Add aggressive mode to CCIAnalyzer

[MER-C]

CCIAnalyzer could do with a second pass mode that, after looking at the original CCI, removes:

• Template arguments
• References
• Comments

Other things that can be removed should be investigated.

[MER-C]

Disambig pages and lists should also go.

[MER-C]

Also consider upping the word threshold to 15 or something.

[MER-C]

• 634eea7 implemented a variable word threshold and removal of list items and references.
• a52526b added a table start culling function and stopped counting "words" consisting of only punctuation characters against the word count limit.
• a226cc7 adds preprocessing functions - included so far are comment and external link removal
• 64feb0f adds culling that acts on page titles, added are functions for disambiguation and list pages

Further ideas, should they become necessary in the future:

• a blacklist that makes sure that all edits that touch a "plot", "synopsis" or "lyrics" section are flagged as major
• filters on full diff content (needs machine readable diffs?)
• safe table row culling