Calibre changes HTML markup during convert

[spiralpath]

Running Calibre 3.13 and when I "Convert books" to EPUB much of my markup changes. Some code I have added to stylesheet.css is removed.

And, it completely removed a file (TableOfContents.html) which seems to be lost forever.

[theducks]

Originally Posted by spiralpath

Running Calibre 3.13 and when I "Convert books" to EPUB much of my markup changes. Some code I have added to stylesheet.css is removed.

And, it completely removed a file (TableOfContents.html) which seems to be lost forever.

Ah! did you have that CSS properly linked to the TOC.HTML ?
AFAIK, calibre only uses what is linked.
AND you DO NOT want calibre to replace the TOC

There are so many settings.

Remember ! Preferences: ... is the DEFAULT that is used for the INITIAL conversion.

Once used, the book retains the settings that were used (may be modified on the conversion start). Fro then on, the Conversion screen shows the settings that were used previously. This is PER BOOK
There is a button (tick in bulk mode) to cause conversion to forget, and grab a fresh 'default'

[kovidgoyal]

The only way calibre will completely remove a file is if you incorrectly marked it as the titlepage or it is not in the spine.

As for removing CSS, calibre completely rewrites all css, flattening it and keeping only the CSS that actually applies to your markup.

[spiralpath]

Originally Posted by theducks

Ah! did you have that CSS properly linked to the TOC.HTML ?
AFAIK, calibre only uses what is linked.
AND you DO NOT want calibre to replace the TOC

There are so many settings.

Remember ! Preferences: ... is the DEFAULT that is used for the INITIAL conversion.

Once used, the book retains the settings that were used (may be modified on the conversion start). Fro then on, the Conversion screen shows the settings that were used previously. This is PER BOOK
There is a button (tick in bulk mode) to cause conversion to forget, and grab a fresh 'default'

Thank you, theducks, for your time and help.

My biggest challenge at the moment is having my HTML markup changed when I run "Convert Book" (to EPUB). I add markup like <span id="chap1-2">some text </span> and after the conversion runs and I "Edit Book," the <span... markup is removed.

Also, I will add <p class="bodyText"> to the beginning of each paragraph and after "Convert Book" that markup is changed to <p class="calibre7"> and the bodyText rule is removed from the css file.

[PeterT]

I hate to be the bearer of bad news but if you do a conversion, calibre will rewrite / merge all the CSS and assign the new classes its own names.

Sent from my Nexus 7 using Tapatalk

[spiralpath]

Originally Posted by PeterT

I hate to be the bearer of bad news but if you do a conversion, calibre will rewrite / merge all the CSS and assign the new classes its own names.

Sent from my Nexus 7 using Tapatalk

Thank you, Peter.

I am not attached to the css classes I name, however I am very partial to the markup that is being removed completely. I am using the following markup to identify specific sub-headings in the text that are linked to the table of contents I created in TableOfContents.html, e.g.:

excerpt from TableOfContents.html:

<p class="toc-level1"><a href="chapter-1.html#chap1-3">Checking Goals at the Door</a></p>

excerpt from chapter-1.html:

<p class="subhead"><span id="chap1-3">Checking Goals at the Door</span></p>

Oddly, the Calibri conversion process does not change the subhead class but it removes the span tags completely. Hence, my strategy to link my table of contents via HTML is foiled.

excerpt from chapter1.html, post-conversion:

<p class="subhead">Checking Goals at the Door</p>

[spiralpath]

Originally Posted by kovidgoyal

The only way calibre will completely remove a file is if you incorrectly marked it as the titlepage or it is not in the spine.

As for removing CSS, calibre completely rewrites all css, flattening it and keeping only the CSS that actually applies to your markup.

Thank you very much, kovidgoyal,

I am not attached to the css classes I name, however I am very partial to the markup that is being removed completely. I am using the following markup to identify specific sub-headings in the text that are linked to the table of contents I created in TableOfContents.html, e.g.:

excerpt from TableOfContents.html:

<p class="toc-level1"><a href="chapter-1.html#chap1-3">Checking Goals at the Door</a></p>

excerpt from chapter-1.html:

<p class="subhead"><span id="chap1-3">Checking Goals at the Door</span></p>

Oddly, the Calibri conversion process does not change the subhead class but it removes the span tags completely. Hence, my strategy to link my table of contents via HTML is foiled.

excerpt from chapter1.html, post-conversion:

<p class="subhead">Checking Goals at the Door</p>

[kovidgoyal]

I dont see how calibre could possibly be removing span tags, unless you have set some conversion setting telling it to do so, such as heuristics or serach and replace, etc.

In any case those span tag are completely superflous, simply put your id on the <p> tag and you can link to it just the same.

[DNSB]

Originally Posted by spiralpath

Thank you very much, kovidgoyal,

I am not attached to the css classes I name, however I am very partial to the markup that is being removed completely. I am using the following markup to identify specific sub-headings in the text that are linked to the table of contents I created in TableOfContents.html, e.g.:

excerpt from TableOfContents.html:

<p class="toc-level1"><a href="chapter-1.html#chap1-3">Checking Goals at the Door</a></p>

excerpt from chapter-1.html:

<p class="subhead"><span id="chap1-3">Checking Goals at the Door</span></p>

Oddly, the Calibri conversion process does not change the subhead class but it removes the span tags completely. Hence, my strategy to link my table of contents via HTML is foiled.

excerpt from chapter1.html, post-conversion:

<p class="subhead">Checking Goals at the Door</p>

A quick look at your code suggests a possibility that your spans do not include a class so they are being removed as do nothing code. My personal feeling is that should only happen to <span> tags that do not include other information.

As suggested, using <p class="subhead" id="chap1-3"> would be cleaner. If you really want to use the spans, something like <span class="dummy" id="chap1-3"> would survive the conversion process.

[DNSB]

Interesting. I added a span with a id to a test epub and then converted it from epub to epub. Not only was a span with an id not removed, it had a class added. My theory about the spans being removed seems to have been shot down. Calibre 3.25, BTW.

Original epub:

<body class="epub">
  <p class="paranon"><span id="tarfu_1.3">This is a sample ebook with a few lines of text though it's hard to say what is actually a line of text when font size, screen width, margin size are variables beyond the control of the author. Of course, we could go to absolute measurements if we really want odd effects. EOT</span></p>
</body>

Converted epub:

  <body class="epub">
  <p class="paranon"><span id="tarfu_1.3" class="calibre">This is a sample ebook with a few lines of text though it's hard to say what is actually a line of text when font size, screen width, margin size are variables beyond the control of the author. Of course, we could go to absolute measurements if we really want odd effects. EOT</span></p>
</body>

The added class:

.calibre {
    line-height: 1.2
    }

[spiralpath]

Thanks, again, kovidgoyal,

Heuristics is turned off and there are no search and replace rules.

I put the id in the <p> tag, as you suggested, and the id was retained in the first occurrence and removed in subsequent occurrences.

This time, I named my table of contents file "Contents.html" and confirmed that it was included in content.opf. However, after conversion, Contents.html was removed.

[spiralpath]

Thank you, David,

Good suggestion to put the id in the <p> tag.

I put the id in the <p> tag, as you suggested, and the id was retained in the first occurrence and removed in subsequent occurrences.

This time, I named my table of contents file "Contents.html" and confirmed that it was included in content.opf. However, after conversion, Contents.html was removed.

[kovidgoyal]

Including it in content.opf is not enough you need to put it in the spine. Anyway, it's too difficult trying to guess what you are doing wrong, see https://www.mobileread.com/forums/sh...d.php?t=186697 for how to provide enough information to get useful answers.