Support multiple models

[tfiedlerdev]

Description

For multiple of our bachelor theses we need to support multiple ml models for one usecase each (also useful for our project vision),
For this we want the "usecase" concept.
Each use case has its own

• directory (to store usecase specific recordings, tensorflow model, labels.json, people.json)

Must Include

• use cases root dir
• UI for switching, displaying and adding a use case(s) (simply adds a new directory inside the use cases root dir)
• load labels.json, people.json from currently selected use case folder
  • when creating a new use case, initialize labels and people.json
• store recordings in currently selected use case folder
• load model from currently selected use case folder for prediction and later on device training screen
  • specifiy input shape of data and sensor amount/tags ( Support on device training #173 )
  • ui-for-training #174
• on first app start, create default use case ie UseCase1 where recordings, ... will be stored in by default
• if case no model is available in use case dir, show warning dialog when trying to predict or train including instructions on where to store a tf lite model on the smartphone
• implement storage for tflite models on the smartphone
  • extract model from assets for default use case