Advanced word typo metrics and statistics

[achalagarwal]

Is your feature request related to a problem? Please describe. An advanced understanding of the letters that are being mistyped in the words and the reasons for the same.

The feature is an extension of the previous one related to combining the two streams https://github.com/achalagarwal/keystrokes/issues/8

Moving ahead in word context errors:

1. Improve cascading metric with the counter value
2. Add per letter flags, is it a cascading error, adjacent error, finger typo, not an error

What's next with this?

1. fix letters on the fly, by detecting the correct word alternatives, might need to include spacy in this but want to first try without spacy
2. the data collection for labels, enable this asap

Describe the solution you'd like A solution is to start sending counter from the typo streams and match letters in the word to the typos being sent, an absent typo means that the letter typed is correct and similarly we can set real typo flags, cascading etc

Send the counter in the typo stream, if there is no error corresponding to a letter, w e Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

An alternative solution would be to somehow ignore the correct letters because they are in the majority, saves a lot of processing.