Provider sources model

[darribas]

This issue describes the model for managing sources.

Currently, we envision at least two sources of providers:

• Automatically scraped from Leaflet providers: these is parsed with parse_leaflet_providers.py and stored under provider_sources/leaflet-providers-parsed.json
• Community contributed additional providers: these are sources provided independently from the official Leaflet list by the community (following guidelines detailed in the docs). They get contributed to provider_sources/community_providers.json, which is a file that should be stored in prettified JSON format. Community contributions come as PRs against that file.

In terms of what gets shipped with the package that can be installed, one idea is to write all the sources into a single, compressed JSON file that ships with the package and loads on import. That way, neither the scraped Leaflet, nor the community, pretty JSON files would be shipped, just the merged and compressed one.

To implement this, we will need:

• Tooling to take all JSON files in provider_sources and merge/compress into a single provider_sources.json
• Load providers on import only from provider_sources.json