Dockerhub

[gsganden]

Closes #78

I pushed a docker image of the prediction serving app to Dockerhub. This PR adds instructions for using it to the repo README. Putting the app image on Dockerhub will make it easier for the Lincoln Park Zoo or anyone else simply to download and run the app without having to build it themselves.

I also simplified the app by removing nginx. My understanding is that nginx is helpful for handling web traffic, but it is unnecessary for local deployments, and it would complicate using Dockerhub because it involves running multiple containers. We can always put nginx back in later if we want to do a web deployment, but at this point I think local deployments work better because sending images over the internet would be a bottleneck.

Pull Request Checklist

• [x] Pull request includes a description of the change and the reason behind it.
• [x] Pull request uses keywords to close relevant issues.
• [ ] Pull request includes unit tests for any new functionality.
• [x] README and docs have been updated.
• [x] ./.ci/test.sh passes locally.

Maintainer's responsibilities:

• [x] _version.py has been updated.
• [x] CHANGELOG.md has been updated.

[gsganden]

@jameslamb I just updated autofocus/predict/example_post.py and autofocus/predict/example_post.R. Could you please take a look? Also, would you be able to make sure that autofocus/predict/example_post.R runs? I don't have an R environment set up.