Class: Doubapi::Douban

Inherits:
Object
  • Object
show all
Defined in:
lib/doubapi.rb

Class Method Summary collapse

Class Method Details

.compare_date(a, b) ⇒ Object

return true if a >= b a,b could be one of the following format 2010.01 2010.1 2010#1 2010-1 2010年1



156
157
158
159
160
 
# File 'lib/doubapi.rb', line 156

def compare_date a , b
	ya, ma = a.scan(/\d{1,4}/)
	yb, mb = b.scan(/\d{1,4}/)
	return true if (ya.to_i * 12 + ma.to_i ) >= (yb.to_i*12+mb.to_i)
end

.douban_get_xml(url) ⇒ Object 



80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
 
# File 'lib/doubapi.rb', line 80

def douban_get_xml url
	puts url
	#I have forgot why i need to specify the user agend
  doc = open(url, :proxy => nil, 'User-Agent' => 'ruby')
  if doc.nil?
    puts "error:failed to open #{url}"
    return nil;
  end
  
  Nokogiri::HTML(doc,nil, "utf-8")
  
  #Nokogiri::HTML(open(url,:proxy => nil,'User-Agent' => 'ruby'),nil, "utf-8")
	#no network access, used to simulator 
	#doc = File.read(File.join(RAILS_ROOT, "app","controllers","event_sample.xml"))
	#Nokogiri::HTML(doc,nil, "utf-8")
  #Nokogiri::HTML(open(url,:proxy => nil,'User-Agent' => 'ruby'),nil, "utf-8")
end

.formate_release_date(release_date) ⇒ Object 



163
164
165
166
167
168
169
170
171
172
173
174
175
176
 
# File 'lib/doubapi.rb', line 163

def formate_release_date release_date
   	#make release data YY-MM-DD style
  	r = release_date.scan(/\d{1,4}/)
  	#if DD was not specified
  	r << "01"         if r.size == 2
  	r << "01" << "01" if r.size == 1

  	y , m , d = r

  	m = "01" unless (1..12).include?(m.to_i) 
  	d = "01" unless (1..30).include?(d.to_i)
  	
  	"#{y}-#{m}-#{d}" 
end

.looks_like_a_live_show?(e, artist) ⇒ Boolean

TODO

Returns:

  • (Boolean) 


134
135
136
137
138
139
140
141
142
143
144
145
146
147
 
# File 'lib/doubapi.rb', line 134

def looks_like_a_live_show? e, artist

	#check e.when should happen
	#2010-08-13F21:30:00+08:00
	_,_,_,hour = e.when.scan(/\d{1,4}/);

	if artist.downcase == "all"
		return true if hour.to_i > 18
	else
		return true if hour.to_i > 18 and e.what.include?(artist) 
	end

	return false
end

.parse_date(date) ⇒ Object

return Time object date format is “时间：2010年8月13日 周五 21:30 - 23:55” or 2010-08-13F21:30:00+08:00



235
236
237
238
 
# File 'lib/doubapi.rb', line 235

def parse_date date
  year, month , day = date.scan(/\d{1,4}/)
  Time.local(year,month,day)
end

.search_ablum(h) ⇒ Object 



122
123
124
125
126
127
128
129
130
 
# File 'lib/doubapi.rb', line 122

def search_ablum h
  artist_chinese = h[:singer]
  max=h[:max_result]||10
  keywords= "%" + artist_chinese.each_byte.map {|c| c.to_s(16)}.join("%")
  uri="http://api.douban.com/music/subjects?tag=#{keywords}&start-index=1&max-results=#{max}"
  #Let's grab it slowly to avoid being baned...	
  sleep(7) 	
  douban_get_xml(uri)
end

.search_albums_of(h) ⇒ Object

search albums tagged with h. It is not quite accurate. I have seen some irrevlant result are returned.



