k2pdfopt: optimizes PDFs for viewing on e-readers

[mikbp]

Thanks for your answer!

The cheating way is not useful for me (or super-very-incredibly tedious) because I want to convert more than 100 pdf's :P
And as the non cheating way doesn't order it well, I will have to stay with the original pdf's.

Originally Posted by willus

I'll try to think about ways I could get k2pdfopt to handle cases like this better. I may even have some options in there to help with this, but I'm not remembering...

What I can think of is this:
Would it work better if there was no picture but just a blank spot? I mean, maybe it would be easier for the software to "see" the column if it could manage to first extract the image and check for the columns later. I see that with your commands line, you manage to define the image box pretty well (I tried it with a bunch of other pdf's and it sometimes also takes some text lines in the box, but many times just the image). Maybe if that is polished to avoid having text in the image box you can in a first step extract the image from the text, then convert the text and finally add again (or not) the image somewhere.

I don't really know how your software works and I'm not into it so I don't think that's going to help you much, but I had to try :P

Thank you again,
m

[markom]

If pictures are not that important to you, you can first quickly crop pdf using Briss app (separating columns cutting through pictures) and then use k2pdfopt.

Here I've left a few letters of the adjoining column partly visible too, but you can easily draw a cropping rectangles closer to the text of each column.

Maybe Willus can find a way to do this page cropping automatically throughout pdf, depending on the number of columns in the current page.

[willus]

Originally Posted by mikbp

Thanks for your answer!

The cheating way is not useful for me (or super-very-incredibly tedious) because I want to convert more than 100 pdf's :P
And as the non cheating way doesn't order it well, I will have to stay with the original pdf's.



What I can think of is this:
Would it work better if there was no picture but just a blank spot? I mean, maybe it would be easier for the software to "see" the column if it could manage to first extract the image and check for the columns later. I see that with your commands line, you manage to define the image box pretty well (I tried it with a bunch of other pdf's and it sometimes also takes some text lines in the box, but many times just the image). Maybe if that is polished to avoid having text in the image box you can in a first step extract the image from the text, then convert the text and finally add again (or not) the image somewhere.

I don't really know how your software works and I'm not into it so I don't think that's going to help you much, but I had to try :P

Thank you again,
m

Sorry for the late response. I still think I can get your case to work more reliably with some minor tweaking--maybe at the source code level. It's a good test case for me. But it does present a case where a processing algorithm like you describe would work better--try to find rectangular figure or table regions--regions that should not be re-flowed--first and then work around them. Right now k2pdfopt doesn't really process like that. I'm not 100% sure of all the details on how I'd do that, but it's on my list of things to work on.

[mikbp]

Hi people,

Originally Posted by willus

It's a good test case for me.

I'm glad you take as a challenge. =)

Originally Posted by markom

If pictures are not that important to you, you can first quickly crop pdf using Briss app (separating columns cutting through pictures) and then use k2pdfopt.

Is it possible to automate it for using it in more than 100 single-paged pdf files? That would be fine for me.

Thanks people
m

[willus]

Originally Posted by mikbp

Hi people,

I'm glad you take as a challenge. =)


Is it possible to automate it for using it in more than 100 single-paged pdf files? That would be fine for me.

Thanks people
m

You can do something similar to Markom's Briss option by just using k2pdfopt as long as all 100 of the PDF files have the same number of columns in the same positions (i.e. cut each page into the three columns like you would with Briss):

k2pdfopt -evl 1 -cbox 0.2,0,2.73,10.85 -cbox 2.92,0,2.41,10.85 -cbox 5.31,0 1.pdf

Each new -cbox option boxes one column. If each file is a little different, those -cbox numbers above would change for each file, so you'd have to manually figure them out each time, which would be tedious. But I still think I can get k2pdfopt to go around the figure more correctly in an automated way. I need some time to investigate.

[markom]

Originally Posted by mikbp

...

Is it possible to automate it for using it in more than 100 single-paged pdf files? That would be fine for me.

