yt_util 0.0.1 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 4b3dfe4b4c3b37181be955ee39eb039d8385844d
-  data.tar.gz: 030dc138c1663339fc1b029689b7486d34fa07fa
+  metadata.gz: e7b1ed94c177fb7d05f71eae24e87db5467d01c1
+  data.tar.gz: 487ea160d323c3b8ef432d91523d2afbc1d44a0b
 SHA512:
-  metadata.gz: a8fe781a3e22f5c1758e3171733ee85183204810563c6cf1602779add34265cc05d25b602101bd0efa60e9170d896e591f0547797f8942c823e0abd4e884a1b0
-  data.tar.gz: 478691e74b25f47ff4a9bfcec18703665c5432b9d071dde56d3e96473b0ab6a06cd48a394c9fb49b7535b5592b4633f159c9fc7b155945a5c0408d29d61f8a81
+  metadata.gz: d529e7cdf11e6b2f4b5e844eddceec0886213894fa87aa70c642042a325a58ac20684c536cb696a5e2ca4111e17c194294a4ead6692f2d940dd16899b5aecf0c
+  data.tar.gz: cfa6c69365757262e89dd62ee777c9a7dcef6eede3093393eec9ada9a3b0dd34270d2a9fdadcdc5d656f48af51b22b7db7cc3ec3fa3ac8523eb8593b19808669

data/lib/yt_util/scrape.rb

@@ -5,69 +5,98 @@ module YtUtil
     # OR range etc https://support.google.com/websearch/answer/136861?hl=en
     # punctuation: https://support.google.com/websearch/answer/2466433
 
-    def self.query(search, filters = "") # Returns an Array of Hash results
-      results = Mechanize.new.
-        tap { |i| i.follow_meta_refresh = true }.
-        get("https://www.youtube.com/results?search_query=#{Addressable::URI.parse(search).normalize + filters}")
-      parse_query(results.parser)
+    def self.raw_query(search, filters = "")
+      raise "Invalid object type" unless search.is_a? String and filters.is_a? String
+      request_query(search,filters)
     end
 
-    def self.video_statistics(video_code) # Video page hit.
-      result = Nokogiri::HTML(open("https://www.youtube.com/watch?v=#{video_code}"))
-      parse_video_page(result)
+    def self.query(search = nil, filters = "", &qry)
+      result = qry.try(:call) || raw_query(search,filters)
+      parse_query(result)
     end
 
-    def self.user_statistics(username)
-      results = Mechanize.new.
-        tap { |i| i.follow_meta_refresh = true }.
-        get("https://www.youtube.com/user/#{username}/about")
-      parse_user(results.parser)
+    def self.raw_video_stats(video_code)
+      raise "Invalid object type" unless video_code.is_a? String
+      raise "Invalid video code" unless video_code.length == 11
+      request_video_stats(video_code)
+    end
+
+    def self.video_stats(video_code = nil, &qry)
+      result = qry.try(:call) || raw_video_stats(video_code)
+      parse_video_page(video_code, result)
+    end
+
+    def self.raw_user_stats(username)
+      raise "Invalid object type" unless username.is_a? String
+      request_user_stats(username)
+    end
+
+    def self.user_stats(username = nil, &qry)
+      result = qry.try(:call) || raw_user_stats(username)
+      parse_user(result)
     end
 
