Large refactoring + cancel when either SRT or WebRTC fails/disconnects + ease of use

Hey Folks 👋 , I'm using Donut to learn (thank you 🙇 ) more about web rtc and srt. This PR is a huge one, and I'm not sure you feel the same about the project's direction, so feel free to close this pull request. _(I can keep experimenting on my fork 😃 )_ You can test this PR by following the instructions. ``` git clone https://github.com/leandromoreira/donut cd donut make run # Make sure to have docker-compose/docker installed # Open http://localhost:8080, click on [Connect] check if it works ``` If it's working, you might want to check the main changes: - Install the [fx library](https://github.com/uber-go/fx) providing dependency injection and build/factory patterns - Extract the core logic (web rtc, srt, streaming, h264, eia608) into controllers - Introduce a simple architecture: - `request --> [server] --> [router] --> [handler] --> [controller]` - Introduce a FAQ section TBH, the only real (beyond moving code around) contributions/changes are: - Make it easier to run/experiment locally with docker-compose - If either the SRT or WebRTC gets disconnected/failed, it closes all related resources - Add more documentation/faq/how_it_works/... My current goals are: - to bring the player to the repo (making it easier to experiment), - support audio, - and play with transcoding media to expand the supported browsers. Thanks cc @Sean-Der