Typing in Context? about rustrict

[callmeclover]

Alright then, I mean, the features wouldn't really be hard to implement myself, so I suppose I'll just swap out with a basic Censor.

[finnbear]

The API for censoring a specific rustrict::Type works like this:

let censored = rustrict::Censor::from_str("some input")
    .with_censor_threshold(typ)
    .with_censor_first_character_threshold(typ) // optional
    .censor();

As for the value of typ, consider the following options:

• rustrict::Type::SEVERE | rustrict::Type::OFFENSIVE includes
  • mildly offensive
  • moderately offensive
  • severely offensive
  • severely profane
  • severely sexual
  • severely mean
  • severely evasive
  • severely spammy
  • severely [any future category]
• rustrict::Type::SEVERE & rustrict::Type::OFFENSIVE includes
  • severely offensive

If you want something other than those two options, please describe it!

[callmeclover]

Does this still work with Context? Would it just be

...process()
    .with_censor_threshold(typ)
    .with_censor_first_character_threshold(typ)

or is there some other way to implement this?

[finnbear]

Ah, I see. That's not currently possible. Context selects a Type at runtime, depending on message history.

While the following advice isn't documented anywhere, I recommend using Context as a model of how to use rustrict, that you can copy and adapt. It only makes sense to use rustrict::Context in your application if you have very similar needs as me. I've received some good advice that it should be removed from the library, possibly leaving it in examples.