Fix Hebrew highlighting regex (hedyorg#3500)

Fixes hedyorg#3476

The regex `CHARACTER` from [highlighting/definition.py](https://github.com/Felienne/hedy/blob/main/highlighting/definition.py) now includes Hebrew characters. The editor no longer recognizes Hebrew characters as end-of-word.

Before (level 11):
![image](https://user-images.githubusercontent.com/61731312/198108602-c46e2f13-6628-4399-9fc8-0e238c673939.png)

After:
![image](https://user-images.githubusercontent.com/61731312/198108713-78d490ee-5b6b-4140-b3b3-4f2f9db55b3d.png)

highlighting/definition.py

@@ -1,10 +1,10 @@
 # This file defines the special regexes
 
 # list of symbols recognized as characters (with non-Latin characters)
-CHARACTER = '0-9_A-Za-zÀ-ÿء-ي'
+CHARACTER = '[\p{Lu}\p{Ll}\p{Lt}\p{Lm}\p{Lo}\p{Nl}_\p{Mn}\p{Mc}\p{Nd}\p{Pc}·]'
 
 # definition of word
-WORD      = '([' + CHARACTER + "]+)"
+WORD      = '(' + CHARACTER + "+)"
 # space
 SPACE     = "( +)"
 
@@ -14,7 +14,7 @@
 
 # beginning and end of words
 START_WORD = '(^| )'
-END_WORD   = '(?![' + CHARACTER + '])'
+END_WORD   = '(?!' + CHARACTER + ')'
 
 DIGIT = '[__DIGIT__]'
 

highlighting/generate-rules-highlighting.py

@@ -1,5 +1,5 @@
 import os
-import re
+import regex as re
 import yaml
 import json