k2pdfopt: optimizes PDFs for viewing on e-readers

[willus]

Originally Posted by atoq

I'm using k2pdfopt 2.34b for converting pdfs to be read on Nook.
The conversion is fine, but when a line contains an equation, and ends
on equal sign, =, for being continued on the next line, the equal sign
gets removed in the converted pdf.

Try putting -hy- in the "Additional Options" box. It will turn off hyphen detection. The equals signs may be getting treated as hyphens? If that doesn't work, please post an example from your source file and list all of the command-line options you are using (or a screen shot of the GUI).

[dhdurgee]

I have been using a recommended command from you to optimize a magazine for reading on my kindle. There are a few cases where it fails, so perhaps my command needs further tweaking. The current command line is:

k2pdfopt -bpc 1 -cbox1 0,.52in -cbox2-9999 0,0 -m .11in,.04in,.14in,.6in -x <input PDF files>

I have attached a page that the above command line fails to detect that the upper portion is in two columns, resulting in giberish in the output file. Perhaps this is unavoidable, as not all pages are in the same layout, and pages like this will need to be handled as individual cases. But I defer to your superior knowledge of your tool to address the case that it might be possible to tweak the column detection in the command line.

Dave

PS: In case it makes a difference, this was page 3 in the PDF being processed

[willus]

Originally Posted by dhdurgee

I have been using a recommended command from you to optimize a magazine for reading on my kindle. There are a few cases where it fails, so perhaps my command needs further tweaking. The current command line is:

k2pdfopt -bpc 1 -cbox1 0,.52in -cbox2-9999 0,0 -m .11in,.04in,.14in,.6in -x <input PDF files>

I have attached a page that the above command line fails to detect that the upper portion is in two columns, resulting in giberish in the output file. Perhaps this is unavoidable, as not all pages are in the same layout, and pages like this will need to be handled as individual cases. But I defer to your superior knowledge of your tool to address the case that it might be possible to tweak the column detection in the command line.

This seems to solve it--add this to the command line (or "Additional Options" box in the MS Windows GUI):

-ch 1

That changes the minimum height of a detectable 2-column region from 1.5 inches to 1 inch. Your 2-column region was just a little shorter than the default for being detected as 2-columns.

[dhdurgee]

Originally Posted by willus

This seems to solve it--add this to the command line (or "Additional Options" box in the MS Windows GUI):

-ch 1

That changes the minimum height of a detectable 2-column region from 1.5 inches to 1 inch. Your 2-column region was just a little shorter than the default for being detected as 2-columns.

Thank you for your quick response. This does fix the problem and I will incorporate it in my command line for future processing.

Dave

[dhdurgee]

Originally Posted by dhdurgee

Thank you for your quick response. This does fix the problem and I will incorporate it in my command line for future processing.

Dave

I just processed a new issue with the revised command line and while it dealt well with most pages there were four pages with problems. I have attached these pages for your inspection in case the command line can be tweaked a bit more without causing problems elsewhere.

The command line as I used it with these pages was:

k2pdfopt -bpc 1 -cbox1 0,.52in -cbox2-9999 0,0 -m .11in,.04in,.14in,.6in -x -ch 1

Of the pages only one was a first page in a pdf file. You can distinguish that one by the extra information at the top above the page image.

Thank you again for your assistance and creating this tool.

Dave

[willus]

Originally Posted by dhdurgee

I just processed a new issue with the revised command line and while it dealt well with most pages there were four pages with problems. I have attached these pages for your inspection in case the command line can be tweaked a bit more without causing problems elsewhere.

The command line as I used it with these pages was:

k2pdfopt -bpc 1 -cbox1 0,.52in -cbox2-9999 0,0 -m .11in,.04in,.14in,.6in -x -ch 1

Of the pages only one was a first page in a pdf file. You can distinguish that one by the extra information at the top above the page image.

Thank you again for your assistance and creating this tool.

Dave

Try -ch 0.5 (instead of -ch 1) and also adding: -evl 2
That will erase the verticle edges of the box around the boxed text on the one page and allow it to be re-flowed. The -ch 0.5 will break apart the shorter columns. You still have trouble with the first page because k2pdfopt is thinking that the text immediately after the black image is a caption for it and so doesn't break it into the separate columns. I don't have a control for this at the present time.

