tomosia_amanaplus_crawl 0.2.3 → 0.2.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e3bc628d6f82c6357d3a4362464870e6e1ac3bc696ad7bc3f91fd7fd5901b7fe
-  data.tar.gz: 7184c0c7edaecb96312ab0650fb4395cc02d50175e51c0d35ce2600e6c9ac61b
+  metadata.gz: 78adb558a362c9594df8a864dad75bf4bc95472eb64e2aa9ef1499d2a39f6837
+  data.tar.gz: d2d670dbc023aa1f1ce265f78245b18a2177da373b0e4436a8a88ced97f9677b
 SHA512:
-  metadata.gz: ab5bc7f49db71fb490d1cf99da244c1321edcf98aec5678415229f4faf6a29fa5edd2589d8abf90b66790e142c1ef276a8e24bb65755c6b2eb5f2ea805dff1f6
-  data.tar.gz: b89334b70ee3693add75baa2e49bff8e2b968c1caac0d21adea45c41a31984b21e713127d8412cde046a478d714da2bd3a9186e0838923c964262967c1f60bf4
+  metadata.gz: 382535d1072a6803ffd0166ee70b99187514b9119656e94c852519ef3c608ed6d829a86d19d4d8a20027a6fc1e1a3fe909defc4644e139002998751a40bb124c
+  data.tar.gz: d18803def75f4efa16e4e4339e3f0b1e64a6a54f77c4a8a6b45d5cb8991e6a3e5a87a0c4d1047c22ad8fdab4266c1c3dedac07660663aeb6557624266e6c6809

data/Gemfile.lock

@@ -1,8 +1,7 @@
 PATH
   remote: .
   specs:
-    tomosia_amanaplus_crawl (0.1.8)
-      httparty (= 0.18.1)
+    tomosia_amanaplus_crawl (0.2.3)
       nokogiri (= 1.10.10)
       spreadsheet (= 1.2.6)
       thor
@@ -11,14 +10,7 @@ GEM
   remote: https://rubygems.org/
   specs:
     diff-lcs (1.4.4)
-    httparty (0.18.1)
-      mime-types (~> 3.0)
-      multi_xml (>= 0.5.2)
-    mime-types (3.3.1)
-      mime-types-data (~> 3.2015)
-    mime-types-data (3.2020.0512)
     mini_portile2 (2.4.0)
-    multi_xml (0.6.0)
     nokogiri (1.10.10)
       mini_portile2 (~> 2.4.0)
     rake (12.3.3)

data/lib/tomosia_amanaplus_crawl.rb

@@ -15,11 +15,11 @@ module TomosiaAmanaplusCrawl
 
       pages = parsed_page.css("div.c-paginate__nums").css('a').last.text.to_i # tổng số page
       images_listings = parsed_page.css("div.p-search-result__body") # danh sách các thẻ div chứa image
-    
+
       # lấy tổng số image
       total = parsed_page.css("h1.p-search-result__ttl").text.split(' ').first
       total = total[(6 + keyword.length)..(total.length - 1)].chop.chop.chop.gsub(',', '').to_i
-      if max > total # nếu max lớn hơn total thì max = total => vẫn lấy hết
+      if max == nil || max > total # nếu max lớn hơn total thì max = total => vẫn lấy hết
         max = total
       end
 
@@ -34,11 +34,11 @@ module TomosiaAmanaplusCrawl
       curr_index = 1
       while curr_page <= pages
         puts "Crawling page #{curr_page}..........."
-    
+
         pagination_unparsed_page = open("#{URL}/#{keyword}?page=#{curr_page}").read
         pagination_parsed_page = Nokogiri::HTML(pagination_unparsed_page)
         pagination_images_listings = pagination_parsed_page.css("div.p-item-thumb")
-    
+
         pagination_images_listings.each do |img|
           if curr_index > max
             return images
@@ -54,7 +54,7 @@ module TomosiaAmanaplusCrawl
           images << current_image
           curr_index += 1
         end
-    
+
         curr_page += 1
       end
       images
@@ -75,7 +75,7 @@ module TomosiaAmanaplusCrawl
               File.open("#{path}/#{curr_image[:url].split('/').last}", "a+") do |file|
                 file.write(image.read) # lưu hình ảnh
                 curr_image[:size] = image.size # cập nhật lại size trong mảng images
-                print "."
+                print '.'
               end
             end # end open
           rescue => exception
@@ -99,10 +99,9 @@ module TomosiaAmanaplusCrawl
       book = Spreadsheet::Workbook.new
       sheet1 = book.create_worksheet
 
-      i = 0
       sheet1.row(0).concat %w{Title Url Size(bytes) Extension}
       puts "Writing..........."
-      images.each do |img|
+      images.each_with_index do |img, i|
         sheet1.row(i += 1).push img[:title], img[:url], img[:size], img[:extension]
       end
       puts "Writed."

data/lib/tomosia_amanaplus_crawl/cli.rb

@@ -8,7 +8,7 @@ module TomosiaAmanaplusCrawl
     option :destination
     option :max
     def crawl(keyword)
-      TomosiaAmanaplusCrawl::Crawler.new.run(keyword, options[:destination], options[:max].to_i)
+      TomosiaAmanaplusCrawl::Crawler.new.run(keyword, options[:destination], options[:max] == nil ? nil : options[:max].to_i)
     end
   end
 end

data/lib/tomosia_amanaplus_crawl/version.rb

@@ -1,3 +1,3 @@
 module TomosiaAmanaplusCrawl
-  VERSION = "0.2.3"
+  VERSION = "0.2.4"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tomosia_amanaplus_crawl
 version: !ruby/object:Gem::Version
-  version: 0.2.3
+  version: 0.2.4
 platform: ruby
 authors:
 - Nhat Huy
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2020-08-10 00:00:00.000000000 Z
+date: 2020-08-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri