clic: Refactor keywords, clusters, non-quotes, and short suspensions

1.4

• Massive refactor to bring keywords to scientific standard
  • Implements a correct Keyword Extraction Algorithm (KEA)
  • Note: Earlier results should be distrusted
  • Tests KEA to check results are identical to match results set by Paul Rayson
  • Documents use and origin of the KEA
• Refactors the clusters
• Updates User Interface to increase modularity and add total counts
  • Makes clusters and keywords tables 100% wide to fix word wrap bug
  • Renames js and GET variables in clusters and keywords to more sane defaults
  • Adds right values for the keywords table (use, p_value, LL)
  • Enables sorting of keywords and clusters on rank
  • Updates book modal dialogue with latest corpus counts
  • Clusters and keywords now include total cluster count
  • Implements descriptions of the corpora based on the user's input
  • Provides a summary of the search settings before giving clusters and keywords tables
• Starts building help pages, About -> Help
• Fixes API code that was unreadable
• Fixes bug that returned 1grams when 5grams were asked
• Enables caching for keywords and clusters
  • Updates beaker cache version to handle args
• Fixes non-quote counts that missed non-quotes before the first quote
  • Correct mistake (before my time) in create_bookcountsjson
  • Update book counts now including total counts/corpus
  • Add note about the wrong q+nq counts
  • Adds selector with help from Rob, Cat, and Matt, but does not use it yet
• Adds consultant to the home page
• Implements concordance searches in short suspensions
• Implements 3 and 4 grams in short suspensions (both keywords and clusters)