Extracting content from encrypted file

[praveentiru]

Please ensure the following:

• I built executable using the latest code (git clone followed by go install)
• I am using Windows 11
• Bug detail described below:
  • I was trying to build a tool to extract financial transactions from Indian custodian for all mutual fund investments
  • The download is encrypted with a user password which I know
  • I used the user password to decrypt the file using pdfcpu
  • Then I extracted the content using pdfcpu in pdfcpu extract -mode content xxx.pdf path
  • Lot of strings in extracted content are non-ascii characters where it seems they are still encrypted

Is this an issue because I do not have access to owner password? Is there a way to work around this issue?

[hhrutter]

Hi there!

This is because you are looking at raw PDF content.
Extracting text ouf of page content is not supported at this time.

[mengstr]

The best/easiest solution I found was to use Ghostscript to extract the text from PDFs. When running on production/staging servers I can call Ghostscript directly, but on my development Macbook the Ghostscript installed from homebrew fails so I have a small docker container with Ghostscript to do the extraction there.

var outb []byte
if os.Getenv("devMode") != "yes" {
	outb, err = exec.Command("gs", "-sDEVICE=txtwrite", "-sOutputFile=-", "-q", "-dNOPAUSE", "-dBATCH", filename).Output()
} else {
	outb, err = exec.Command("docker", "run",
		"--rm",
		"-v", "/Users/mats/Desktop/HealthManager/hm-storage/labtests/:/app",
		"-w", "/app",
		"minidocks/ghostscript",
		"-sDEVICE=txtwrite", "-sOutputFile=-", "-q", "-dNOPAUSE", "-dBATCH",
		filepath.Base(filename),
	).Output()
}
out := string(outb)

[praveentiru]

Ok. I have found that after decrypting even Apache Tika was able to extract text properly without issue.