删除JSONP以在HTTParty中解析

使用自定义解析器.在这个例子中,我使用子字符串去除填充,但如果您愿意,可以使用正则表达式:

require 'httparty'

class JsonpParser < HTTParty::Parser
  SupportedFormats = {"text/javascript" => :jsonp}

  def jsonp
    JSON.load(body[2..-11], nil)
  end

end

class Dictionary
  include HTTParty

  parser JsonpParser

  def self.define(word)
    get "http://www.google.com/dictionary/json?callback=a&sl=en&tl=en&q=#{word}"
  end

end

Dictionary.define 'epitome'

这很容易做到,除了常规的Ruby以及它附带的JSON和OpenURI类之外什么都没有.考虑一下:

require 'open-uri'
require 'json'

URL = 'http://www.google.com/dictionary/json?callback=a&sl=en&tl=en&q=epitome'
jsonp = open(URL).read

此时,jsonp包含来自Google API的响应.你想要的部分用花括号包裹{...}.

正则表达式是贪婪的,并且会吞噬它们所能做的一切.通常这是一个问题,但在这种情况下它是好的.一个简单的模式/{.+}/将找到从第一个{到最后一个的所有内容},即JSON有效负载,此时,您已准备好解析:

payload = jsonp[/{.+}/]
# => "{\"query\":\"epitome\",\"sourceLanguage\":\"en\",\"targetLanguage\":\"en\",\"primaries\":[{\"type\":\"headword\",\"terms\":[{\"type\":\"text\",\"text\":\"e·pit·o·me\",\"language\":\"en\",\"labels\":[{\"text\":\"Noun\",\"title\":\"Part-of-speech\"}]},{\"type\":\"phonetic\",\"text\":\"/i?pit?m?/\",\"language\":\"und\"},{\"type\":\"sound\",\"text\":\"http://www.gstatic.com/dictionary/static/sounds/de/0/epitome.mp3\",\"language\":\"und\"}],\"entries\":[{\"type\":\"related\",\"terms\":[{\"type\":\"text\",\"text\":\"epitomes\",\"language\":\"und\",\"labels\":[{\"text\":\"plural\"}]}]},{\"type\":\"meaning\",\"terms\":[{\"type\":\"text\",\"text\":\"A person or thing that is a perfect example of a particular quality or type\",\"language\":\"en\"}],\"entries\":[{\"type\":\"example\",\"terms\":[{\"type\":\"text\",\"text\":\"she looked the \\x3cem\\x3eepitome\\x3c/em\\x3e of elegance and good taste\",\"language\":\"en\"}]}]},{\"type\":\"meaning\",\"terms\":[{\"type\":\"text\",\"text\":\"A summary of a written work; an abstract\",\"language\":\"en\"}]},{\"type\":\"meaning\",\"terms\":[{\"type\":\"text\",\"text\":\"A thing representing something else in miniature\",\"language\":\"en\"}]}]}]}"

解析:

data = JSON.parse(payload)
# => {"query"=>"epitome",
#     "sourceLanguage"=>"en",
#     "targetLanguage"=>"en",
#     "primaries"=>
#      [{"type"=>"headword",
#        "terms"=>
#         [{"type"=>"text",
#           "text"=>"e·pit·o·me",
#           "language"=>"en",
#           "labels"=>[{"text"=>"Noun", "title"=>"Part-of-speech"}]},
#          {"type"=>"phonetic", "text"=>"/i?pit?m?/", "language"=>"und"},
#          {"type"=>"sound",
#           "text"=>
#            "http://www.gstatic.com/dictionary/static/sounds/de/0/epitome.mp3",
#           "language"=>"und"}],
#        "entries"=>
#         [{"type"=>"related",
#           "terms"=>
#            [{"type"=>"text",
#              "text"=>"epitomes",
#              "language"=>"und",
#              "labels"=>[{"text"=>"plural"}]}]},
#          {"type"=>"meaning",
#           "terms"=>
#            [{"type"=>"text",
#              "text"=>
#               "A person or thing that is a perfect example of a particular quality or type",
#              "language"=>"en"}],
#           "entries"=>
#            [{"type"=>"example",
#              "terms"=>
#               [{"type"=>"text",
#                 "text"=>
#                  "she looked the x3cemx3eepitomex3c/emx3e of elegance and good taste",
#                 "language"=>"en"}]}]},
#          {"type"=>"meaning",
#           "terms"=>
#            [{"type"=>"text",
#              "text"=>"A summary of a written work; an abstract",
#              "language"=>"en"}]},
#          {"type"=>"meaning",
#           "terms"=>
#            [{"type"=>"text",
#              "text"=>"A thing representing something else in miniature",
#              "language"=>"en"}]}]}]}

现在你已经有了一个很好的ol'Ruby hash,你可以正常访问:

data['query'] # => "epitome"

总结一下:

require 'open-uri'
require 'json'

URL = 'http://www.google.com/dictionary/json?callback=a&sl=en&tl=en&q=epitome'

data = JSON.parse(open(URL).read[/{.+}/])

data['query'] # => "epitome"
data['primaries'].size # => 1

而且,它也适用于常规JSON结果,因此您无需执行任何特殊操作即可.