I was just looking at SolrCellBuilder, and it looks like there's an assumption that documents will not have attachments/embedded objects. Unless I misunderstand the code, users will not be able to search documents inside zips, or attachments in msg/ doc/pdf/etc (cf. SOLR-7189).
Are embedded documents extracted in a step before hitting SolrCellBuilder? Bug or feature? Thank you! Cheers, Tim