Couple of newbie questions (text cut off, wrong fonts)

Shohreh · 08-20-2024, 10:29 AM

Hello,

To make them more readable, I'm trying to replace fonts in PDFs from serif to serif.

After running pdf2html and editing the file in the latest (2.3.0) release of Sigil, the EPUB I get has two issues:
1. The text is cut off on the right

2. It still uses serif instead of sans serif

Code:

.ft0{font: 14px 'Verdana';line-height: 21px;}

Any idea why?

Thank you.

KevinH · 08-20-2024, 10:57 AM

My guess is the pdf to html tool created a fixed layout file, where css provides the exact location of each word start or even each character. It is non-reflowable. There is no good reason to create fixed layout epubs/html as pdf handles fixed layout better.

Does your pdf to html tool have an option to create a reflowable html file?

retiredbiker · 08-21-2024, 01:32 AM

It seems you are trying to get the text out of a pdf just to replace the font, then make a new pdf..right? So getting the text out properly is the problem; try the things below.

Once you have the text out, by far your best bet to get a new pdf is to put the text into Writer or Word, format it the way you want, including the font, and export the new pdf from the word processor app. Trying to go pdf-->html-->epub-->pdf is just too hard and fraught with problems. No pdf on the planet will have styling information to allow you to get the text out and put it back with the same appearance. They just aren't made that way.

Calibre uses pdftohtml, a Popper tool, in it's attempts to convert pdf files. It is not always successful. There is a pdftotext tool that will sometimes work when the html tool fails. You called it "pdf2html" so you may have some completely different tool. There is no shortage of similar tools. Try here: https://github.com/elswork/poppler-utils

The bottom line is there can be anything, any kind of garbage, in a pdf file and some of them are just about impossible to get good text out of. Another thing you might try is using an OCR tool to give the pdf your own text layer in one go, and then converting. Again there are many tools.

The last gasp is to do your own OCR page by page from screen shots. Or pull out png images with pdftopng, then OCR. Lots of work.

PDFs are mostly evil things in the ebook world, especially old ones made with who-knows what tools.

retiredbiker · 08-21-2024, 01:55 AM

I downloaded your picture so I could enlarge it and see it better. The text as shown at the bottom, in the Sigil preview, actually seems pretty complete and flowing. Try opening the html text you got directly with Writer...that might work well after all (Or Word, it probably does the same but I don't use it.) Then you could fiddle the font and export a new pdf.

What are you using to view the epub? With the preview looking as good as it does, that could be a problem.

BetterRed · 08-21-2024, 02:01 AM

Quote:

Originally Posted by retiredbiker

It seems you are trying to get the text out of a pdf just to replace the font, then make a new pdf..right? So getting the text out properly is the problem; try the things below.

Once you have the text out, by far your best bet to get a new pdf is to put the text into Writer or Word, format it the way you want, including the font, and export the new pdf from the word processor app. Trying to go pdf-->html-->epub-->pdf is just too hard and fraught with problems. No pdf on the planet will have styling information to allow you to get the text out and put it back with the same appearance. They just aren't made that way.

Calibre uses pdftohtml, a Popper tool, in it's attempts to convert pdf files. It is not always successful. There is a pdftotext tool that will sometimes work when the html tool fails. You called it "pdf2html" so you may have some completely different tool. There is no shortage of similar tools. Try here: https://github.com/elswork/poppler-utils

The bottom line is there can be anything, any kind of garbage, in a pdf file and some of them are just about impossible to get good text out of. Another thing you might try is using an OCR tool to give the pdf your own text layer in one go, and then converting. Again there are many tools.

The last gasp is to do your own OCR page by page from screen shots. Or pull out png images with pdftopng, then OCR. Lots of work.

PDFs are mostly evil things in the ebook world, especially old ones made with who-knows what tools.

Reading the text in the OPs screen shot I see it's about Lithium ion batteries, so it's a recently created PDF. I would try opening the PDF in current Word or maybe LO Writer, change the font and save as PDF.

If the OP wants EPUB then they can save as DOCX and convert that to EPUB… via the Sigil plugin, the ePub-Tools addin for Word, or via calibre. The last is probably the easiest.

BR

Shohreh · 08-21-2024, 06:01 AM

Thanks for the infos.

The pdftohtml I mentioned is the one from Apryse.

I view EPUBs through SumatraPDF (Windows) or a 6" e-reader.

I'll try the poppler → LibreOffice solution.

KevinH · 08-21-2024, 08:51 AM

Yes the Apryse pdf to html tool creates a fixed layout epub not a reflowable one according to its website. The fixes the location of every word on thar page so it will not reflow to adhjust to screen width.