Thanks people
m

If you are using Briss or pdf-scissors you should merge all your pdf pages into one big pdf beforehand, using Acrobat or some free pdf merging/splitting tool.

If you are using k2pdfopt as Willus explained above by using three -cbox commands next to each other, you can put all the separated pdf files(pages) into one folder and open that folder in k2pdfopt.

You can quickly get precise coordinates for every cropping box by using that k2pdfopt' blue Margin Select" button, and drawing the first box then writing down those numbers shown on the top, then drawing the second and the third box getting their respective coordinates.

Then just enter those three -cbox commands into "Additional Options" field and k2pdfopt shall crop every pdf in the folder.

If some pdf pages contain wider or thinner columns you can just use wider columns cropping boxes for all the pages, because even if another column is partly visible, it's still comfortable to read the main column in portraite mode, 'cause 9 cm screen width of 6" reader is wide enough.

[mike2003]

I'm sorry, but I'm with the same request.
If you do not want to make your program completely portable (with the settings in the file in the program folder).
Can you at least keep the settings are not in the overall basket of garbage?
HKEY_CURRENT_USER\Environment

but somewhere in a normal place like all other programs?
HKEY_CURRENT_USER\Software\Adobe
for example
HKEY_CURRENT_USER\Software\k2pdf

It's so uncomfortable, how it now works!
You can not even save the settings and transfer them to another computer!
(reg EXPORT or regedit do not work with one key)

but I do not want to lose my profile in case of failure of OS.

[willus]

Originally Posted by mike2003

I'm sorry, but I'm with the same request.
If you do not want to make your program completely portable (with the settings in the file in the program folder).
Can you at least keep the settings are not in the overall basket of garbage?
HKEY_CURRENT_USER\Environment

but somewhere in a normal place like all other programs?
HKEY_CURRENT_USER\Software\Adobe
for example
HKEY_CURRENT_USER\Software\k2pdf

It's so uncomfortable, how it now works!
You can not even save the settings and transfer them to another computer!
(reg EXPORT or regedit do not work with one key)

but I do not want to lose my profile in case of failure of OS.

If you don't want to lose your k2pdfopt profile, then you can backup your environment variable settings, e.g. in MS Windows, from the command line:

set > d:\backup_folder\my_environment.txt

I do this anytime I switch to a new computer or re-install an OS. It's a helpful protective measure.

[mike2003]

I do not need a file with a bunch of information on any topic. Only my profile. how can I do it? no way! Yesterday I asked for my profile. I can not transmit it to another person! and that in 2015!
--
for fun question:
why do not you develop a program for the development of rules for the windows? or the rules of portability, as do almost all? What could be the reason to come up with something new instead of the proven and convenient methods ???

[willus]

Originally Posted by mike2003

I do not need a file with a bunch of information on any topic. Only my profile. how can I do it? no way! Yesterday I asked for my profile. I can not transmit it to another person! and that in 2015!

There are options under the "File" menu to save and restore your settings (see attached screen shot), making it possible to transfer your settings to another person or PC, and providing another way of making a backup.

[mike2003]

Yes, of course, but we must do it hands each time.
and the other person will lose all its settings should if I want to give him only one profile.

[willus]

Originally Posted by mike2003

Yes, of course, but we must do it hands each time.
and the other person will lose all its settings should if I want to give him only one profile.

The settings are stored as a simple text file. It is not any harder to mix and match profiles from two different settings than it is to modify or save registry settings.

[mike2003]

this man does not even know what the registry. convenience for you in last place. I do not understand you, but you are the author. Good luck.

[mikbp]

Thank you Willnus and Markom!
With your suggestions it is much easier to read

Originally Posted by willus

I still think I can get k2pdfopt to go around the figure more correctly in an automated way. I need some time to investigate.

Great! For my own personal use it is fine already, but I plan to make a present for which I would very gladly use this extra function

Thank you guys again,
m

[mahpooya]

please create an option like -neg but without images inverse
only texts and background inverse