很早之前实现的，不知道有没有更好的支持同义词映射的词频统计方案。

tags = YAML.load_file('./tag_synonyms.yml')

 text = "this is ruby run ubuntu is github ruby, ror is rails rails4 rails5 ror3"

TagClassifier.new(text).classify(tags)

# =>
{
    "ruby on rails" => 5,
             "ruby" => 2,
           "github" => 1,
           "ubuntu" => 1
}

• 同义词映射表 tag_synonyms.yml

  ---
  - ubuntu
  - centos
  - redhat
  - ruby on rails:
  - ror
  - rails
  - mysql
  - redis
  - ruby
  - java
  - python
  - foreman
  - pry
  - apache
  - nginx
  - node.js:
  - node-js
  - nodejs
  - openssh
  - debian
  - docker
  - github
  - amazon
  

• 过滤器 tag_classifier.rb

  # -*- encoding: utf-8 -*-
  

class TagClassifier

def initialize(text) @text = text @words = {} end

def classify(tags, name=nil) return(@words) unless @text.to_s.size >= 3

tags.map do |tag| if tag.is_a?(Hash) term = tag.flatten(2) tag = term[0] end

count = scan(term || tag)

if count > 0 keyword = name || tag @words[keyword] = count end end return Hash[@words.sort_by { |tag, rank| rank }.reverse] unless @words.empty? @words end alias :split :classify

private def scan(term) term = term.join('|') if term.is_a?(Array) @text.scan(/#{term}/i).count end end