Search for unicode character (ranges)

[Azraelo]

Hello,

i've got an epub with multiple weird unicode characters and wanted to use a regex to get rid of it in the epub editor.

Example of text in the epub:
𝒷.𝓬𝑶𝐦

According to my research these character should represent the following unicode characters:
\u1D4B7
\u002E
\u1D4EC
\u1D476
\u1D426

But no matter what I try, the search function never matches those characters.
I've opened a text file within the epub editor, put "\u1D4B7" into the search part and changed the modus to "Regex".
When searching, nothing is found.
If I search for "[\u1D400-\u1D4FF]", then all normal characters are listed as match (a-zA-Z).

What is the logic behind this?

My intention was, to search for something like this:"[\u1D400-\u1D4FF\u002E]{4,20}" and replace it with nothing.
Can you please give me a hint, how to accomplish this?

Regards
Azraelo

[Sarmat89]

Try \U instead.

[Azraelo]

If i try "\U1D4B7", then I get the error message:
calibre, version 6.21.0
FEHLER: Ungültige RegEx: <p>Der reguläre Ausdruck, den Sie eingegeben haben, ist ungültig: <pre>\U1D4B7</pre> mit Fehler: incomplete escape \U1D4B7 at position 2

So this doesn't really help

[DNSB]

Did you try escaping the \? i.e. \\U1D4B7 since most regex implementations use the backslash as a special character. Though I'm not sure if calibre allows use of that format for a Unicode character.

Alternatively, just copy/paste the bscr ( �� ) into the search box. Interesting while I can see the character while entering the message, it does not show when I post the message and if I quote the OP's message it still doesn't show.

Originally Posted by Azraelo

��.������

[Quoth]

Originally Posted by Azraelo

Example of text in the epub:
��.������

Hmm...
Your browser is being strange?
Copy & paste ��.������

[Quoth]

Originally Posted by Azraelo

Example of text in the epub:
𝒷.𝓬𝑶𝐦

Hmm...
Browser is being strange?

[Quoth]

If I just quote it's OK

If I copy and paste it makes entire post bad?
𝒷.𝓬𝑶𝐦

OK in preview

[Quoth]

Baffling!
��.������
��.������

[Quoth]

Baffling!
𝒷.𝓬𝑶𝐦
𝒷.𝓬𝑶𝐦


This time "Go Advanced"

[Quoth]

So it's messed up unless you "Go Advanced" to see preview.

A site bug.

[Karellen]

It is strange how searching for "Mathematical Script Small B" \u1D4B7 won't work.

But \u00A0 for a No Break Space will work.

(Calibre Win10)

[Azraelo]

Unfortunately searching for "\\U1D4B7" didn't provide any results.
I have created this ebook from a website which randomly adds its domain into the text and relaces the characters with similar looking unicode characters. This is also done randomly, so the combination is never the same.
So just copying those characters directly for searching won't work for for more than one occurence.

Due to this I want to use the unicode range to catch those strange unicode characters using regex to remove them all at once.

[Quoth]

Spelling Check?

[Azraelo]

Nice idea, didn't think of this yet.
Unfortunately, there are hundreds of combinations of the url and the spelling check doesn't seem to provide a batch edit function for similar words.
If I wanted to use the editor, I would still have to hundreds of entries manually.

[BetterRed]

Originally Posted by Azraelo

Nice idea, didn't think of this yet.
Unfortunately, there are hundreds of combinations of the url and the spelling check doesn't seem to provide a batch edit function for similar words.
If I wanted to use the editor, I would still have to hundreds of entries manually.

Try the Sigil epub editor, IIRC it uses a different regex engine (but I could be wrong), failing that try using a text editor like Notepad++ (Win) or TextMate (Mac) via Calibre's (or Sigil's) Open with feature.

BR