reading_length.rb

# -*- coding: utf-8 -*-
require 'httpclient'
require 'nokogiri'

def get text
  endpoint = 'http://jlp.yahooapis.jp/MAService/V1/parse'
  appid = 'RZv4ed6xg67n15cyyGFF8Io0r3i.o0uwISXfrFOZYyMghbdeNA10_M6KemHLqz0laQ--'
  params = {
    :appid => appid,
    :sentence => text,
  }

  client = HTTPClient.new

  Nokogiri (client.get endpoint, params).body
end

class Word
  attr_accessor :surface, :reading, :pos

  def self.new_from_node node
    me = new
    me.surface = (node.at 'surface').text
    me.reading = (node.at 'reading').text
    me.pos     = (node.at 'pos').text
    me
  end

  def is_suffix
    %w{ 連体詞 接続詞 接尾辞 助詞 助動詞 特殊 }.include? pos
  end

  def is_special
    pos == '特殊'
  end

  def reading_length
    return 0 if is_special
    reading.gsub(/[ぁぃぅぇぉゎゃゅょ]/, '').length
  end
end

class Array
  def body
    map{ |item|
      item.surface
    }.join ''
  end

  def reading_length
    inject(0) { |a, b| a + b.reading_length }
  end
end

text = ARGV.first

unless text
  warn "usage: #{$0} (text)"
  exit 1
end


res = get text

state = 0
buffer = []
length_count = Hash.new 0

words = (res.search 'word')

words.each_with_index{ |node, index|
  is_last = index+1 == words.length

  word = Word.new_from_node node

  if state == 1 and not word.is_suffix
    state = 0

    puts "#{buffer.reading_length}\t#{buffer.body}"
    length_count[buffer.reading_length] += 1
    buffer = []

  elsif state == 0 and word.is_suffix
    state = 1
  end

  buffer << word

  if is_last
    puts "#{buffer.reading_length}\t#{buffer.body}"
    length_count[buffer.reading_length] += 1
  end
}

puts

(1 .. length_count.keys.max).each{ |key|
  puts "#{key}\t#{'*' * length_count[key]}"
}

xexample.txt

% ruby reading_length.rb "「この味がいいね」と君が言ったから七月六日はサラダ記念日"
2       「この
3       味が
4       いいね」と
3       君が
5       言ったから
3       七月
4       六日は
7       サラダ記念日

1
2       *
3       ***
4       **
5       *
6
7       *