k2pdfopt: optimizes PDFs for viewing on e-readers

[kmoll]

Hi, I cannot seem to find an option to get a reflowed output for use with 6 inch ereaders, but which contains selectable text that I can highlight on my ereader.
Is it not possible at all with k2pdfopt? I would be surprised, since Koreader, which uses your code, does exactly that. The documents I read are A4 in size and with one single column, so using "native pdf" and "2 columns" to cut the page in half vertically is not a solution, as each line of text runs horizontally across the page's width.
By the way, thank you so much for creating and maintaining this marvelous tool!!!
Regards

[willus]

Originally Posted by kmoll

Hi, I cannot seem to find an option to get a reflowed output for use with 6 inch ereaders, but which contains selectable text that I can highlight on my ereader.
Is it not possible at all with k2pdfopt? I would be surprised, since Koreader, which uses your code, does exactly that. The documents I read are A4 in size and with one single column, so using "native pdf" and "2 columns" to cut the page in half vertically is not a solution, as each line of text runs horizontally across the page's width.
By the way, thank you so much for creating and maintaining this marvelous tool!!!
Regards

It may depend on your reader. I ran your document with no special options and got the attached. The text is selectable using my PC PDF viewer, SumatraPDF. See the screen shot.

[vasilas7]

Hello Willus!
Congratulations for your work! I am trying to find the optimal options for k2pdfopt for my Kindle Paperwhite 4, unfortunately I can't.. Whenever I try to convert it, the result is impossible to read.
The pdf is attached, the language is Greek. I know that the pdf isn't the best, but I need it for my work.
Can you please help me?
Thanks in advance!

[willus]

Originally Posted by vasilas7

Hello Willus!
Congratulations for your work! I am trying to find the optimal options for k2pdfopt for my Kindle Paperwhite 4, unfortunately I can't.. Whenever I try to convert it, the result is impossible to read.
The pdf is attached, the language is Greek. I know that the pdf isn't the best, but I need it for my work.
Can you please help me?
Thanks in advance!

Actually, that PDF is very good for processing. It's perfectly straight, consistent from page to page, and very clean. First, I recommend downloading the Greek Tesseract OCR data set and installing it per these instructions. Then you can run one of the following commands depending how large you want the text.

1. Separate each page into two but don't do any text re-flow.
k2pdfopt -grid 2x1 -n- -ocr t -lang grc source.pdf

2. Same but with text re-flow
k2pdfopt -grid 2x1 -fc- -n- -f2p 0 -wrap -ocr t -lang grc source.pdf

3. Even larger text (50% larger with -mag 1.5)
k2pdfopt -grid 2x1 -fc- -n- -f2p 0 -wrap -ocr t -lang grc -mag 1.5 source.pdf

I've attached the results of these three methods for just page 5 of your PDF. You'll notice the text is selectable and searchable, unlike your original.

[vasilas7]

Originally Posted by willus

Actually, that PDF is very good for processing. It's perfectly straight, consistent from page to page, and very clean. First, I recommend downloading the Greek Tesseract OCR data set and installing it per these instructions. Then you can run one of the following commands depending how large you want the text.

1. Separate each page into two but don't do any text re-flow.
k2pdfopt -grid 2x1 -n- -ocr t -lang grc source.pdf

2. Same but with text re-flow
k2pdfopt -grid 2x1 -fc- -n- -f2p 0 -wrap -ocr t -lang grc source.pdf

3. Even larger text (50% larger with -mag 1.5)
k2pdfopt -grid 2x1 -fc- -n- -f2p 0 -wrap -ocr t -lang grc -mag 1.5 source.pdf

I've attached the results of these three methods for just page 5 of your PDF. You'll notice the text is selectable and searchable, unlike your original.

Thank you so much, Willus. If I select to crop the borders, what I have to do? Because in this pdf file the borders are not exactly in the same position.

[willus]

Originally Posted by vasilas7

Thank you so much, Willus. If I select to crop the borders, what I have to do? Because in this pdf file the borders are not exactly in the same position.

Do you mean crop the document like I've shown here so you don't get the page headers and page numbers (like I show in the first attachment)? Most pages are consistent, so the best way is to cover those pages and just live with the ones that it misses. You can pick specific pages to have different croppings, but that gets pretty involved. The GUI allows you to pick three different page ranges for applying different crop boxes. See attached screen shot. Here, instead of using the -grid 2x1 option, I have selected two crop boxes to apply to all pages.

[kmoll]

Originally Posted by willus

It may depend on your reader. I ran your document with no special options and got the attached. The text is selectable using my PC PDF viewer, SumatraPDF. See the screen shot.

