Problem

Task import fails when visual/office conversion is involved because of comma in a filename, e.g. when Support-, Wartungs- und Hostingvertrag_benutzerbasiert_final.zip import archive contains

workfiles/Support-, Wartungs- und Hostingvertrag_benutzerbasiert_final.docx

It is supplied in t5visual-conversion.json as 2 files:

"Support-" and " Wartungs- und Hostingvertrag_benutzerbasiert_final.docx":

t5visual-conversion.json
{"files":[{"file":"Support-","type":"","pdf":"","error":""},{"file":" Wartungs- und Hostingvertrag_benutzerbasiert_final-.docx","type":"docx","pdf":"","error":""}],"timestamp":1764237442,"step":"office-to-pdfs"} 

PDFs have the same Problem.

Solution

In our Model (LEK_visualreview_files.source, {}LEK_visualreview_files.officeSource) the files are stored comma-separated in the database. Evaluate two solutions:

a) moving to a JSON-column
b) using special char (e.g. slash) instead comma as separator

JSON format was chosen for 2+ filenames storing