Read time based on language

Problem/Motivation

Currently the read time is calculated and saved only for the default language. It would be better if it saves the values for each translation.

Steps to reproduce

Create a translation for the wanted node with different amount of words. You'll see that the read time doesn't change.

Proposed resolution

Add language as another value to distinguish between the entires in the database (table read_time). It has to work on node insert and also on node update.
And also when the node is deleted -> remove the correct entry in the database.