<p class="block_15">, etc, etc...

[RMOP]

I recently converted a book found only in PDF format to MS Word, and then using Calibre, to ePub. Much of the conversion went quite well. However, the formatting is occasionally "off." Usually, this is a matter of a missing indentation of the first line of a new paragraph. I've traced this to the presence of this tag:

  <p class="block_16">

, whereas the properly indented paragraph first lines usually (but not invariably) use

<p class="block_15">

. In correcting this, I've also found numerous other variants of the

<p class="block_**">

. Some seem to work just like

<p class="block_15">

(e.g.,

<p class="block_38">

), but others are quite different in their effects. I have not enumerated all of the variants, but they are probably at least 10 in number.

FWIW, the

<p class="block_16">

occurs both at the beginning of a new paragraph, but also in mid-sentence of the last line of a paragraph. The result of the latter is to force a new, unindented paragraph consisted of the last fragment of that sentence.

1) Where is the best source for understanding the function of the many

<p class="block_**">

variants?

2) other than a search/replace of known offending variants, is there any way to fix these formatting glitches?

3) is there any way to reduce the likelihood of their being introduced in the first place?

Many thanks.

[Quoth]

Double check the paragraph styles in the MS Word docx.

Docx to epub conversion is just about perfect if the styles are correct.

[RMOP]

Quoth,

Thanks. I was so relieved to find all of the pdf text and images displayed nicely in a .docx file that I didn't even think to examine what styles were in effect in the document before converting to ePub. As it is, I managed a tedious but effective search/replace of the numerous odd CLASS codes in my ePub files. All 50 chapters...

[Quoth]

I forget how Word works, but you can search and replace styles in LO Writer. Styles in LO Writer* or MS Word map one-to-one with CSS classes when docx is converted to epub.

So I KNOW these spurious classes are spurious paragraph styles or direct formatting in the docx.

Of course direct formatting is very bad and harder to get rid off. Simply resetting format to paragraph style of identical as big as possible chunks of a chapter works.

[* Only edit LO Writer in odt format saves/loads after initial docx import, then for epub do an extra Save As in docx also. This is because otherwise if you edit docx it's a conversion every time with multiple similar page, paragraph, heading styles etc that need fixed each time the docx is opened. This is also an issue exchanging docx between different versions of MS Word!]

[RMOP]

Quoth,

Helpful. Thank you. I'm just using MS Word again after years of LO, but have minimal use of either for creating .docx documents intended for conversion to ePub format. The only clear upshot of my approach is that the embedded poetry in my document did get the necessary tags to appear properly in ePub. That was, unfortunately, offset by the many spurious tags which were eliminated by a tedious "search & destroy" mission. I'll pay more attention to styles in the LO or MS Word documents in future.

[Sarmat89]

I usually convert PDFs by hand with LibreOffice and regular expressions. The result can be easily converted to EPUB or FB2, and you can easily detect and unify inconsistent formatting.

[JSWolf]

Create a stye for just p such as

p {
  margin-top: 0;
  margin-bottom: 0;
  text-indent: 1.2em;
}

and edit out the classes you don't need and leave them as just <p>.