Way to test PDF to see if there is any text?

[spedinfargo]

I'm on Windows and am using ocrmypdf under WSL with some great success.

I've been playing around with the --sidecar option to see if I can "test" a PDF for ocr-ness. I'm doing this in my Powershell script using PdfToText.exe similar to this:

foreach ($pdffile in get-childitem -recurse -filter *.pdf){
    $pdftext=invoke-expression ("C:\Temp\Xpdftools\pdftotext.exe -q '"+$pdffile.fullname+"' -");
    if ($pdftext -ne $null) {
        if ($pdftext.ToCharArray().length -lt 20){

That works very well - I can then run the ocrmypdf command on the files that match:

wsl ocrmypdf '/mnt/c/OneDrive/scan0014.pdf' '/mnt/c/OneDrive/scan0014.pdf' --output-type pdf

However, it would be nice if I could skip the installation of pdftotext and just use ocrmypdf for everything. Is there any way to do this or would it be a feature request?

[jbarlow83]

You can use ocrmypdf's API to obtain the text, which mainly relies on pdfminer.six. It's in the ocrmypdf.pdfinfo module. There are no command line bindings to access it, however.