Add autocorrect threshold to model-metadata branch

See merge request alex/latinime!6

* If the first letter is capital, only capitalized first tokens will be sampled. If the whole text is capitalized, then only fully capital tokens will be sampled for the whole word
* If a word is an exact match, it gets boosted relative to others
* Probability threshold for autocorrect is now 18.0
* Add "clueless" threshold, if it's less than 1.3 then just show the user's typed word in the middle instead.

Fix finetuning, add a finetuning screen, handle errors during importing model, update metadata format, add model exporting

Fix embedded tokenizer loading, implement new model management methods, implement model info loading, model importing

Popup would leave leftover dead popups forever after screen orientation change

* Use RecyclerView for significantly better performance
* Skin tone selector popup

* Shift + swiping space now lets you select text
* Text editor uses more consistent selection logic instead of sending dpad and relying on apps to implement shift+dpad selection, which many of them don't. Dpad is still used for up/down