183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
 
# File 'lib/doubapi.rb', line 183

def search_albums_of h 
  artist = h[:singer]
  after_date = h[:since]||"1900.01"
  doc = search_ablum h
 
  if(doc.nil?) 
    return [0,[]];
  end
  
  #the totalResult here trying
  #totalResults = doc.at_xpath(".//totalresults").text.to_i
   
  albums=[]
  doc.xpath("//entry").each do |entry|
  	title = entry.at_xpath(".//title").text
  	#author
  	authorItem = entry.at_xpath(".//name")
  	author = if authorItem.nil? then artist else authorItem.text end
  	#link - pc web
		link =  entry.at_xpath(".//link[@rel='alternate']")["href"]
	  cover_thumnail = entry.at_xpath(".//link[@rel='image']")["href"]
	  
	  #cover big 
	  #example:
	  #thumbnail http://img1.douban.com/spic/s1461123.jpg
	  #big       http://img1.douban.com/lpic/s1461123.jpg
	  cover_big = cover_thumnail.gsub("spic","lpic");
	  
	  #publisher
  	pubItem = entry.at_xpath(".//attribute[@name='publisher']")
  	publisher = if pubItem.nil? then "unknow" else pubItem.text end
    #link - mobile_site
  	mobile_site = entry.at_xpath(".//link[@rel='mobile']")["href"]
    #release_date
    pubDateItem = entry.at_xpath(".//attribute[@name='pubdate']")
    release_date = if pubDateItem.nil? then "0000-00" else pubDateItem.text end
  	formated_release_day = formate_release_date(release_date)
  	#check the release date
  	if compare_date release_date, after_date			
  		albums << Doubapi::Album.new(author, title, formated_release_day, link, cover_thumnail,cover_big ,publisher,mobile_site)
  	end
  end
  #improve ME
  [albums.size,albums]
end

.search_event(h) ⇒ Object

return Atom Douban search : will return results that does not match



102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
 
# File 'lib/doubapi.rb', line 102

def search_event h
  puts h.inspect

  key_chinese = h[:key]
  location    = h[:location] || "shanghai"
  start_index = h[:start_index] || 1
  max         = h[:max_result]|| 20

  if (key_chinese.downcase == "all")
	uri="http://api.douban.com/event/location/#{location}?type=music&start-index=#{start_index}&max-results=#{max}"
  else
  keywords= "%" + key_chinese.each_byte.map {|c| c.to_s(16)}.join("%")
  uri="http://api.douban.com/events?q=#{keywords}&location=#{location}&start-index=#{start_index}&max-results=#{max}"
  end

  #Let's grab it slowly to avoid being baned...	
  sleep(7) 	
  douban_get_xml(uri)
end

.search_events_of(h = {}) ⇒ Object 



240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
 
# File 'lib/doubapi.rb', line 240

def search_events_of(h={})
  artist = h[:key]
  after_date =  h[:after_date]||Time.now.strftime("%Y-%m")
  doc = search_event h
  events=[]
  
  #NOTES:all the key will be downcase-d
  totalResults = doc.at_xpath(".//totalresults").text.to_i
  doc.xpath("//entry").each do |entry|
    #pp entry
    title = entry.at_xpath(".//title").text
    #attribute is starttime NOT startTime as specified in the xml
    start_time = entry.at_xpath('.//when')["starttime"]
    #  city = entry.at_xpath('.//location').text
    where = entry.at_xpath('.//where')["valuestring"]
    link =  entry.at_xpath(".//link[@rel='alternate']")["href"]
    what = entry.at_xpath(".//content").text
    posterItem = entry.at_xpath(".//link[@rel='image-lmobile']")
    poster_mobile = if posterItem.nil? then "empty" else posterItem["href"] end
    authItem = entry.at_xpath(".//author")
    iconLink = if authItem.nil? then nil else authItem.at_xpath(".//link[@rel='icon']") end
    bar_icon = if (iconLink.nil?) then nil else iconLink["href"] end

	#check the date
	if parse_date(start_time) > parse_date(after_date)
	    events << Doubapi::Event.new(title, start_time, where, what, link, poster_mobile, bar_icon)
    end
  end
	
  #filtering of the results
  events.select!{|e| looks_like_a_live_show?(e,artist)}
  
  [totalResults, events]
end