├── Gemfile
├── lib
    ├── bosonnlp
    │   ├── version.rb
    │   ├── mixin.rb
    │   └── util.rb
    └── bosonnlp.rb
├── .gitignore
├── bosonnlp.gemspec
├── examples
    ├── usage.rb
    └── usage.rb,
└── README.md


/Gemfile:
--------------------------------------------------------------------------------
1 | source "https://rubygems.org"
2 | 
3 | gem "httpclient"
4 | 


--------------------------------------------------------------------------------
/lib/bosonnlp/version.rb:
--------------------------------------------------------------------------------
1 | class Bosonnlp
2 |   VERSION = '0.1.1'
3 | end
4 | 


--------------------------------------------------------------------------------
/.gitignore:
--------------------------------------------------------------------------------
 1 | *~
 2 | tags
 3 | pkg
 4 | coverage
 5 | doc
 6 | reports
 7 | Gemfile.lock
 8 | .rbx
 9 | .bundle
10 | *.gem
11 | *.swp
12 | 


--------------------------------------------------------------------------------
/lib/bosonnlp/mixin.rb:
--------------------------------------------------------------------------------
 1 | # -*- coding: utf-8 -*-
 2 | require 'bosonnlp'
 3 | 
 4 | # Mixin module
 5 | module BosonnlpMixin
 6 |   def method_missing(name, *args)
 7 |     case name.to_s
 8 |     when /^(c_|m_|s_)(.+)$/
 9 |       @@nlp ||= Bosonnlp.new
10 |       data = self
11 |       data = [data] if self.class == String
12 | 
13 |       @@nlp.send name, data, *args
14 |     else
15 |       super
16 |     end
17 |   end
18 | end
19 | 
20 | # Mixin it
21 | class Array
22 |   include BosonnlpMixin
23 | end
24 | 
25 | # Mixin it
26 | class String
27 |   include BosonnlpMixin
28 | end
29 | 


--------------------------------------------------------------------------------
/bosonnlp.gemspec:
--------------------------------------------------------------------------------
 1 | require 'rake'
 2 | 
 3 | require_relative 'lib/bosonnlp/version'
 4 | 
 5 | Gem::Specification.new do |s|
 6 |   s.name = 'bosonnlp'
 7 |   s.version = Bosonnlp::VERSION
 8 |   s.date = '2014-10-07'
 9 |   s.authors = ['CC']
10 |   s.email = 'chcoalc@gmail.com'
11 |   s.homepage = 'http://github.com/alal/bosonnlp'
12 |   s.platform = Gem::Platform::RUBY
13 |   s.summary = 'Bosonnlp Ruby SDK'
14 |   s.description = 'bosonnlp.com'
15 |   s.files = FileList['lib/*.rb', 'lib/bosonnlp/*.rb']
16 |   s.require_path = 'lib'
17 |   s.add_runtime_dependency 'httpclient', '~> 2.5', '>= 2.5.0'
18 |   s.license = 'Apache'
19 |   s.required_ruby_version = '>= 1.9'
20 | end
21 | 


--------------------------------------------------------------------------------
/lib/bosonnlp/util.rb:
--------------------------------------------------------------------------------
 1 | class Bosonnlp
 2 |   # 'Utils'
 3 |   module Util
 4 |     def get_hash_args(*args)
 5 |       if Hash === args[-1]
 6 |         h = args[-1]
 7 |       else
 8 |         h = {}
 9 |       end
10 |       h
11 |     end
12 | 
13 |     # :query, :body, :header, :follow_redirect
14 |     def _http_request(method, url, *args)
15 |       request_args = get_hash_args(*args)
16 |       url = get_url url unless url.to_s.start_with?('http://')
17 |       request_args[:header] = @head
18 |       request_args[:body] = \
19 |         JSON.dump(request_args[:body]) if request_args[:body]
20 | 
21 |       res = @http_session.request(method, url, request_args)
22 |       status = res.status
23 |       raise HTTPError, 'HTTPError: %s %s'\
24 |         % [status, (JSON.load(res.content)['message'] rescue res.reason)]\
25 |         if status >= 400 && status < 600
26 |       res
27 |     end
28 |   end
29 | end
30 | 


