Regex for Marking Text?

[Turtle91]

Does anyone know of a way to restrict a find/replace to marked text....within the search parameter? I know you can manually mark the text with Ctrl + Shift + M, but I'm hoping to do it as part of a saved search group.

eg.

Step one:
find: <table(.*?)>(.*?)</table> and set as marked text

Step two:
find (within marked text): <p>(.*?)></p> and replace: \1

Step three:
Change the 'marked text' selector back to "All HTML Files"

Thanks,

[Doitsu]

If it's usually only one section that you want to update, simply highlight it, right-click it and select Mark Selected Text from the context menu.

All Find & Replace actions will only be applied to that section.

[Turtle91]

Yes, I'm familiar with the manual method. In this particular instance I am trying to clean out all the <p> tags inside of a table....and there are over 200 tables If I can build it as part of a saved search group that would save a lot of highlighting and right-clicking.

My other option, of course, is to have a much more complex search, but I was worried about having too many saved groups (.*?)...isn't there a limit of like 5 or 6 within a given regex?

[Klecks]

I don't have a solution for a sigil regex, but if you want to try the calibre editor, you can use the function mode.
The function mode gives you the possibility of more advanced text processing.
It should be perfect for your problem:

Open your book with the calibre editor.
Choose one of the html-files on the left panel and click Ctrl+F.
Insert "<table[^>]*>.*?</table>" in the field "Find:"
Choose Mode "Regex-function" and click create/edit.


In the new window you will have the basic structure of a function.
In the upper field you can insert a name (maybe "clean_tables") and the the given basic function you should replace with:

import regex
def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    return regex.sub(r'<p[^>]*>(.*?)</p>',r'\1',match.group())

Confirm with OK and you can use the new function.

Klecks

[Turtle91]

Wow - that's awesome! I had played with regex functions a while back and knew they were powerful, but hadn't thought about them in a while. I guess that shows we can get stuck in our ruts and fail to think outside of the comfort box.

Thanks!

[Turtle91]

I don't know the proprieties about sharing code between two open-source projects like Sigil & Calibre.

Is it a thing/possibility to implement the regex-functions capability within Sigil?

[Tex2002ans]

Originally Posted by Turtle91

Yes, I'm familiar with the manual method. In this particular instance I am trying to clean out all the <p> tags inside of a table....and there are over 200 tables If I can build it as part of a saved search group that would save a lot of highlighting and right-clicking.

I also have a set of 5 Saved Search Regexes I use to clean out Finereader table cruft:

Search: <td>\s+<p>([^<]+)</p>\s+</td>
Replace: <td>\1</td>

[...]

Search: <td colspan="([0-9]+)">\s+<p>([^<]+)</p>\s+</td>
Replace: <th colspan="\1">\2</th>

You could adapt something similar for your specific case.

[DiapDealer]

Originally Posted by Turtle91

I don't know the proprieties about sharing code between two open-source projects like Sigil & Calibre.

Is it a thing/possibility to implement the regex-functions capability within Sigil?

I'm guessing a Sigil plugin could be worked up that would allow similar functionality. You're basically creating a library of small custom python routines (using predefined parameters) to alter text matched by a regexp.

[Turtle91]

I've been a 'programmer' my whole life...not in any way professionally... basic/fortran/pascal/vba and wrote a little web app with Google Apps Script this week... which forced me to learn JS over the last couple of days. I haven't looked at python at all.

If it's OK with Kovid and the Calibre community I could peruse their code and give it my best shot at plug-in-ifying it???

[Turtle91]

Originally Posted by Tex2002ans

I also have a set of 5 Saved Search Regexes I use to clean out Finereader table cruft:

Search: <td>\s+<p>([^<]+)</p>\s+</td>
Replace: <td>\1</td>

[...]

Search: <td colspan="([0-9]+)">\s+<p>([^<]+)</p>\s+</td>
Replace: <th colspan="\1">\2</th>

You could adapt something similar for your specific case.

Thanks Tex!

[DiapDealer]

Originally Posted by Turtle91

If it's OK with Kovid and the Calibre community I could peruse their code and give it my best shot at plug-in-ifying it???

If you feel so moved, have it at!

Happy to offer some tips if you get snagged.

[Doitsu]

Originally Posted by Turtle91

I haven't looked at python at all.

IMHO, one of best Python crash courses is Al Sweigarts's Creative Commons licensed book Automate the Boring Stuff with Python, which is available online.

You also might want to check out the Beautiful Soup Python module, which is bundled with Sigil.

BTW, the last topic in the Sigil plugin development thread contains minimal code for a very simple plugin that boldens "the" in all HTML files.

[Turtle91]

Great - thanks!

That'll give me something to do while I'm sitting around waiting for Florence to get a move on...

[Turtle91]

"I know kung-fu."


Que training montage to get my fingers familiar with what's in my brain...