Get the character index as the animator sees it.

[Fabrice Leconte]

Hello, did you try to play with normalize method?

str = '\uFB01ne'
normalized = str.normalize('NFKC')
str + ' = ' + str.length + ' chars | ' +
normalized + ' = ' + normalized.length + ' chars'

[Andrei Popa]

Hi Fabrice, thanks for your reply.

This is indeed interesting, but my problem is kinda backwards. I have the “fine” source text. When I check “Ligatures”, it becomes ‘\uFB01ne’. I wonder if I can somehow get that string, the one that is seen after I check ligatures (the 3 chars long one), from “fine”.

I need this because I want to make some words to change color in arabic languages. In the latest update, Adobe made it so animators correctly count characters (right-to-left, before it counted character left-to-right even if the text layer was RTL). The problem is that arabic has a lot of ligatures, and the Javascript that counts the characters to find the match and the animator that changes the colors count differently. Javascript counts 2 characters per ligature and animator counts one. I get the string via expressions, so I don’t know how to transform it from no ligatures to ligatures, so the Javascript counts characters the same way that the animator does.

I don’t know arabic so I work with some text I took from a lorem ipsum site. I have isolated some of the exceptions. One of them is this لإ which is seen as 2 characters by text.sourceText.length and as 1 by the animator. I observed the same behavior when dealing with the fi ligature so I figured out it’s the same problem and I am trying to fix the fi one.

I hope all of this make sense.

[Fabrice Leconte]

I probably missed something but just using this expression I have a good result.

it doesn’t matter if the ligature box is checked or not.

// Source Text Expression
'unde\uFB01ned '.repeat(2)
// Expression Selector Expression
var str = text.sourceText;
str.charAt(textIndex-1) == 'n'?100:0;

[Andrei Popa]

The problem is that I do not have the ‘unde\uFB01ned ‘.repeat(2) as source text.

I have the source text <i style=”font-family: inherit; font-size: inherit;”>’undefined ‘.repeat(2). But it looks like this due to ligatures in font.

I need a way to get the string as it is after the text engine process it. ‘fi’ is not the only ligature in this font. It also has fl, and maybe some more, and they are automatically transformed.

In this example I used this expressions:

// Source Text Expression
'undefined '.repeat(2)
// Expression Selector Expression
var str = text.sourceText;
str.charAt(textIndex-1) == 'n'?100:0;

[Fabrice Leconte]

Using Roboto font, I have the same result as you.

But if I change to Arial or some others fonts, it works correctly.

[Fabrice Leconte]

I fixed the problem, I created a function to convert a regular string to ligature string

Just 1 problem, the expression is about 8700 lines. 😂

You’ll probably need to remove some useless entries in the array such as asian characters if you don’t need.

https://imgur.com/a/KfZ3fWH

View post on imgur.com

[Fabrice Leconte]

The code:

ligature_expression.zip 12 KB Zip - Click to view

• Copy Download Link

[Andrei Popa]

This is really impressive.

May I ask where you got the array with the ligatures from?

[Fabrice Leconte]

I extracted ligatures from this list Unicode Data