--------------------------------------------------------------------------------
/examples/usage.rb:
--------------------------------------------------------------------------------
 1 | require 'bosonnlp'
 2 | 
 3 | # run `export BOSON_API_TOKEN="<your token from bosonnlp.com>"` first. e,g,
 4 | #     `export BOSON_API_TOKEN="bmw.1999.eW91ciB0b2tl"`
 5 | 
 6 | # -------------- string and list style ----------------
 7 | p '美好'.c_sentiment
 8 | p ['美好', '悲惨'].c_sentiment
 9 | 
10 | p ['很好吃, 有点贵', '贵了点, 好吃',
11 |    '价格贵, 但很好吃', '很好吃, 但很贵'].m_comments
12 | 
13 | p ['病毒式媒体网站：让新闻迅速蔓延'].s_keywords
14 | 
15 | # ------------------- normal style --------------------
16 | # You can pass your token like `Bosonnlp.new("<your token from bosonnlp.com>")`
17 | nlp = Bosonnlp.new
18 | 
19 | p nlp.c_sentiment(['美好','悲惨'])
20 | p nlp.c_ner(['对于该小孩是不是郑尚金的孩子，目前已做亲子鉴定，结果还没出来，',\
21 |              '纪检部门仍在调查之中。成都商报记者 姚永忠'])
22 | p nlp.c_depparser(['我以最快的速度吃了午饭', '先留长头发再剃个秃子'])
23 | p nlp.c_tag(['这个世界好复杂', '计算机是科学么'])
24 | p nlp.c_classify(['俄否决安理会谴责叙军战机空袭阿勒颇平民',
25 |                   '邓紫棋谈男友林宥嘉：我觉得我比他唱得好',
26 |                   'Facebook收购印度初创公司'])
27 | 
28 | # ------------ mutiple texts API ------------------
29 | p nlp.m_cluster(['今天天气好', '今天天气好', '今天天气不错', '点点楼头细雨',\
30 |                  '重重江外平湖', '当年戏马会东徐', '今日凄凉南浦'])
31 | p nlp.m_comments(['很好吃, 有点贵', '贵了点, 好吃',
32 |                   '价格贵, 但很好吃', '很好吃, 但很贵'])
33 | 
34 | # -- push more than one time (Handle large amount of texts with mutiple API) --
35 | mh = nlp.create_multiple(:comments)
36 | mh.push(['很好吃, 有点贵', '贵了点, 好吃'])
37 | mh.push(['价格贵, 但很好吃', '很好吃, 但很贵'])
38 | mh.analysis  # Start computing on the server.
39 | p mh.result  # Call this will block the client until get result from the server.
40 | 
41 | mh.push(['很好吃, 有点贵', '贵了点, 好吃']) #  Yes, play incrementally!
42 | mh.analysis
43 | p mh.result
44 | 
45 | mh.clear  # Clear the texts.
46 | 
47 | # -------------------------------------
48 | #p nlp.s_time(['2013年二月二十八日下午四点三十分二十九秒'])
49 | p nlp.s_keywords(['病毒式媒体网站：让新闻迅速蔓延'])
50 | 
51 | # ----------- Handle extra param ------------
52 | query = { 'top_k' => 3 }
53 | p nlp.s_suggest(['粉丝'], :query => query)
54 | 


