I find keywords extraction is a big selling point for DTP. It works pretty well for English. However, the analysis results for Chinese texts do not make much sense to me as a Chinese native speaker as Chinese (and Japanese) needs to be segmented (or tokenized) first before any further analysis. You may consider using specified algorithm or scripts (such as Pynlpir or Jieba) for processing Chinese (or Japanese) doc.
Thanks for the heads-up. No promises, of course, but we will take a look at these things. Cheers!