Add logic to the search and replace words script.

Good day. I will send you the files in personal messages when I see your bid and your profile. I also need a professional in JS for ongoing collaboration. I have a script in js and a file in json, in which the words are specified for replacement. It needs to be corrected. I will send you 2 js files with the script itself. The script code starts from line 261 replace.

And now what exactly needs to be done.

Currently, the search and replacement of words works like this: in the Json file from = word and to = word

And there is a problem with short words of three letters, for example, руб should be replaced with грн. 

The script does this. Example:

In the article, it finds 100 руб and replaces it with 100 грн. And this is correct.

BUT! When the script encounters any word that contains 3 letters = руб, it replaces them with грн. And it creates a lot of spelling mistakes. It turns normal words into abnormal ones. I showed this in the screenshot.

It simply found the letters that need to be replaced and blindly replaces them in any words, whether they have 3 letters or 45 letters, thus creating abnormal words. For example:

It found the word аеротруба and replaced it with аеротгрну because it found the 3 letters руб in the word and replaced them with 3 letters грн. Thus, it distorted the word. 

It found the word рубашки and replaced it with грнашки. 

And this happened with hundreds of words in different articles.

It is necessary to make the script precise regarding the number of letters it searches for. For example, if the word in from is a three-letter word, then it should search for words that have no more than 3 letters. And if the word has more than three letters, then it should not replace letters in that word, because it is a different word. I think you understand the essence.

Only make sure it doesn't break my other conditions. There are conditions where words have more than three letters. For example, not руб, but рублів. So here the script should understand that we are making replacements in words that have no more than 6 letters.

In other words, if in the from condition: the word that needs to be replaced consists of three letters, then the script searches for words that have 3 letters and replaces them. Not all words that contain those three letters.

If in the from condition: the word that needs to be replaced consists of 10 letters, then the script should search for these letters only in words that have no more than 10 letters.

Can you IMPROVE THE SCRIPT THAT EXISTS? It works fine. It just lacks precision regarding the number of letters in which it should make replacements. It makes replacements in all words, regardless of the number of letters, and thus in some words the replacement is correct. 100 руб to 100 грн = THIS IS NORMAL! But the word рубашки was replaced with грнашки = THIS IS NOT NORMAL.

I was told earlier that it is necessary to make its logic based on regex. And then it will work more accurately.
Also, in the script, there is code for escaped tags, do not break this line of code. Because! It is responsible for replacing escaped tags, because if something is redone and this piece of code is forgotten, then replacements will not occur in all words. I will explain in detail if needed.

I can test your version of the script myself. Or connect to me via anydesk. And we will test it together. 

Evaluate the task. And I will give you the project.