Once integration is achieved, this is the next critical step. Tika is waiting for this: https://issues.apache.org/jira/browse/TIKA-1727
This is done at https://github.com/virtuald/poi/tree/xdgf, ready for initial review.
In particular, the text extractor can be found at https://github.com/virtuald/poi/commit/e871eea8f9ee14b59052133cf21f89a5902accf3
Looks good on the overall changes, for the Visio specifics I cannot say much, unfortunately.
Good enough for now, thanks for looking! Will merge into trunk with everything else after the release.
I merged this in with the rest of the XDGF changes in revision 1709361, forgot to mark this done.