Quote:
Originally Posted by dgvirtual
I have a problem with a three-column text. Inside one page an image is located in a way to overlap the two columns. And the program does not read the page correctly - it does not recognize the text as three column. Now, it renders the second page fine. Now, if I cut the image out by specifying 3,4" bottom margin, the columns get recognized (although the lines separating the columns does not get ignored, which is a mino problem, though).
could something be done about pages like this, or is it just too much play? Here is the file I have a problem with:
http://www.nzidinys.lt/files/various...iene%20txt.pdf |
Layouts like this make me want to re-think the way I order regions in k2pdfopt, or at least to provide a couple more options, but I was able to get something reasonable with the existing version:
k2pdfopt -col 4 -cgr .4 -evl 1 -sm -mb 1.1 -ch 0.5 Az.pdf
-col 4 enables detection of up to 4 columns (2 levels of recursion).
-cgr .4 limits the horizontal search range for the column divider. The value of .4 gets k2pdfopt to treat the left column divider as the first divider, which is the key to correct layout on page 1.
-evl 1 erases the vertical lines, which helps k2pdfopt find the column dividers.
-sm shows you how k2pdfopt is flowing your document (in the
..._marked.pdf file). You can take that out on the final conversion since it slows things down considerably.
-mb 1.1 ignores the page numbers / footer on the bottom of each page by cropping off the bottom 1.1 inches from each source page.
-ch 0.5 allows regions as short as 0.5 inches in height to be separated into multiple columns, which is important for page 1 (the default is 1.5 inches).