Alternate cleaned up pipeline ready for integration

[alecristia]

This version of the pipeline includes:

• code optimization for running time (corpus extraction done once), improved dealing with foreign words
• final version of cleaning to increase comparability across morphemes and words
• final version of phonologization with clearer rule annotation
• final version of cutting, occurring after phonologization
• no real changes to analyses
• simpler results collapse
• analyses in knittable version

Remain to be sorted:

• code for corpus extraction works in mac but not linux
• cutting does not match across Chintang and Japanese for corpus length
• analyses running so there could be bugs not yet spotted