How did I create PDF from the Scans sent over By Email
First off, the scanned documents that I received from my email were split into
many documents. Let’s square this off right away. Combine multiple pdf documents
into a single document using the
pdfunite command like so:
pdfunite sample_* single.pdf
On Ubuntu the
pdfunite command is part of the poppler-utils package. If you
are on Ubuntu, you can install the package using the
sudo apt-get install poppler-utils
Next up, I wanted to create a bookmarked table of contents for the pdf. Doing this manually would have taken me a long time. The idea of spending my lazy afternoon transcribing pdf documents did not, to put it mildly, excite me. My search led me to tesseract, an OCR program that does a wonderful job of extracting text from images.
As soon as you decide to use tesseract, you hit another roadblock.
Tesseract works on
.tiff format and most of the examples from the documentation
were in tiff format. I used Imagemagick’s
convert command to convert pdf to tiff.
I installed imagemagick and ran the convert command only to find myself in another
I ran into this problem.
unable to create temporary file
/some/pathPermission denied @ error/pdf.c/ReadPDFImage/465
On further digging, I realized that the convert command used Ghostscript under the hood. A security update prevented this access. I fixed it by going through this link: https://alexvanderbist.com/posts/2018/fixing-imagick-error-unauthorized
I got the document converted to tiff and from there I could extract the
text using Tesseract. Lastly, I had to do manual labor to get the
bookmarks and page number in correct order but extracting the text from
the pdf document gave me a good lead. Lastly, I updated the bookmarks
Here are the steps in short:
If you have multiple documents, then combine them using the
Convert the document in
.tiffformat using the
convertcommand from Imagemagick.
tesseractcommand to generate text from your tiff files.
Generate the table of contents using the format support by the
pdftkcommand. The extracted text will give you some idea.
Generate the bookmarks and update the documents using the
update_infosubcommand from the
You can also try updating the index of the document now that you have the text inside it, but that’s for later.