I am using unoconv (https://github.com/dagwieers/unoconv) to convert DOCX and DOC file to PDF, but will often get strange results on certain characters when they are rendered in the PDF.
One particular problem is numbers translating oddly for example, the section label of:
Section 2.3 (http://note.io/1Q33RX6)
Get's turned into a roman numeral:
Section II.3 (http://note.io/1b6MDs5)
I have a feeling this has to do with installed character sets but have no idea how to debug it.
The setting for the issue is a Django app making call to a unix shell script to convert a document on disk.