Requesting help to figure out how to get this to work using regular expressions.

[F4in7_]

I'm bad at describing how to make this work, but right now I have more than a dozen saved searches trying to align ellipses to each other including quotation marks.

Here's a list of them, and I'm figuring out a way to simplify them.

• ' …' to '…'
• '. . .' to '… '
• ' . . . ' to '…'
• ' . . .' to '…'
• '*…' to '…'
• '([A-z])…([A-z])' to '\1… \2'
• '‘… ' to '‘…'
• '… ’' to '…’'
• '‘. . . ' to '‘…'
• ' . . .’' to '…’'
• '“… ' to '“…'
• '… ”' to '…”'
• '‘ … ' to '‘…'
• ' … ’' to '…’'
• ' . . . ’' to '…’'
• ' … ”' to '…”'
• ' . . . ”' to '…”'

[hihohahi]

I have similar issue but with Arabic numbers. They randomly show up in my epub. I don't know any way to solve it so I made a simple macro AHK script and just run it every file.

Send, ^f ;press ctrl+F
Send {٢} ;type number 2 in arabic in FIND
Send {Enter} ;navigate to REPLACE
Send {Tab}
Send {2} ;type number 2
Send {Enter} ;navigate to REPLACE ALL
Send {Tab}
Sleep 300 ;wait 300 milisec inbetween otherwise it clicks too fast
Send {Tab}
Sleep 300
Send {Tab}
Sleep 300
Send {Tab}
Sleep 300
Send {Space}
Sleep 300
Send {Enter}
Sleep 300

The above is to replace ٢ with 2. repeat this for 0-9. it runs in about 20secs.

[EbookMakers]

Originally Posted by hihohahi

I made a simple macro AHK script and just run it every file.

This macro executes a find/replace. It seems to me that you use it to memorize the find/replace. You could run the Search / Saved search command: it displays a window that also allows you to memorize find/replace between different sessions.

You can have 10 saved searches to successively make changes 0-9 with Replace all.

Example:
Add search in the Saved search window.
name: Arabic_to_Latin_number_2
Find: ٢
Replace : 2
Mode: Normal (or regex)
and click on: Done

Same for the other numbers.

You can successively select the saved searches and click on Replace all.
You can also select the 10 searches then click on Replace all, The 10 searches are executed successively. But I cannot guarantee that the result will be the same.

[BetterRed]

Originally Posted by F4in7_

I'm bad at describing how to make this work, but right now I have more than a dozen saved searches trying to align ellipses to each other including quotation marks.

Here's a list of them, and I'm figuring out a way to simplify them.

• ' …' to '…'
  . . . <snip> . . .
• ' . . . ”' to '…”'

Perhaps you could do something with this plugin -->> Editor Chains

Automate various tasks in Calibre's ebook editor. Plugin allows to chain multiple actions together. You can choose from Plugin specific actions, calibre builtin actions, or create you own actions through the plugin's module editor.

Or the Editor's Regex function mode, see ==>> Function mode for Search & replace in the Editor

BR

[lomkiri]

Originally Posted by F4in7_

I'm bad at describing how to make this work, but right now I have more than a dozen saved searches trying to align ellipses to each other including quotation marks.

Here's a list of them, and I'm figuring out a way to simplify them.

If I understand correctly, you want to transform 3 dots (with or without spaces) in ellipsis, always remove a space before, and always remove a space after except if it's a letter.

You can do this with a regex-function :

— search : 
\s?(?:(?:\. ?\. ?\. ?)|…) ?(.)?

— replace (in mode regex-function):

def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    end = match[1] or ''   # be sure is not None
    space = ' ' if end.isalpha() else ''
    return '…' + space + end

I don't understand lines 2 and 5
> '…' to '… '
It seems that you want to put a space where everywhere else (except letter) you want to remove it ? I didn't consider this case because I didn't understand the purpose.

> '*…' to '…'
I don't understand this *. It is a jocker ? I didn't consider it either, but if you want to match a star, the search should be :

\s?\*?(?:(?:\. ?\. ?\. ?)|…) ?(.)?

