Inconsistent styling of comments in book details

[Jellby]

I have books with apparently the same comment formatting, but which are displayed differently in the book details panel: some have spaces between paragraphs, some don't. I can't find a reason for this. A couple of examples:

No spacing:


Spacing:


(left: book details panel; right: edit metadata dialog, HTML source tab.)

I don't see anything relevant in the HTML, do you? This is just after starting up calibre, no edits have been done.

There's the global styling mentioned here (https://www.mobileread.com/forums/sh...d.php?t=325867). But is there also some per-book styling? Or what is going on?

[kovidgoyal]

It's likely something in the underlying html that gets stripped out byt the conversion betwene normal and HTML view. try switching to normal view then back to html then clicking ok, does rendering normalise?

[Jellby]

Hmm... Going to HTML and and back to normal view makes the spacing appear in normal view, but no change in the panel. Of course, because I didn't make any edit. If I make some change (e.g. delete a letter), then the spacing appears alright. Interestingly, if in normal view I select all text and clear formatting... the spacing disappears, but that's a red herring, because the spacing is still there once I go to HTML and back or save.

Is there some easy way to apply this normalization to all books?

[chaley]

Originally Posted by Jellby

Is there some easy way to apply this normalization to all books?

It might be too big a hammer, but the Comments Cleaner plugin might do what you want.

[chaley]

FWIW: I found an example in my test library. Starting out, the raw HTML for a comment is

<div>05/07/13<div>Dernière page lue : Emplacement 6514 (99%)</div></div>

Looking at the comment in Edit Metadata I see the text shown in this image:


Clicking the HTML source tab I see:


The source HTML is now transformed. <p> tags are added around the date and the nested <div> is changed to <p>.

The Normal view is now:


Adding a space after the date to force the comment to be saved, the raw HTML becomes

<div>\n<p>05/07/13 </p>\n<p>Dernière page lue : Emplacement 6514 (99%)</p></div>

It seems like the transformation also includes pretty-printing the HTML.

[kovidgoyal]

Yes, the transformation is quite extensive, it has to be because the Qt rich text editor really badly mangles HTML, so calibre tries hard to reverse as much of the damage as possible. I dont think there is a bulk tool to apply that normalisation builtin but as chaley suggested the Comment cleaner plugin might do the trick.

[chaley]

Here is an action chain that uses the same method as calibre's comments editor to transform (clean) comments. Add it to Action Chains, which will create a chain "Clean comments for selected books". Apply it by selecting some books then running the chain. If you are brave, select your entire library.

As it uses calibre's method it shouldn't change anything that calibre's editor wouldn't change. That said, make a backup of your library before using it.

This is the template it uses to do the work.

python:
def evaluate(book, context):
	from Qt.core import QTextEdit
	from calibre.gui2.comments_editor import fix_html

	e = QTextEdit()
	e.setHtml(book.comments)
	ans = fix_html(e.toHtml(), e.toPlainText())
	return ans