Discussion about tags

[bunny-therapist]

Starting this discussion to move it from 3f9be61#r152275098 because discussing in a commit is troublesome.

The point is, @xamgore - I believe the priority order of the tags matter and reverting back will change the output. (Though the further we get from the named tags in the original yake, the harder it is for me to actually figure out what is going on, so don't expect me to get all the details right.)

[bunny-therapist]

Also, I feel like this new "tags.rs" module is problematic. I did not create such a module because I knew that some tag functions needed word, other occurrence, some need config, etc, so there is no way to have a clean interface. Now our config is effectively also config for tags.rs, but we have to pass it to some functions and not other (like is_capital needs the strict_capital config for example).

[xamgore]

Yeah, that's true. But extracting them is still a positive change, as it's much easier now to understand, what is what. I forget all those tags constantly. n lost me.

[bunny-therapist]

I struggle to come up with a test to show the impact of tag priority order because it is only the maximum of tf_a and tf_n that matters. The different ways of determining them (get_tags vs separate functions) change the values of tf_a and tf_n, but not the maximum of both. Though I cannot conclusively prove that this is always true. I would like to see such a proof before we start reverting stuff as in your todo.

[xamgore]

Yeah, I agree. Tests turn red if the other tags (n, d) are ignored.