A histogram view on contributor statistics

developer/bin/the_long_tail

@@ -0,0 +1,241 @@
+#!/usr/bin/env ruby
+#
+# the_long_tail
+#
+# A histogram view on contributor stats
+#
+# notes
+#
+# Since this script does not track file-renames in the git history, the
+# dependence of Casks upon occasional contributors/non-maintainers can
+# only be expressed as a range or lower bound.
+#
+
+###
+### dependencies
+###
+
+require 'open3'
+require 'set'
+
+###
+### configurable constants
+###
+
+BINS = [
+        (1..10).to_a,
+        100,
+        1000,
+       ].flatten
+
+OCCASIONAL_CUTOFF = 5
+
+CASK_PATH = 'Casks'
+
+MAINTAINERS = %w[
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                  [email protected]
+                 ]
+
+###
+### git methods
+###
+
+def cd_to_project_root
+  Dir.chdir File.dirname(File.expand_path(__FILE__))
+  @git_root ||= Open3.popen3(*%w[
+                                 git rev-parse --show-toplevel
+                                ]) do |stdin, stdout, stderr|
+    begin
+      stdout.gets.chomp
+    rescue
+    end
+  end
+  Dir.chdir @git_root
+  @git_root
+end
+
+def authors
+  @authors ||= Open3.popen3(*%w[
+                                git log --no-merges --format=%ae --
+                               ]) do |stdin, stdout, stderr|
+    h = {}
+    stdout.each_line do |line|
+      line.chomp!
+      h[line] ||= 0
+      h[line]  += 1
+    end
+    h
+  end
+end
+
+def casks_by_author
+  @casks_by_author ||= Open3.popen3(*%w[
+                                        git log --no-merges --name-only --format=%ae --
+                                       ],
+                                    CASK_PATH) do |stdin, stdout, stderr|
+    email = nil
+    h = {}
+    stdout.each_line.to_a.join('').split("\n\n").each do |paragraph|
+      if paragraph.include?('Casks/')
+        lines=paragraph.split("\n")
+        email = lines.pop
+        h[email] ||= Set.new
+        h[email].merge(lines.compact)
+      else
+        email = paragraph.chomp
+      end
+    end
+    h
+   end
+end
+
+###
+### filesystem methods
+###
+
+def all_casks
+  @all_casks ||= Open3.popen3('/usr/bin/find',
+                               CASK_PATH,
+                               *%w[-type f -name *.rb]
+                              ) do |stdin, stdout, stderr|
+    stdout.each_line.map(&:chomp)
+  end
+end
+
+###
+### analysis and report methods
+###
+
+def histogram
+  if @histogram.nil?
+    @histogram = Hash[*BINS.map{ |elt| [elt, 0] }.flatten]
+    authors.each do |name, num_commits|
+      bottom = 0
+      BINS.each do |top|
+        if num_commits >= bottom and num_commits < top
+          @histogram[bottom] += 1
+        end
+        bottom = top
+      end
+    end
+  end
+  @histogram
+end
+
+def historic_occasional_cask_set
+  @historic_occasional_cask_set = authors.each.collect do |name, num_commits|
+    if num_commits > OCCASIONAL_CUTOFF
+      nil
+    elsif ! casks_by_author.key?(name)
+      nil
+    else
+      casks_by_author[name].to_a
+    end
+  end.flatten.compact.to_set
+end
+
+def extant_occasional_cask_count
+  # avoid double-counting renames by intersecting with extant Casks
+  historic_occasional_cask_set.intersection(all_casks).count
+end
+
+def historic_nonmaintainer_cask_set
+  @historic_nonmaintainer_cask_set = authors.each.collect do |name, num_commits|
+    if MAINTAINERS.include?(name)
+      nil
+    else
+      casks_by_author[name].to_a
+    end
+  end.flatten.compact.to_set
+end
+
+def extant_nonmaintainer_cask_count
+  # avoid double-counting renames by intersecting with extant Casks
+  historic_nonmaintainer_cask_set.intersection(all_casks).count
+end
+
+def extant_occasional_cask_percentage
+  @extant_occasional_cask_percentage ||= (100 * extant_occasional_cask_count / all_casks.count).to_i
+end
+
+def historic_occasional_cask_percentage
+  @historic_occasional_cask_percentage ||= (100 * historic_occasional_cask_set.count / all_casks.count).to_i
+end
+
+def extant_nonmaintainer_cask_percentage
+  @extant_nonmaintainer_cask_percentage ||= (100 * extant_nonmaintainer_cask_count / all_casks.count).to_i
+end
+
+def historic_nonmaintainer_cask_percentage
+  # this is so large, it might cross 100%
+  @historic_nonmaintainer_cask_percentage ||= [100, (100 * historic_nonmaintainer_cask_set.count / all_casks.count).to_i].min
+end
+
+def onetime_author_percentage
+  @onetime_author_percentage ||= (100 *
+                                  histogram[1] /
+                                  authors.length).to_i
+end
+
+def occasional_author_percentage
+  # why is it so hard to slice a hash?
+  @occasional_author_percentage ||= (100 *
+                                 (1 .. OCCASIONAL_CUTOFF).to_a.collect{ |bin| histogram[bin] }.reduce(:+) /
+                                 authors.length).to_i
+end
+
+def graph_width
+  if @graph_width.nil?
+    @graph_width  = `/bin/stty size 2>/dev/null`.chomp.split(" ").last.to_i
+    @graph_width  = 80 if @graph_width <= 0
+    @graph_width -= 20 if @graph_width > 20
+  end
+  @graph_width
+end
+
+def graph_normalization
+  @graph_normalization ||= histogram.values.max.to_f
+end
+
+def print_header
+  puts "Commits\tContributors"
+  puts "---------------------"
+end
+
+def print_table
+  BINS.each do |bin|
+    plural = (bin % 10) == 0 ? "'s" : ''
+    graph = '.' * ((histogram[bin]/graph_normalization) * graph_width)
+    puts "#{bin}#{plural}\t#{histogram[bin]}\t#{graph}"
+  end
+end
+
+def print_footer
+  puts %Q[\n#{occasional_author_percentage}% of contributors are "occasional" (with <= #{OCCASIONAL_CUTOFF} commits)]
+  puts "\n#{onetime_author_percentage}% of contributors commit only once"
+  puts "\n#{extant_occasional_cask_percentage}% - #{historic_occasional_cask_percentage}% of Casks depend on an occasional contributor"
+  puts "\n#{extant_nonmaintainer_cask_percentage}% - #{historic_nonmaintainer_cask_percentage}% of Casks depend on a contributor who is not a maintainer"
+  puts "\n"
+end
+
+def generate_report
+  print_header
+  print_table
+  print_footer
+end
+
+###
+### main
+###
+
+cd_to_project_root
+generate_report