Shohreh · 08-21-2024, 01:10 PM

That figures

08-20-2024, 10:29 AM	#1
Shohreh Groupie Posts: 181 Karma: 304158 Join Date: Jan 2016 Device: none	Couple of newbie questions (text cut off, wrong fonts) Hello, To make them more readable, I'm trying to replace fonts in PDFs from serif to serif. After running pdf2html and editing the file in the latest (2.3.0) release of Sigil, the EPUB I get has two issues: 1. The text is cut off on the right 2. It still uses serif instead of sans serif Code: .ft0{font: 14px 'Verdana';line-height: 21px;} Any idea why? Thank you. Attached Thumbnails

08-20-2024, 10:57 AM	#2
KevinH Sigil Developer Posts: 8,160 Karma: 5450818 Join Date: Nov 2009 Device: many	My guess is the pdf to html tool created a fixed layout file, where css provides the exact location of each word start or even each character. It is non-reflowable. There is no good reason to create fixed layout epubs/html as pdf handles fixed layout better. Does your pdf to html tool have an option to create a reflowable html file? Last edited by KevinH; 08-20-2024 at 11:18 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Text formatting (newbie questions)	kamanza	Sigil	9	08-17-2011 07:50 PM
My kindle was cut in its prime a couple of days ago	tzooka	Amazon Kindle	10	01-31-2011 11:42 AM
Couple of newbie questions about Jetbook	greenapple	Ectaco jetBook	10	01-30-2010 01:27 AM
Newbie here - couple of questions	puremagic	Sony Reader	19	01-05-2009 06:29 PM
A couple of more newbie questions	reader_newb	iRex	4	04-21-2008 03:47 PM

08-21-2024, 01:32 AM	#3
retiredbiker Evangelist Posts: 420 Karma: 2737916 Join Date: May 2013 Location: Ontario, Canada Device: Kindle KB, Oasis, Pop_Os!, Jutoh, Kobo Forma	It seems you are trying to get the text out of a pdf just to replace the font, then make a new pdf..right? So getting the text out properly is the problem; try the things below. Once you have the text out, by far your best bet to get a new pdf is to put the text into Writer or Word, format it the way you want, including the font, and export the new pdf from the word processor app. Trying to go pdf-->html-->epub-->pdf is just too hard and fraught with problems. No pdf on the planet will have styling information to allow you to get the text out and put it back with the same appearance. They just aren't made that way. Calibre uses pdftohtml, a Popper tool, in it's attempts to convert pdf files. It is not always successful. There is a pdftotext tool that will sometimes work when the html tool fails. You called it "pdf2html" so you may have some completely different tool. There is no shortage of similar tools. Try here: https://github.com/elswork/poppler-utils The bottom line is there can be anything, any kind of garbage, in a pdf file and some of them are just about impossible to get good text out of. Another thing you might try is using an OCR tool to give the pdf your own text layer in one go, and then converting. Again there are many tools. The last gasp is to do your own OCR page by page from screen shots. Or pull out png images with pdftopng, then OCR. Lots of work. PDFs are mostly evil things in the ebook world, especially old ones made with who-knows what tools.

08-21-2024, 01:55 AM	#4
retiredbiker Evangelist Posts: 420 Karma: 2737916 Join Date: May 2013 Location: Ontario, Canada Device: Kindle KB, Oasis, Pop_Os!, Jutoh, Kobo Forma	I downloaded your picture so I could enlarge it and see it better. The text as shown at the bottom, in the Sigil preview, actually seems pretty complete and flowing. Try opening the html text you got directly with Writer...that might work well after all (Or Word, it probably does the same but I don't use it.) Then you could fiddle the font and export a new pdf. What are you using to view the epub? With the preview looking as good as it does, that could be a problem.

08-21-2024, 06:01 AM	#6
Shohreh Groupie Posts: 181 Karma: 304158 Join Date: Jan 2016 Device: none	Thanks for the infos. The pdftohtml I mentioned is the one from Apryse. I view EPUBs through SumatraPDF (Windows) or a 6" e-reader. I'll try the poppler → LibreOffice solution.

08-21-2024, 08:51 AM	#7
KevinH Sigil Developer Posts: 8,160 Karma: 5450818 Join Date: Nov 2009 Device: many	Yes the Apryse pdf to html tool creates a fixed layout epub not a reflowable one according to its website. The fixes the location of every word on thar page so it will not reflow to adhjust to screen width.

08-21-2024, 01:10 PM	#8
Shohreh Groupie Posts: 181 Karma: 304158 Join Date: Jan 2016 Device: none	That figures