filter git repo to only files interested in by regex

filter_reduce.rb

#!/usr/bin/env ruby

# filter git repo to only files interested in by regex
#
# Get all files ever in the repo.
# git log --pretty=format: --name-only --diff-filter=A | sort -u > list.txt
# ruby filter_reduce.rb | tr '\r\n' ' ' > reduced_list.txt
# git filter-branch --index-filter "git rm -rfq --cached --ignore-unmatch $(cat reduced_list.txt)" --prune-empty

list = []
File.open("#{__dir__}/list.txt").each_line do |line|
  list << line.chop
end

$stderr.puts " Total: #{list.length}"

# @todo Make this an arg... or not.
keep = list.grep(/(ah-task-server\.rb|ah-run-site-hook\.rb|task-server-config\.json|taskserver\/files\/config\.json|[\w-]+\.task\.\w+)$/)
if keep.empty?
  raise "Nothing to keep"
end

$stderr.puts "  Keep: #{keep.length}"

# Files to delete.
delete = []
list.each do |line|
  last_line = nil
  loop do
    # We reduce to the most general path that doesn't delete things we like, otherwise too many arguments.
    if keep.grep(/#{line}/).empty?
      last_line = line.clone
      line = File.dirname(line)
    else
      break
    end
  end
  if last_line
    delete << last_line
  end
end

delete.uniq!
$stderr.puts "Delete: #{delete.length}"
delete.each { |line| puts line }