Add Dockerfiles for web API & headless processing

[mihow]

We likely need 2 containers:

• One to run ami api on every server that needs to run an instance of the ML processing backend
• One to run the headless CLI processing tools, especially for processing a queue from a shared database with multiple compute nodes
• Optionally one to run the Gradio demo? (ami gradio)

Each has the same dependencies, so they all use same container with different commands. Can the ami CLI tool be the entrypoint?

[f-PLT]

@mihow is this still relevant and if so does it need to be on the radar for October?