Recommended clean-up before HTML → EPUB?

[Shohreh]

Hello,

I'd like to concat a bunch of web pages into a single EPUB to read on my e-reader.

I tried pandoc, but it's very slow and pretty much freezes my computer, so I tried Calibre which at least kept my computer responsive:

copy /b *.html full.html

pandoc -o full.epub  full.html

"C:\Program Files\Calibre2\ebook-convert.exe" full.html full.epub

Regardless, how do you clean up HTML files before joining them into a single file? Any good practices?

"-h" returns a bewildering number of otptions.

Alternatively, what about first converting HTML files into simpler layouts (Markdown?) before joining them into a single file, and calling an HTML to EPUB converter?

Thank you.

[Turtle91]

I think the answer would depend on how comfortable you are working with the raw HTML code...

If you are OK with it, then Sigil (and I'm pretty sure Calibre) has a 'merge' feature that will remove the separate headers/footers and leave them combined into a single file. That process works well if the css is similar - or you make the css similar before merging.

If you are saying "what's raw html code" then I would suggest leaving the pages separate. You can still bundle them into an ePub - there is no requirement to have everything as a single page. It is actually more preferred to keep the files in ePubs separated logically, such as chapters. Both Sigil and Calibre editor can perform this function admirably. When you read the ePub, with the pages as separate files, it just requires a swipe/tap when transitioning from one file to the next.

[Shohreh]

Thanks. I'm used to working with HTML with Python.

I'm looking for a way to automate the process, and end up with pages that are as clean as possible on e-readers.

Can an EPUB contain multiple HTML pages?

--
Edit: Yup.

https://www.reddit.com/r/Calibre/com...a_single_epub/

https://manual.calibre-ebook.com/faq...specific-order

[najgori]

Before starting to work on epub I prefer to clean HTML to basic tags without any styling.
Step 1 is to open epub in Calibre editor, delete all css files, go to Remove unused css rules tool
Step 2 is Custom cleaner plus in Sigil for the rest of bookmarks, ids...
From Sigil you can export (x)html to editor of your choice or continue working in Sigil editor.

[Shohreh]

Thanks. I'll look into automating the process with a Python script.

[Doitsu]

Originally Posted by Shohreh

Thanks. I'll look into automating the process with a Python script.

Have a look at BeautifulSoup. It's a Python library for manipulating (X)HTML files.

[BobC]

Originally Posted by najgori

Before starting to work on epub I prefer to clean HTML to basic tags without any styling.
Step 1 is to open epub in Calibre editor, delete all css files, go to Remove unused css rules tool
Step 2 is Custom cleaner plus in Sigil for the rest of bookmarks, ids...
From Sigil you can export (x)html to editor of your choice or continue working in Sigil editor.

If you delete all css files then remove css rules you may well have got rid of important formatting. Some books, instead of

 <i> some italics></i>

will instead use something like

<span class="it"> some italics </span>

and use

.it {
 font-style: italic;
}

in styles.css or in css defined in the file header.

by using a brute-force technique as suggested all italics would be lost. There may also be "headers" that are just paragraphs styled to centred, bold and larger than the main text.

Look before you carry out drastic surgery.

BobC

[najgori]

You are right about italics. I don't care about the headers, though.