Write a script that prepares the data end-to-end

Which means:

[x] Extract conversation, actor, dialogue entry data from the dialogue bundle (#6)
[x] Use the extracted data to label the audio clips extracted from the game asset files (#2)
[ ] Use the audio clip table above to create files used for training the voder in the format used by Kaldi (#7)
[ ] Combine all of the above into a single mix task (#11)

ngtban / wavenet_de_data_prep