Problem
When uploading a PDF that shall be vectorized, it is inconvenient to take the amount of text in the PDF into account as "marker" - as it requires the uploader to achieve this reliably

Solution

• based on feature https://jira.translate5.net/browse/TRANSLATE-5526 we will have a way to have workflow-step dependant params/configs
• if such a config is set, the pdfToHtml conversion step will be configured with the additional config "vectorizePdfs=1" for the container
• this leads to a vectorization of the pdfs in the pdfconverter with ghostscript (which can fail! -> error-handling is crucial)
• the pdf2htmlEX extraction in the pdfconverter also will extract a SVG instead of a JPEG (based on the new container-config "vectorizePdfs")
• the rest of the process (printing of visual, email-generation for approval) will work like it is currently

FRONTEND TESTS

Testdata is in nextcloud/Developer/test-data/BY-ISSUE (hidden from AI)/TRANSLATE-5527