Arabic Shaping Support?

[Ellpeck]

Hi, thanks for this very cool library! Definitely the most succinct bidi library that I've found that isn't too much of a hassle to set up (because it doesn't require native bindings).

Are there any plans to support arabic shaping as part of this library (even though I realize it's not necessarily directly related to bidi support)? ICU implements this in Java as ArabicShaping (the docs here also contain an explanation for what it does!) and in the original C library as ushape. The data seems to be available as ArabicShaping.txt.

I also found an existing C# port of the Java version of ArabicShaping in an archived project. Maybe it can be a good starting point, since it's MIT licensed: https://github.com/sherif-elmetainy/DotnetGD/blob/master/src/CodeArt.Bidi/ArabicShaping.cs

Thanks for this lib either way! ❤

[roddyrap]

Hi! I am currently busy but I think I could take a look at it around next weekend.

I am in favour of supporting more Unicode related data natively in C# (As long as I can do it right).

I will take a look at it when possible and return a time estimate if I think I can manage it.

Thanks for the suggestion! I will also take a look at the linked Java implementation.

[roddyrap]

I found some time to look at the Arabic shaping algorithm & data.

Getting the Unicode information is easy enough, and I have already constructed a basic implementation, though the final data structures will probably be altered before a final version.

I think the Arabic shaping algorithm is specified here (I always prefer to use official Unicode documents as references): https://www.unicode.org/versions/Unicode16.0.0/core-spec/chapter-9/#G26848 (Unicode Chapter 9). I need to check that I am not missing more rules.

It doesn't look hard to implement, so I hope to have a "working" prototype by the end of the week (Though I do not promise). I still need to figure out how to test it. Currently, I can manually check for correctness myself, but in the future, I hope to find a complete test dataset that I can use to automatically ensure correctness someday.

Do you have any specific texts you want me to use/check? Will you be able to test the implementation I will (Hopefully) provide?

[Ellpeck]

Thanks so much for the quick responses!! Your plan sounds awesome.

I can absolutely test the implementation you'll provide! At the moment, I use a FriBidi wrapper around C# that already works pretty well, but obviously isn't as cross-platform as this lib would be. If you need me to, I can send comparative screenshots of text in the project I use these for & whatnot :)