[umeshg]

I am using Windows 10 and I have tried all 64 Bit version but I am getting the error message when I select the ' convert' option

Can someone please assist

Thank you

[willus]

Originally Posted by umeshg

I am using Windows 10 and I have tried all 64 Bit version but I am getting the error message when I select the ' convert' option

Can someone please assist

Thank you

Can you post a screen shot of your GUI before you start the conversion? It's a long shot, but are you certain that (1) you have enough disk space for the conversion, and (2) you are saving into a folder where you have write permissions?

[Masacroso]

Hi. First: thank you very much willus! Your application is awesome, I converted some PDF mathematical texts with this and I can read it in my ereader.

But I must says that I tried to convert somes djvu but the program dont read the files, when I press convert the log screen show that the program cant read the djvu files (but I can read it using windjview or any other djvu viewer).

Im using the windows 64 bit version of the program (the last one, the 2.34b) over windows 7. Ijust need to report that, maybe Im doing something wrong. I tried to convert in the default settings and some others but no one works.

[willus]

Originally Posted by Masacroso

But I must says that I tried to convert somes djvu but the program dont read the files, when I press convert the log screen show that the program cant read the djvu files (but I can read it using windjview or any other djvu viewer).

Im using the windows 64 bit version of the program (the last one, the 2.34b) over windows 7...

Can you post or PM me an example DJVU file which does not convert correctly?

[Masacroso]

Originally Posted by willus

Can you post or PM me an example DJVU file which does not convert correctly?

Yes. I must say that it doesnt convert at all, the program just cant read the file.

I attached two files that I tried convert.

Anyway I must say that this problem is minor, I can export the djvu files to pdf and after make the conversion. Anyway it will be interesting to check what happen about djvu conversion.

[Masacroso]

Ok, I solved the problem: it seems that the program dont open files with spaces in it name, so just making a copy and changing the name to "aaa.djvu" it works.

The weird thing is that this doesnt happen when I choose a pdf file with spaces in it name.

[willus]

Originally Posted by Masacroso

Ok, I solved the problem: it seems that the program dont open files with spaces in it name, so just making a copy and changing the name to "aaa.djvu" it works.

The weird thing is that this doesnt happen when I choose a pdf file with spaces in it name.

Strange. I had no issues, even with the original file names that had spaces. Did you use the GUI or the command line? If you used the command line, you would have had to put double-quotes around the file names if they had spaces in them. Otherwise, is the file "aaa.djvu" in the same folder as the other files? Does the full path of the file name contain any non-ASCII characters (e.g. special European or foreign characters)? If you used the command line, can you send the exact command that failed, or if you used the GUI, can you post a screen shot?

I also notice that on the Kurtz file, k2pdfopt claims an error converting the blank pages. Not a big deal, but I'll check that before the next release.

[Masacroso]

Originally Posted by willus

Strange. I had no issues, even with the original file names that had spaces. Did you use the GUI or the command line? If you used the command line, you would have had to put double-quotes around the file names if they had spaces in them. Otherwise, is the file "aaa.djvu" in the same folder as the other files? Does the full path of the file name contain any non-ASCII characters (e.g. special European or foreign characters)? If you used the command line, can you send the exact command that failed, or if you used the GUI, can you post a screen shot?

I also notice that on the Kurtz file, k2pdfopt claims an error converting the blank pages. Not a big deal, but I'll check that before the next release.

I tested: it is the path, yes... I make a copy of a djvu file somewhere, without changing it name, and it was converted with no problem. The weird thing is that I converted some pdf from the same folder and it doesnt had any problem. Only when I try to convert djvu the path problem appear.

The name of the folder have some non-common character. I will take a screenshot using the GUI:

http://i.imgur.com/5Kb1Z58.png

[willus]

Originally Posted by Masacroso

I tested: it is the path, yes... I make a copy of a djvu file somewhere, without changing it name, and it was converted with no problem. The weird thing is that I converted some pdf from the same folder and it doesnt had any problem. Only when I try to convert djvu the path problem appear.

The name of the folder have some non-common character. I will take a screenshot using the GUI:

http://i.imgur.com/5Kb1Z58.png

The PDF file handling routines support UTF-8 strings, whereas the DJVU routines do not (yet). So this is consistent. Thank you for checking.