--------------------------------------------------------------------------------
/examples/usage.rb,:
--------------------------------------------------------------------------------
 1 | require 'bosonnlp'
 2 | 
 3 | # run `export BOSON_API_TOKEN="<your token from bosonnlp.com>"` first. e,g,
 4 | #     `export BOSON_API_TOKEN="bmw.1999.eW91ciB0b2tl"`
 5 | 
 6 | # -------------- string and list style ----------------
 7 | p '美好'.c_sentiment
 8 | p ['美好', '悲惨'].c_sentiment
 9 | 
10 | p ['很好吃, 有点贵', '贵了点, 好吃',
11 |    '价格贵, 但很好吃', '很好吃, 但很贵'].m_comments
12 | 
13 | p ['病毒式媒体网站：让新闻迅速蔓延'].s_keywords
14 | 
15 | # ------------------- normal style --------------------
16 | # You can pass your token like `Bosonnlp.new("<your token from bosonnlp.com>")`
17 | nlp = Bosonnlp.new
18 | 
19 | p nlp.c_sentiment(['美好','悲惨'])
20 | p nlp.c_ner(['对于该小孩是不是郑尚金的孩子，目前已做亲子鉴定，结果还没出来，',\
21 |              '纪检部门仍在调查之中。成都商报记者 姚永忠'])
22 | p nlp.c_depparser(['我以最快的速度吃了午饭', '先留长头发再剃个秃子'])
23 | p nlp.c_tag(['这个世界好复杂', '计算机是科学么'])
24 | p nlp.c_classify(['俄否决安理会谴责叙军战机空袭阿勒颇平民',
25 |                   '邓紫棋谈男友林宥嘉：我觉得我比他唱得好',
26 |                   'Facebook收购印度初创公司'])
27 | 
28 | # ------------ mutiple texts API ------------------
29 | p nlp.m_cluster(['今天天气好', '今天天气好', '今天天气不错', '点点楼头细雨',\
30 |                  '重重江外平湖', '当年戏马会东徐', '今日凄凉南浦'])
31 | p nlp.m_comments(['很好吃, 有点贵', '贵了点, 好吃',
32 |                   '价格贵, 但很好吃', '很好吃, 但很贵'])
33 | 
34 | # -- push more than one time (Handle large amount of texts with mutiple API) --
35 | mh = nlp.create_multiple(:comments)
36 | mh.push(['很好吃, 有点贵', '贵了点, 好吃'])
37 | mh.push(['价格贵, 但很好吃', '很好吃, 但很贵'])
38 | mh.analysis  # Start computing on the server.
39 | p mh.result  # Call this will block the client until get result from the server.
40 | 
41 | mh.push(['很好吃, 有点贵', '贵了点, 好吃']) #  Yes, play incrementally!
42 | mh.analysis
43 | p mh.result
44 | 
45 | mh.clear  # Clear the texts.
46 | 
47 | # -------------------------------------
48 | #p nlp.s_time(['2013年二月二十八日下午四点三十分二十九秒'])
49 | p nlp.s_keywords(['病毒式媒体网站：让新闻迅速蔓延'])
50 | 
51 | # ----------- Handle extra param ------------
52 | query = { 'top_k' => 3 }
53 | p nlp.s_suggest(['粉丝'], :query => query)
54 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # bosonnlp - Boson NLP SDK for Ruby
 2 | 
 3 | ### Install
 4 | 
 5 |     gem install bosonnlp
 6 | 
 7 | ### Usage
 8 | 
 9 | ```ruby
10 | # run `export BOSON_API_TOKEN="<your token from bosonnlp.com>"` first.
11 | 
12 | # ----- string and list style -----
13 | require 'bosonnlp'
14 | p '美好'.c_sentiment
15 | p ['很好吃, 有点贵', '贵了点, 好吃', '价格贵, 但很好吃', '很好吃'].m_comments
16 | 
17 | # ----- normal style (Handle extra param) -----
18 | nlp = Bosonnlp.new
19 | query = { 'top_k' => 3 }
20 | p nlp.s_suggest(['粉丝'], :query => query)
21 | 
22 | # -- push more than one time (Handle large amount of texts with mutiple API) --
23 | mh = nlp.create_multiple(:comments)
24 | mh.push(['很好吃, 有点贵', '贵了点, 好吃'])
25 | mh.push(['价格贵, 但很好吃', '很好吃, 但很贵'])
26 | mh.analysis  # Start computing on the server.
27 | p mh.result  # Call this will block until receive result from the server.
28 | 
29 | mh.push(['很好吃, 有点贵', '贵了点, 好吃'])  # Yes, play it incrementally!
30 | mh.analysis
31 | p mh.result
32 | 
33 | mh.clear  # Clear the texts.
34 | ```
35 | 
36 | You must have noticed the prefix **'c_','m_','s_'** before API names, e.g. sentiment API is called with the name 'c_sentiment'.
37 | 
38 | There are three types of APIs provided by bosonnlp.com.
39 | 
40 | This SDK supports APIs that even don't exists yet. It just need to know what kind of API it's handling by given the prefix. It's logic is not related to the API's name.
41 | 
42 | - Start with "m_": Multiple texts ones, like cluster API, it's meaningless to
43 |     cluster single text.
44 | - Start with "s_": Single ones, that limited to one text a time, and retrun one
45 |     result, e.g. keywords API.
46 | - Start with "c_": Common single ones, just like single ones, but can accept
47 |     multiple texts, and retrun multiple result, those results are not related
48 |     to each other, e.g. sentiment API.
49 | 
50 | 
51 | Check http://docs.bosonnlp.com for API details.
52 | 
53 | Check [examples/usage.rb](https://github.com/alal/bosonnlp/blob/master/examples/usage.rb) for more examples.
54 | 


