Unable to detect references with non-Latin conventions.

[JoshTheDerf]

Great work on fetch(bible)!
We're looking at using it to detect and link references for some of our non-English articles.

Detection works great for references using standard latin formatting conventions, but not as well for CJK languages. I suspect different languages may need different sets of regexes in https://github.com/gracious-tech/fetch/blob/master/references/src/detect.ts#L15-L25

For example, Genesis 1:1-2 is commonly formatted differently for the following languages:

• Genesis 1:1-2 - English
• 创世记1章1–2节 - Chinese (Simplified)
• 創世記1：1ー2 - Japanese
• 창세기 1장 1–2절 - Korean

Could we brainstorm a way to add regular expression override configuration to either the enhance script or the language/translation definitions themselves?

[shadow-light]

Hi Josh, yes happy to try improve support for other languages. I've only tested it out on some basic stuff so far.

There's several layers to this:

1. reference detector (docs): which already has useful settings for other languages, like matching book name from a single char, and matching anywhere in a word and not just start. You can see a test for Chinese here.
2. client: which auto supplies book names depending on the user's language which are essential for reference detection. The client (and therefore also enhancer) will auto-set Chinese relevant props here.
3. enhancer: which walks the page for you and transforms references it finds, but is less configurable.

Let me know if any of that helps with the above examples you gave, and where the logic goes wrong. Sorry, I don't have time right now myself. But if you can work out what logic needs changing I'll be happy to assign time to addressing it when I can.

Thanks for the feedback