Description
I'm finalizing the XSLF api in Apache POI and would like to improve the corresponding part on the Tika side.
Current XSLFPowerPointExtractorDecorator is too low-level and doesn't correlate much with the XSLF usermodel. My patch eliminates direct manipulations with low-level XmlBeans and uses high-level usermodel objects instead.
Cheers,
Yegor