Can't detect name. | Bug reports

Can't detect name.

complete

Internal Quail

I wrote in Thai, we do not have space between all words in sentences. So it cannot detect any name in the codex. For example, the name is "นิกซ์" ("Nyx"), but when I wrote "นิกซ์กล่าว" ("Said Nyx") it cannot understand that this is a name. Big problem. Please help asap.

August 26, 2024

spaceemotion

marked this post as

complete

spaceemotion

marked this post as

in next release

Found a way to guesstimate the word counting for Thai (based on https://gotranscript.com/word-count) and updated how Asian languages are being handled so this becomes easier to update in the future.

The example of "นิกซ์กล่าว" is supported by the codex now, although the word count is "3" (though it should be 2, but hey..)

Internal Quail

spaceemotion Wow, this is incredible! Thank you so much for your help!

Can't wait for the next release to see this came true! :D

spaceemotion

Internal Quail: Managed to find the issue with the word counting - any tone marks weren't taken into consideration. If you want, you can give me some more examples to test with :)

Internal Quail

spaceemotion  You work so fast! Sure, here are some more examples you can use for testing : 
นิกซ์ก็ร้องขึ้นว่า (Nyx loudly said) - 5 words, 1 name
“จับเจ้าโบนไว้!” ("Catch Bone!") - 4 words, 1 name
แล้วนิกซ์ก็ใช้มือข้างหนึ่งจับแขนซัมเมอร์ (then, Nyx used one hand to grab Summer's arm) - 10 words, 2 names
“ฟรานเชสก้าห้ามคุณทำร้ายผม” ("Francesca forbade you to harm me") - 5 words, 1 name
"...แน่นอนว่ายกเว้นกับฉัน ก็ฉันเป็นเทพีแห่งกาลเวลานี่นะ” ("Except for me, of course. I am the goddess of time, after all.") - 13 words, No name [but I assigned "เทพีแห่งกาลเวลา" (the goddess of time) as one of the character's Aliases/Nicknames in Codex]
แอนดราสและดาร์บี้หันกลับไปมองนิกซ์ (Andras and Darbi turn back to Nyx) - 8 words, 3 names
“คุณเป็นแอนิมัสเหรอ” แอนดราสถาม ("Are you an Animus?" asked Andras) - 6 words, 1 character's name, +1 name from lore (Codex Entry)
I hope this help!  Again, I am very happy and very impress of how fast you found ways to fix things. Thank you so much!
Please let me know if I can do anything else to help! :)

spaceemotion

marked this post as

planned

Hi there, currently the system only supports Chinese for these situations as I have to handle them individually. I can add another special case for Thai.
However, that would also affect word counting. Is NC accurate with word counting for Thai right now? or wayy off? (I would need to fix it along with the support for inline names for Thai)
If not, are there any good guides on how to count words in Thai? Thanks!

Internal Quail

spaceemotion Thank you SO MUCH for your reply. Here are my answer,

Yes, please add another special case for Thai. That would be a big help. :)
No, Novelcrafter's word counting for Thai is not accurate (you're right, it's wayyy off now).
There are no easy indicators for how to separate Thai words without recognizing Thai vocabularies as context. Sorry about that. Believe me, I've been trying to come up with rules for the last 2 hours.
I understand that the algorithm is now (detecting words) ---> (match each word with Codex), if that's the case, can we eliminate the (detecting words) for now? And just go straight to matching alphabets/marks from the whole text? For example, can it find "Sarah" from "hellomynameissarah" right away without having to separate each word first?

Just so we can at least use the Codex properly for now. The word count is not working anyway, right?

We can figure out how to count Thai words later, and I will help you with anything I can. I'll work with you. It's my language, so it will be easy for me. For now, I just need Novelcrafter for AI, and AI need to recognize names to work efficiently.

Please let me know. Thank you again! :)