--------------------------------------------------------------------------------
/lib/bosonnlp.rb:
--------------------------------------------------------------------------------
  1 | # -*- coding: utf-8 -*-
  2 | 
  3 | require 'json'
  4 | require 'httpclient'
  5 | require 'securerandom'
  6 | 
  7 | require 'bosonnlp/util'
  8 | require 'bosonnlp/mixin'
  9 | require 'bosonnlp/version'
 10 | 
 11 | # Main class
 12 | class Bosonnlp
 13 |   include Util
 14 |   # Operated task not found on server
 15 |   class TaskNotFoundError < StandardError
 16 |   end
 17 |   # Other task error return by server
 18 |   class TaskError < StandardError
 19 |   end
 20 |   # Http error
 21 |   class HTTPError < RuntimeError
 22 |   end
 23 |   # Task time out.
 24 |   class TimeoutError < StandardError
 25 |   end
 26 | 
 27 |   # Multiple text engine handler
 28 |   class MultipleHandler
 29 |     include Util
 30 |     def initialize(name, http_session, head, multiple_url)
 31 |       @name = name
 32 |       @http_session = http_session
 33 |       @head = head
 34 |       @id = SecureRandom.uuid
 35 |       @multiple_url = multiple_url
 36 |       @alpha = 0.8
 37 |       @beta = 0.45
 38 |     end
 39 | 
 40 |     def pre_process(data)
 41 |       fail ArgumentError, 'Wrong args, data must be a list'\
 42 |           ' which contains more than one element.'\
 43 |         unless Array === data and data.size > 1
 44 |       fail ArgumentError, 'Wrong args, data elements must be list of string '\
 45 |             'or {_id, text} Hash.'\
 46 |         unless String === data[0] or Hash === data[0]
 47 | 
 48 |       processd = []
 49 |       if String === data[0]
 50 |         data.each do |d|
 51 |           hash_t = {}
 52 |           hash_t['_id'] = SecureRandom.uuid
 53 |           hash_t['text'] = d
 54 |           processd << hash_t
 55 |         end
 56 |       else
 57 |         processd = data
 58 |       end
 59 | 
 60 |       processd
 61 |     end
 62 | 
 63 |     def get_url(method_name)
 64 |       @multiple_url % [@name, method_name.to_s, @id]
 65 |     end
 66 | 
 67 |     def push(*args)
 68 |       processd = pre_process args[0]
 69 |       request_args = get_hash_args(*args)
 70 | 
 71 |       is_ok = false
 72 |       processd.each_slice 100 do |s|
 73 |         request_args[:body] = s
 74 |         res = _http_request(:post, __method__, request_args)
 75 |         is_ok = res.ok?
 76 |         # TODO, fail for one exception?
 77 |       end
 78 |       is_ok
 79 |     end
 80 | 
 81 |     def analysis(*args)
 82 |       request_args = get_hash_args(*args)
 83 |       request_args['alpha'] ||=  @alpha
 84 |       request_args['beta'] ||=  @beta
 85 |       res = _http_request(:get, __method__, *args)
 86 |       res.ok?
 87 |     end
 88 | 
 89 |     def status(*args)
 90 |       res = _http_request(:get, __method__, *args)
 91 |       status = JSON.load(res.content)['status']
 92 |       case status.to_s.downcase
 93 |       when 'not found'
 94 |         raise TaskNotFoundError, "Operated task #{@id} not found on server."
 95 |       when 'error'
 96 |         raise TaskError, "Task #{@id} failed on server."
 97 |       end
 98 |       status
 99 |     end
100 | 
101 |     def _wait(timeout)
102 |       time_elapsed = 0.0
103 |       sleep_lenth = 1.0
104 |       while true
105 |         sleep sleep_lenth
106 |         serv_status = status
107 |         return if serv_status.downcase == "done"
