vacuum_zsh_history.rb

#!/usr/bin/env ruby

require 'optparse'
require 'logger'

HISTORY_PATTERN       = /\A: (?<started>\d+):(?<finished>\d+);(?<text>.+)\z/o.freeze
IGNORE_ENTRY_PATTERN  = /\A(?:l[sla]?|man|git st\z)/o.freeze
REQUIRED_OPTION_NAMES = [:output_file, :input_file, :lines_to_read]

Config = Struct.new(:output_file, :input_file, :lines_to_read).new

def logger
  @logger ||= Logger.new($stderr)
end

# maybe_matched:Maybe[MatchData] => entry:Hash
def ensure_history_entry(maybe_matched)
  maybe_matched ? extract_matched_data(maybe_matched) : {}
end

# matched:MatchData => data:Hash
def extract_matched_data(matched)
  matched.names.each_with_object({}) {|name, hash| hash[name] = matched[name] }
end

# line:String => entry:Hash
def extract_entry(line)
  matched = line.match(HISTORY_PATTERN) rescue nil # see also Src/zsh.h
  entry = ensure_history_entry(matched)
  entry.merge('line' => line)
end

# entry:Hash => line:String
def serialize_entry(entry)
  ': %d:%d;%s' % entry.values_at('started', 'finished', 'text')
rescue => e
  logger.error("fail: #{entry['line']}; because: #{e}")
  ''
end

def fetch_lines(enum, n)
  n < 0 ? enum : enum.take(n)
end

option_parser = OptionParser.new do |o|
  o.on('-o', '--output-file OUTPUT_FILE') {|v| Config.output_file = v }
  o.on('-i', '--input-file INPUT_FILE') {|v| Config.input_file = v }
  o.on('-n', '--lines-to-read N') {|v| Config.lines_to_read = v.to_i }
end

option_parser.parse!(ARGV)

REQUIRED_OPTION_NAMES.each do |name|
  option_parser.abort("Missing argument: #{name}") unless Config.send(name)
end

lines = fetch_lines(IO.foreach(Config.input_file), Config.lines_to_read)
out = lines.lazy.
  map {|line| extract_entry(line.strip) }.
  reject {|entry| IGNORE_ENTRY_PATTERN === entry['text'] }.
  force.
  uniq {|entry| entry['text'] }.
  sort_by {|entry| entry['started'].to_i }.
  map {|entry| serialize_entry(entry) }

open(Config.output_file, 'w') do |f|
  f.puts(*out)
end