Yes, indeed, I hadn't noticed that the text is editable... Sorry for making you loose your time...

BTW, I have another problem with a few files like the one attached, where the words are not cut properly. Do you happen to know how to fix this? Many thanks.

[willus]

Originally Posted by kmoll

Yes, indeed, I hadn't noticed that the text is editable... Sorry for making you loose your time...

BTW, I have another problem with a few files like the one attached, where the words are not cut properly. Do you happen to know how to fix this? Many thanks.

When I run k2pdfopt v2.51a on your source document, I get the attached--it seems to break the words okay. What version are you running, or what command-line options are you using?

[kmoll]

Originally Posted by willus

When I run k2pdfopt v2.51a on your source document, I get the attached--it seems to break the words okay. What version are you running, or what command-line options are you using?

Please find she options in used in the attached image. I am using the same version as you.

[willus]

Originally Posted by kmoll

Please find she options in used in the attached image. I am using the same version as you.

Check the "Smart line breaks" option. That is on by default. Not sure how you got it turned off.

[kmoll]

Originally Posted by willus

Check the "Smart line breaks" option. That is on by default. Not sure how you got it turned off.

That worked, indeed! I must have disabled that when playing with the options without having fully read the command line options detailed explanation.
A minor issue I have is that, in my workflow, I read pdfs on my ereader and annotate (highlight) them, and then continue working on them on my pc, annotating other stuff. The inconvenience I experience when using k2pdfopt is that, once the file is converted, the font looks too big on the pc screen. It would be great if we could somehow restore the original layout of the document to continue working with the file in windows (while not loosing the highlights previously made)... I got spoiled by the "on-the-fly" conversion made in koreader (which I don't use because there are bugs with the highlighting function).
Again, thank you so much for this wonderful tool and for your patient guidance.

[willus]

Originally Posted by kmoll

That worked, indeed! I must have disabled that when playing with the options without having fully read the command line options detailed explanation.
A minor issue I have is that, in my workflow, I read pdfs on my ereader and annotate (highlight) them, and then continue working on them on my pc, annotating other stuff. The inconvenience I experience when using k2pdfopt is that, once the file is converted, the font looks too big on the pc screen. It would be great if we could somehow restore the original layout of the document to continue working with the file in windows (while not loosing the highlights previously made)... I got spoiled by the "on-the-fly" conversion made in koreader (which I don't use because there are bugs with the highlighting function).
Again, thank you so much for this wonderful tool and for your patient guidance.

Sorry--not much I can do to help you with the highlighting issue. There's not really a way to do what you suggest above with k2pdfopt.

[kmoll]

Originally Posted by willus

Sorry--not much I can do to help you with the highlighting issue. There's not really a way to do what you suggest above with k2pdfopt.

OK, nevermind, thank you for your answer, I realized I can reduce the font by setting a fixed font size or by downgrading the resolution.
However, I noticed that, when trying to select more than one word, in some places, the whole document gets selected, like in the attached image. Even in the version you posted previously of this file, it also happens, for example when trying to select several words from the word "aboutit", line 10, onwards. This really is problematic because I can't highlight words...

Another question : is it possible to reduce the space between words? This time, to avoid asking a question which is already answered, I have looked carefully at the command line options documentation, but haven't seen anything related. As you can see in the attached image, the space between words is sometimes very wide... I have tried solving this both by setting a fixed font size and by playing with the resolution, to no avail.

[willus]

Originally Posted by kmoll

OK, nevermind, thank you for your answer, I realized I can reduce the font by setting a fixed font size or by downgrading the resolution.
However, I noticed that, when trying to select more than one word, in some places, the whole document gets selected, like in the attached image. Even in the version you posted previously of this file, it also happens, for example when trying to select several words from the word "aboutit", line 10, onwards. This really is problematic because I can't highlight words...

Another question : is it possible to reduce the space between words? This time, to avoid asking a question which is already answered, I have looked carefully at the command line options documentation, but haven't seen anything related. As you can see in the attached image, the space between words is sometimes very wide... I have tried solving this both by setting a fixed font size and by playing with the resolution, to no avail.

I sent you a private message. The highlighting thing is a known bug. I'm working on a bug-fix release, but it's slow. To get the words closer together, you need to turn off full justification:

k2pdfopt -j 0- ... (that's a zero, not an 'O').

[kmoll]

Originally Posted by willus

I sent you a private message. The highlighting thing is a known bug. I'm working on a bug-fix release, but it's slow. To get the words closer together, you need to turn off full justification:

k2pdfopt -j 0- ... (that's a zero, not an 'O').

Awesome! Both issues solved. Again, thank you very much!!