Re-make of central functions

[Kostis-S-Z]

Major updates

• Re-implement SDK functions following functional programming, instead of OOP to remove necessity of setting up a client. The idea behind this feature is that we make it as easy as possible for devs to integrate the SDK to their codebase.

Before

from datacollective import DataCollective
client = DataCollective()
dataset = client.load_dataset("<dataset-id>")

After

from datacollective import load_dataset
dataset = load_dataset("your-dataset-id")

• Create registry to handle dataset-specific loading functions. See README

• Single source of truth for environment variables under api_utils.py

• Remove setting multiple environments

• Remove duplicate code by implementing a generic api_request() function to handle GET/POST requests

• Remove scripts dir + code @moz-johanndiedrick was there a use for it in this repo?

Minor updates

• Add mkdocs + GH page + CI for docs
• Add CI for tests
• Add logo
• Add PyCharm files (.idea dir) to .gitignore
• Prettify and simplify README. Docs are moved to GH page.

[moz-tello-wharton]

The changes to .github/workflows/publish.yml... would you be able to share a little about the intended flow? It looks like it is getting quite complex.

Specifically, I am curious about how the flow is expected to work when it is main, and when it is release... and if perhaps those should be two separate files?

[moz-johanndiedrick]

LGTM!