     private
+    def self.request(web_request)
+      try { Mechanize.new.tap { |i| i.follow_meta_refresh = true }.get(web_request).parser } ||
+        Nokogiri::HTML(open(web_request))
+    end
+
+    def self.request_query(search, filters = "")
+      web_request = "https://www.youtube.com/results?search_query=#{Addressable::URI.parse(search).normalize + filters}"
+      request(web_request)
+    end
+
+    def self.request_video_stats(video_code)
+      web_request = "https://www.youtube.com/watch?v=#{video_code}"
+      request(web_request)
+    end
+
+    def self.request_user_stats(username)
+      web_request = "https://www.youtube.com/user/#{username}/about"
+      request(web_request)
+    end
+
     def self.parse_query(query_result)
       query_result.css("ol.item-section > li")[1..-1].map do |result|
         {
-          title: result.css("div:nth-child(1)").css("div:nth-child(2)").css("h3").text,
-          video: result.css("div:nth-child(1)").css("div:nth-child(2)").css("h3 > a").first[:href].dup.tap{|i|i.replace i[(i.index("=").to_i+1)..-1]},
-          views: Integer(String(/(\d+)/.match(result.css("div:nth-child(1)").css("div:nth-child(2)").css("li:nth-child(3)").text.gsub(",", "")))),
-          new: !!result.css("div:nth-child(1)").css("div:nth-child(2)").css("div:nth-child(4)").css("ul:nth-child(1)").text["New"],
-          hd: !!result.css("div:nth-child(1)").css("div:nth-child(2)").css("div:nth-child(4)").css("ul:nth-child(1)").text["HD"],
-          description: result.css("div:nth-child(1)").css("div:nth-child(2)").css("div:nth-child(3)").text,
-          length: result.css("div:nth-child(1)").css("div:nth-child(1)").css("a:nth-child(1)").css("span:nth-child(2)").text
+          title: try {result.css("div:nth-child(1)").css("div:nth-child(2)").css("h3").text},
+          video: try {result.css("div:nth-child(1)").css("div:nth-child(2)").css("h3 > a").first[:href].dup.tap{|i|i.replace i[(i.index("=").to_i+1)..-1]}},
+          views: try {result.css('li').select {|i| i.text =~ /^[\d,]{1,} views/ }.first.text.split.first.gsub(",","_").to_i},
+          new: try {!!result.css("div:nth-child(1)").css("div:nth-child(2)").css("div:nth-child(4)").css("ul:nth-child(1)").text["New"]},
+          hd: try {!!result.css("div:nth-child(1)").css("div:nth-child(2)").css("div:nth-child(4)").css("ul:nth-child(1)").text["HD"]},
+          description: try {result.css("div:nth-child(1)").css("div:nth-child(2)").css(".yt-lockup-description").text},
+          length: try {result.css("div:nth-child(1)").css("div:nth-child(1)").css("a:nth-child(1)").css("span:nth-child(2)").text}
         }
       end
     end
 
-    def self.parse_video_page(query_result)
+    def self.parse_video_page(video_code, query_result)
       {
-        video: video_code,
-        user_name: query_result.css('a.yt-user-name').text,
-        description: query_result.css('p#eow-description').text,
-        category: query_result.css('li.watch-meta-item:nth-child(1)').css('ul:nth-child(2)').css('li:nth-child(1)').css('a:nth-child(1)').first[:href].try(:[], 1..-1),
-        # Views are rounded to nearest thousand.
-        views: String(/(\d+)/.match(query_result.css('div.watch-view-count').text.strip.gsub(',', ''))).to_i*1000,
-        # Likes are rounded to nearest thousand.
-        likes: String(/(\d+)/.match(query_result.css('button#watch-like').text.strip.gsub(',', ''))).to_i*1000,
-        # Dislikes are rounded to nearest thousand.
-        dislikes: String(/(\d+)/.match(query_result.css('button#watch-dislike').text.strip.gsub(',', ''))).to_i*1000,
-        published: query_result.css('strong.watch-time-text').text[13..-1],
-        license: query_result.css('li.watch-meta-item:nth-child(2)').text.gsub("\n", '').strip.tap {|i| i.replace i[i.index(' ').to_i..-1].strip}
+        video: video_code || try {query_result.css('meta').select {|i| i.attributes["property"].try(:value) =~ /og:url/}.first["content"].match(/\?v=(.+)/)[1]},
+        user_name: try {query_result.css('li').css('a').select{|i| i.text =~ / by [a-z0-9]{1,}/i}.map {|i| i.text.match(/by ([a-z0-9]{1,})/i)[1]}.first},
+        description: try {query_result.css('p#eow-description').text},
+        category: try {query_result.css('.watch-meta-item').css('a').text},
+        views: try {String(/(\d+)/.match(query_result.css('div.watch-view-count').text.strip.gsub(',', ''))).to_i},
+        likes: try {String(/(\d+)/.match(query_result.css('button#watch-like').text.strip.gsub(',', ''))).to_i},
+        dislikes: try {String(/(\d+)/.match(query_result.css('button#watch-dislike').text.strip.gsub(',', ''))).to_i},
+        published: try {query_result.css('strong.watch-time-text').text.match(/ on ([a-z0-9, ]{11,})/i)[1]},
+        license: try {query_result.css('li.watch-meta-item:nth-child(2)').text.gsub("\n", '').strip.tap {|i| i.replace i[i.index(' ').to_i..-1].strip}}
       }
     end
 
