Filter out title case matches from dictionary rule matches

[rhystmills]

What does this change?

This PR filters out Dictionary rule matches that match against Title Case phrases.

Currently, most false positive spellcheck errors are for proper nouns - usually non-english names and places.

This PR filters out such matches before they are sent to Composer (though the DictionaryMatcher LanguageTool instance still generates such matches).

It also adds some tests which try to cover as many edges cases as possible.

In action in Composer:

How to test

Run locally:

• Composer, with config pointing to local Typerighter Checker
• Typerighter Checker service
• Typerighter Rule Manager service
• Make sure you have dictionary rules published from the rule manager
• Use Typerighter in an article in your local checker.
• Title Case words and phrases should not be spellchecked

[rhystmills]

Ran tests locally, works as expected.

One question – we'd expect corrections at the start of a sentence to be startcased, e.g. ... end of sentence. Strat of next sentence would have correction Start. Would this PR account for sentence starts?

There's a bit of prior art for detecting and managing sentence starts in RegexMatcher if that's relevant.

This did cross my mind. It will not catch typos when they're the first word of a sentence or paragraph, currently.

We'd still have the problem of false positive names in these instances, so it's up to us to decide whether that risk is worth spellchecking the first word. What are your thoughts?

[jonathonherbert]

@rhystmills, do we still need this after #443?

[rhystmills]

@rhystmills, do we still need this after #443?

We do not - now closed.