___________
Edit 1: made optional the last char.
The cavit is that it will transform all ellipsis to themselves, but it shouldn't be a handicap in terms of time. If you don't want this, it can be avoid using 2 different regex-functions.

Edit 2: added how to remove a star

[lomkiri]

Originally Posted by hihohahi

I have similar issue but with Arabic numbers. They randomly show up in my epub. I don't know any way to solve it so I made a simple macro AHK script and just run it every file.

You can do it with a "replace all" and a simple regex-functon:

search: 
[\u0660\u0661\u0662\u0663\u0664\u0665\u0666\u0667\u0668\u0669]

Replace: 
def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    str_in  = '\u0660\u0661\u0662\u0663\u0664\u0665\u0666\u0667\u0668\u0669'
    str_out = '0123456789'
    table = str.maketrans(str_in, str_out)
    return match[0].translate(table)

[enuddleyarbl]

Originally Posted by lomkiri

If I understand correctly, you want to transform 3 dots (with or without spaces) in ellipsis, always remove a space before, and always remove a space after except if it's a letter.

You can do this with a regex-function :

— search : 
\s?(?:(?:\. ?\. ?\. ?)|…) ?(.)?

— replace (in mode regex-function):

def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    end = match[1] or ''   # be sure is not None
    space = ' ' if end.isalpha() else ''
    return '…' + space + end

It's been on my agenda to try to standardize the various ellipses I run across in books. Your Search, above, pretty much solves that for me. But, for my use, I find that I don't need the Regex-Function. With a minor modification to your search, a plain old Replace works for what I need:

Search: \s?(?:(?:\.\s?\.\s?\.\s?)|…)(\s?.)? (<-- EDIT: I've replaced the spaces with \s because sometimes they use non-breaking-spaces and plain space didn't pick it up)
Replace (as Regex): …\1

For the Search, all I've done is moved the opening parenthesis of your 2nd group (well, the only "selecting" group) to include the space if present. For the Replace, I just replace whatever form of spaces and ellipses the book already has, with just an actual ellipsis and tack on whatever the book currently has following it.

The problem with ellipses is that there doesn't seem to be any hard-and-fast rules for them. The best I've found (and there are a lot of contradictions) are:

1. if it's the old-fashioned 3-dot ellipse with spaces between the dots, then there should be spaces before and after the ellipse.
2. If the 3-dots are without intervening spaces, then there's shouldn't be spaces before or after.
3. sentences ending with elllipses and a period or comma should have the period/comma first and then the ellipsis.
4. sentences ending with question or exclamation marks should have the ellipsis first and then the question/exclamation mark.

Personally, I think rule 3, above, is for the birds. In general, trailing ellipses seem to be for thoughts trailing off. Not for sentences trailing off (which is what the rule implies to me). And, for rules 1 and 2, we're replacing those 3 dots with an actual ellipsis, so they don't really apply.

Soft rules I've found for ellipses seem to say there should always be spaces before and after unless one bumps up against a closing quote. In that case, the ellipsis bumps right up to the closing quote.

But, most of what I see in actual books have opening ellipses right up against the start of sentences, closing ellipses right up against the end of the sentences, and embedded ellipses adjacent to the previous bit of sentence and having a space before the sentence resumes (I think).

And, that's why I changed your search and replace. If I've worked through it correctly, all I've got it doing is stripping off any leading spaces, replacing any 3-dot ellipses with real ellipses and using whatever trailing spaces the author/publisher decided to use.

That seems to do what I need. So, thanks for the regex. It sure made things easier for me.

EDIT: and what is it with Pratchett and ellipses? I just opened a book of his to test this and there are 828 ellipses in the book.

[lomkiri]

The regex-function was necessary because the OP was asking to put a space if the following char was a letter, and no space elsewhere. But if you decide to keep it as the editor wrote it (with or without a space), you're right, no need for of a function.

But if the purpose is to unify the notations, maybe it's good to decide when to put or to strip out a space? In this case, a function can make the job (for example, if we want to strip spaces between ellipsis and quotation marks, and to force a space elsewhere).

You're also perfectly right about \s instead spaces, I should have done it, but as I was replying to his specific request…

Anyway, I'm glad I could help you with this regex.