108 |         time_elapsed += sleep_lenth
109 |         raise TimeoutError if timeout and time_elapsed > timeout
110 |         sleep_lenth *= 1.5 if sleep_lenth < 2**6
111 |       end
112 |     end
113 | 
114 |     def result(*args)
115 |       _wait get_hash_args(*args)['timeout']
116 |       res = _http_request(:get, __method__, *args)
117 |       JSON.load(res.content)
118 |     end
119 | 
120 |     def clear(*args)
121 |       res = _http_request(:get, __method__, *args)
122 |       res.ok?
123 |     end
124 |   end
125 | 
126 |   def initialize(token = nil)
127 |     @token = token
128 |     @env_token = ENV['BOSON_API_TOKEN']
129 |     proxy = ENV['HTTP_PROXY']
130 |     @http_session = HTTPClient.new(proxy)
131 | 
132 |     @readable = false
133 |     @fail_on_exception = false
134 |     @base_url = 'http://api.bosonnlp.com'
135 |     @analysis_url =  @base_url + '/%s/analysis'
136 |     @multiple_url =  @base_url + '/%s/%s/%s'
137 | 
138 |     @token ||= @env_token
139 |     fail 'No API token given or found in environment variables, run '\
140 |         '`export BOSON_API_TOKEN="<your api token>"` in your shell first.'\
141 |       unless @token
142 | 
143 |     headers = {}
144 |     headers['Accept'] = 'application/json'
145 |     headers['User-Agent'] = "bosonnlp-ruby #{Bosonnlp::VERSION} "\
146 |                             "(httpclient #{HTTPClient::VERSION})"
147 |     headers['Content-Type'] = 'application/json'
148 |     headers['X-Token'] = @token
149 |     @head = headers
150 | 
151 |     # TODO, args to change default values
152 |   end
153 | 
154 |   def common_api(name,  *args)
155 |     fail ArgumentError, 'Wrong args, first arg must be a list of string.'\
156 |       unless Array === args[0] and String === args[0][0]
157 | 
158 |     url = @analysis_url % name
159 | 
160 |     request_args = args[1] || {}
161 |     request_args[:body] = args[0]
162 | 
163 |     res = _http_request(:post, url, request_args)
164 |     JSON.load(res.content)
165 |   end
166 | 
167 |   def create_multiple(name)
168 |     MultipleHandler.new name.to_s, @http_session, @head, @multiple_url
169 |   end
170 | 
171 |   def multiple_api(name, *args)
172 |     begin
173 |       mh = create_multiple(name)
174 |       mh.push(*args)
175 |       mh.analysis(*args)
176 |       return mh.result
177 |     rescue => e
178 |       puts e.message
179 |       puts e.backtrace.inspect
180 |     ensure
181 |       mh.clear(*args) if mh
182 |     end
183 |   end
184 | 
185 |   def single_api(name, *args)
186 |     fail ArgumentError, 'Wrong args, data must be a list'\
187 |         ' which contains one and only one string element.'\
188 |       unless Array === args[0] and String === args[0][0] and args[0].size == 1
189 | 
190 |     url = @analysis_url % name
191 | 
192 |     request_args = get_hash_args(*args)
193 |     request_args[:body] = args[0][0]
194 | 
195 |     res = _http_request(:post, url, request_args)
196 |     JSON.load(res.content)
197 |     # :query, :body, :header, :follow_redirect
198 |   end
199 | 
200 |   def method_missing(name, *args)
201 |     case name.to_s
202 |     when /^c_(.+)$/
203 |       common_api($1, *args)
204 |     when /^m_(.+)$/
205 |       multiple_api($1, *args)
206 |     when /^s_(.+)$/
207 |       single_api($1, *args)
208 |     else
209 |       super
210 |     end
211 |   end
212 | end
213 | 


--------------------------------------------------------------------------------