EPUB e-book bought on Kobo website - random word spacings when displayed on Calibre

[garland667]

Hi,

I bought this book on the Kobo website, and I'd prefer to read it on my PC as I'm using it to prepare an exam.

The layout is great on my reader, but after downloading it from ADE and importing it to Calibre, it's barely readable, the layout is messy, with a lot of unneeded spaces between words.

I'm a complete newbie to Calibre, but I tried and checked the HTML code using the editor, and I noticed that whenever there's a word in bold type or some kind of special character, there's a superfluous space before and after.

Also these spaces are shown in the code as "<div class="pcalibre1 pcalibre t41">. And this "t41" goes all the way to "T9260" by the end of the book.

Any idea how I could fix all that?

[theducks]

What is the CSS for those "t###" selectors?

Spacing between words, might be JUSTIFICATION, padding to fill the line.
Folk that hate the gaps, usually prefer Left Justification (AKA Ragged Right)

Another possibility is the CSS has some weird Letterspacing set (My Kobo does spaces wrong and runs words together. Someone may have tried to force it. Browser based viewers like Calibre process letterspacing correctly)

[garland667]

Thanks for your reply.

Changing the alignment requires a conversion, right? I tried converting to EPUB with the option "Left Align" ticked but the resulting file still shows the same defects...

The CSS in "stylesheet.css" for the "t" selectors is the the following:

.t1 {
-moz-font-feature-settings: "liga" 0;
-ms-transform: matrix(0.759826, 0, 0, 0.759826, 0, 0);
-ms-transform-origin: 0 100%;
-webkit-transform: matrix(0.759826, 0, 0, 0.759826, 0, 0);
-webkit-transform-origin: 0 100%;
border-bottom: 0;
border-top: 0;
bottom: 712.386738px;
color: rgb(0, 0, 0);
display: block;
font-family: ff2;
font-size: 1.33333em;
font-style: normal;
font-weight: normal;
height: 31.898749px;
left: 366.766739px;
letter-spacing: 0;
line-height: 1.2;
margin-bottom: 0;
margin-top: 0;
padding-bottom: 0;
padding-top: 0;
position: absolute;
text-indent: 1.5em;
text-shadow: -0.015em 0 transparent, 0 0.015em transparent, 0.015em 0 transparent, 0 -0.015em transparent;
transform: matrix(0.759826, 0, 0, 0.759826, 0, 0);
transform-origin: 0 100%;
visibility: visible;
white-space: pre-wrap;
width: 7545px;
word-spacing: 0;

Similarly, there are 4596 ".pcalibre" entries, which include this stuff:

}
.pcalibre4596 {
-moz-font-feature-settings: "liga" 0;
-ms-transform: matrix(0.737031, 0, 0, 0.759826, 0, 0);
-ms-transform-origin: 0 100%;
-webkit-transform: matrix(0.737031, 0, 0, 0.759826, 0, 0);
-webkit-transform-origin: 0 100%;
border-bottom: 0;
border-top: 0;
bottom: 491.460946px;
color: rgb(0, 0, 0);
display: block;
font-family: ffe;
font-size: 1.33333em;
font-style: normal;
font-weight: normal;
height: 31.05px;
left: 990.830985px;
letter-spacing: -0.375px;
line-height: 1.2;
margin-bottom: 0;
margin-top: 0;
padding-bottom: 0;
padding-top: 0;
position: absolute;
text-indent: 1.5em;
text-shadow: -0.015em 0 transparent, 0 0.015em transparent, 0.015em 0 transparent, 0 -0.015em transparent;
transform: matrix(0.737031, 0, 0, 0.759826, 0, 0);
transform-origin: 0 100%;
visibility: visible;
white-space: pre-wrap;
width: 7545px;
word-spacing: 0.825px;
}

If I understand well, those are parameters that define the characteristics of each character in the text, like font, alignement, size...?

If so, why are there so many of them?

[theducks]

Garbage original code. Somethin is slightly different in each.
Calibre scans the book, building a class list for each unique use.
Then it assigns it where called for.
I have seen OCR come out with an excess (nothing like yours), were a margin is 1 or 2 pt different as a side effect of the document OCR

[Karellen]

Looks like the book needs to be completely re-coded.

[kovidgoyal]

Convert it to MOBI and then convert the MOBI back to EPUB that will likely discard all the problematic CSS automatically.

[garland667]

I've tried converting the EPUB to MOBI, but the resulting file is...empty, it mostly consists of black pages along with a bunch of white ones with some lines...

I'm grateful for your inputs, any other idea? Is there anything I could provide you with to come up with a solution?

[PeterT]

How about sharing the name of the book..

[kovidgoyal]

Then it will almost certainly be a fixed layout book. These arent supported in calibre

[JSWolf]

Originally Posted by kovidgoyal

Then it will almost certainly be a fixed layout book. These arent supported in calibre

I agree that it's a fixed layout ePub.

[garland667]

How about sharing the name of the book..

Of course! I didn't think it could help.

Fondamentaux - La Constitution commentée 2023-2024

Hachette Éducation
Release Date: August 30, 2023
Imprint: Hachette Éducation
ISBN: 9782017235576
Language: French
Download options: EPUB 3 (Adobe DRM)

Here's also a screenshot of what it looks like in Calibre reader:

[Karellen]

Maybe you could scramble the book... https://www.mobileread.com/forums/sh...d.php?t=267998
Then upload the scrambled version.

[garland667]

Originally Posted by Karellen

Maybe you could scramble the book... https://www.mobileread.com/forums/sh...d.php?t=267998
Then upload the scrambled version.

Done! Here's the scrambled version of the book.

Fondamentaux - La Constitution commentee 2 - Simon-Louis Formery_scrambled.epub

[DNSB]

I took a look at a sample of the book I downloaded from Kobo.

<meta property="rendition:layout">pre-paginated</meta>
<meta property="rendition:orientation">auto</meta>
<meta property="rendition:spread">auto</meta>

I found the above lines in the content.opf file which says that this is a ePub 3 fixed layout book.

<meta name="viewport" content="width=1508, height=2048"/>

This viewport size declaration in each of the xhtml files tells you the same thing.

Converting it to any other format is not an easy task. You will notice that every page in the book has it's own xhtml file and everything on that page is given a fixed location within that viewport.

You can use Thorium on your computer or sending it to a Kobo ereader as a kepub file to read the FLO book. Your scrambled version corrupted quite a bit of the fixed layout information and is unlikely to be viewable on anything. I'm not how well ADE 4.5 will deal with this since it's support for FLO is not state of the art.

[PeterT]

Have you thought of using Kobo Desktop to read this on your PC? Kobo Desktop