stylemooncat 0.0.8 → 0.0.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d22a25a0df1771109c5519e3bd76209381878a77
-  data.tar.gz: 21b0f00b49821cf758437aacdff9af2121663b9c
+  metadata.gz: 15167c685cf84f8384b77774e8131bf5e2858221
+  data.tar.gz: e4b76b8567f8b3d98fa1b6835708d35f00378b9a
 SHA512:
-  metadata.gz: ac9206c21e680963f0fa251a499c776f47715668953ba428c8d8071d05c39091a1771edbd8e3f25914fe130c921f13b4e99bcac5544a226e601b86eacf32bc5f
-  data.tar.gz: c532ddf1a5d9a03be1e73e22ec1190a87a56853413d6ce63e7a96fe13d567426a0a0190d89cc4b4bf4a98bee0c8e672344902a67a386ffb48be51e0cb9229de5
+  metadata.gz: 01cceb18d13a90075f03e839437846befc18220e8a6d740b27f5d6cdf399527bcb15ba5d7d5b1991c91988b53052567a9dabdca8d3bae2039c53a65b5ad4a450
+  data.tar.gz: 53f473894440e01eb118911574554f33594a0c0aed19280f9f1dec5ab336980eaf00b6e0e7159236d48f00b91c146df8545185765c2ec6407c2d93297e55e0b0

data/bin/stylemooncat

@@ -13,6 +13,13 @@ price_boundary.push(ARGV[4])
 options[:keyword]=ARGV[2]
 options[:price_boundary]=price_boundary
 
-#puts @scraper.scrape(ARGV[0],ARGV[1],ARGV[2],ARGV[3],ARGV[4],ARGV[5])
-puts @scraper.scrape(ARGV[0],options)
+puts @scraper.scrape(ARGV[0],ARGV[1],ARGV[2],ARGV[3],ARGV[4],ARGV[5])
+if ARGV[5]=='color'
+ puts @scraper.scrape_contain_color(ARGV[0],options)
+else
+  puts @scraper.scrape(ARGV[0],options)
+end
+
 #puts @scraper.scrape("shoes",{:keyword=>"none",:page_limit=>3,:price_boundary=>[0,600]})
+#puts '-----------'
+#puts @scraper.scrape("shoes",{:keyword=>"跟鞋",:page_limit=>1})

data/lib/stylemooncat/scraper.rb

@@ -2,7 +2,7 @@
 require 'oga'
 require 'open-uri'
 require 'open-uri-s3'
-
+require 'iconv'
 # scrape data
 module StyleMoonCat
   class Scraper
@@ -60,7 +60,7 @@ module StyleMoonCat
 
     @@COLOR_ITEM_XPATH = "//option"
 
-    def scrape_contain_color(category,options)
+    def scrape_contain_color(category,options={})
         @@IsScrapeColor=1
         filter_results = scrape(category,options)
         filter_results_with_color = filter_results.each do |x|
@@ -87,10 +87,12 @@ module StyleMoonCat
         item.text.split(" ")[0].split("：")[1]
     end
 
-    def scrape(category,options)
+    def scrape(category,options={})
+      options[:keyword]
+
+    #  keyword= Iconv.conv('big5','utf-8',options[:keyword])
       keyword= options[:keyword]
       page_limit=options[:page_limit]
-      puts options
 
       if options[:price_boundary]!= nil && options[:price_boundary].length ==2
           if options[:price_boundary][0].to_i>options[:price_boundary][1].to_i
@@ -211,7 +213,7 @@ module StyleMoonCat
 
     private
     def uri_with_keyword(uri, keyword)
-      "#{uri}&keyword=#{keyword}"
+      "#{uri}&keyword=#{URI.escape(keyword)}"
     end
 
     def uri_with_page(uri, page)
@@ -266,7 +268,7 @@ module StyleMoonCat
 
     def extract_images(item)
       result=[]
-      result.push(item.xpath(@@IMAGE_XPATH).attribute(:src).first.value)
+      result.push('http://www.stylemooncat.com.tw'+item.xpath(@@IMAGE_XPATH).attribute(:src).first.value)
 
     end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: stylemooncat
 version: !ruby/object:Gem::Version
-  version: 0.0.8
+  version: 0.0.9
 platform: ruby
 authors:
 - Even Chang
@@ -11,7 +11,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-01-03 00:00:00.000000000 Z
+date: 2016-01-05 00:00:00.000000000 Z
 dependencies: []
 description: This is a gem scraping StyleMoonCat's website.Input category name,page
   limit,searcing keyword,and price range,and it will return the items with title,price,image,and