[ https://issues.apache.org/jira/browse/TIKA-4015?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17711544#comment-17711544 ]
Tim Allison commented on TIKA-4015: ----------------------------------- This is a start on font-based charsets in ooxml: [https://c-rex.net/projects/samples/ooxml/e1/Part4/OOXML_P4_DOCX_charset_topic_ID0EH2CU.html] This tells us that we should map the "abc" to abc > Extract symbols as symbols from .docx > ------------------------------------- > > Key: TIKA-4015 > URL: https://issues.apache.org/jira/browse/TIKA-4015 > Project: Tika > Issue Type: New Feature > Reporter: Tim Allison > Priority: Minor > Attachments: symbol.docx.zip > > > [~chetab] raised this issue on the user list.and supplied an example document. > The Font is symbol and the text should be: abcedefghijklmnopqrstuvwxyz > However, the text as literally stored in the docx and extracted by Tika is: > abcedefghijklmnopqrstuvwxyz > > We may need to add processing for unicode mappings or the equivalent in > ooxml. I haven't seen this before. :P -- This message was sent by Atlassian Jira (v8.20.10#820010)