one likes to clutter their working environment. That's why this repository installs all the requirements in self contained containers and virtual environments. This way you can quickly clone an interesting project, toy around with the code and quickly nuke the project and all of it's requirements out of existence when you're done. 💥
- Requirements
- Setup
- Working with this setup
- Using this repo as a template for your own project
- Cleaning up after yourself
For all of this to work you need to have three things installed and set up on your computer:
makedockerhttps://docs.docker.com/install/virtualenvhttps://virtualenv.pypa.io/en/latest/installation/
After you've installed and tested the above requirements you're ready to setup the project. This is how you do that:
- Clone the repository.
git clone git@github.com:svenanderzen/data-science-template-repo.git
- (Optional) Setup a python virtual environment, install
nbstripoutand configuregitto use it in the local repository. We provide a script in order to ease this installation. If you're just checking out the project and know that you're not going to provide any pull requests you can skip this step.
./scripts/setup.sh
- Start a containerized JupyterLab instance and get cracking on the project. :hammer_and_wrench: We provide a Makefile target in order to provide the correct flags so that you can save those precious keypresses to something more useful.
make
- Follow the instructions in the terminal to open the JupyterLab instance.
Don't worry about someone else on the network accessing your notebook
without your permission. We bind the Docker container to only listen on
127.0.0.1i.e.localhost. This way your work is safe from prying eyes. :eyes:
Now, since we installed nbstripout inside a virtual environment we need to
activate this environment before we go ahead and use the git commands for
committing and diffing .ipynb files. Just do:
source venv/bin/activate
before you start to working with git and you're good to go. 🎉
The aim of this repository is to serve as a template for quickly getting up and running for your own data science projects. But how do you go about and do that?
Simple, just follow the below steps and you should be good to go 👍
- Clone this repository.
git clone git@github.com:svenanderzen/data-science-template-repo.git
- Rename your cloned repository
mv data-science-template-repo <your_awesome_name_here>
- Change the
gitremoteoriginto point to your own project
git remote rename origin svenanderzen/data-science-template-repo
git remote add origin <your_origin_url_here>
You don't need to save the old origin as
svenanderzen/data-science-template-repo unless you want to. However, it could
prove useful to keep it around in case you want to pull down template updates
and fixes in the future. This way you can just update your projects repository
with a simple merge commit:
git pull svenanderzen/data-science-template-repo master
If you no longer want to have the project or any of it's dependencies on your computer simply delete the repository folder and everything is gone!
This works because the Docker container is started with the --rm flag. This
way the container is completely removed from your system as soon as you shut
down the container. But what about the git configurations and virtual
environment? Well, the configurations are stored in the local .git folder and
the dependencies nbstripout and nbconvert are stored in the venv folder,
so when you remove the repository folder, the git configurations and python
dependencies are removed as well. Simple as that 🤷♂️