     def self.parse_user(query_result)
-      views_n_subs = query_result.css('.about-stats').
+      views_n_subs = try {query_result.css('.about-stats').
         css('li').take(2).map{|i| i = i.text.strip; {
         i.match(/[a-z]+/)[0] => i.match(/[\d,]+/)[0]}
-      }.inject(:update)
+      }.inject(:update)}
 
       {
-        description: query_result.css('.about-description').css('p').text,
-        link: query_result.css('a[title="Google+"]')[0]["href"],
-        views: views_n_subs["views"],
-        subscribers: views_n_subs["subscribers"],
-        joined: query_result.css('.about-stats').css('.joined-date').text.strip
+        description: try {query_result.css('.about-description').css('p').text},
+        link: try {query_result.css('a[title="Google+"]')[0]["href"]},
+        views: try {views_n_subs["views"]},
+        subscribers: try {views_n_subs["subscribers"]},
+        joined: try {query_result.css('.about-stats').css('.joined-date').text.strip}
       }
     end
 

data/lib/yt_util/try.rb

@@ -0,0 +1,102 @@
+unless defined? try
+  class Object
+    # Invokes the public method whose name goes as first argument just like
+    # +public_send+ does, except that if the receiver does not respond to it the
+    # call returns +nil+ rather than raising an exception.
+    #
+    # This method is defined to be able to write
+    #
+    #   @person.try(:name)
+    #
+    # instead of
+    #
+    #   @person.name if @person
+    #
+    # +try+ calls can be chained:
+    #
+    #   @person.try(:spouse).try(:name)
+    #
+    # instead of
+    #
+    #   @person.spouse.name if @person && @person.spouse
+    #
+    # +try+ will also return +nil+ if the receiver does not respond to the method:
+    #
+    #   @person.try(:non_existing_method) # => nil
+    #
+    # instead of
+    #
+    #   @person.non_existing_method if @person.respond_to?(:non_existing_method) # => nil
+    #
+    # +try+ returns +nil+ when called on +nil+ regardless of whether it responds
+    # to the method:
+    #
+    #   nil.try(:to_i) # => nil, rather than 0
+    #
+    # Arguments and blocks are forwarded to the method if invoked:
+    #
+    #   @posts.try(:each_slice, 2) do |a, b|
+    #     ...
+    #   end
+    #
+    # The number of arguments in the signature must match. If the object responds
+    # to the method the call is attempted and +ArgumentError+ is still raised
+    # in case of argument mismatch.
+    #
+    # If +try+ is called without arguments it yields the receiver to a given
+    # block unless it is +nil+:
+    #
+    #   @person.try do |p|
+    #     ...
+    #   end
+    #
+    # You can also call try with a block without accepting an argument, and the block
+    # will be instance_eval'ed instead:
+    #
+    #   @person.try { upcase.truncate(50) }
+    #
+    # Please also note that +try+ is defined on +Object+. Therefore, it won't work
+    # with instances of classes that do not have +Object+ among their ancestors,
+    # like direct subclasses of +BasicObject+. For example, using +try+ with
+    # +SimpleDelegator+ will delegate +try+ to the target instead of calling it on
+    # the delegator itself.
+    def try(*a, &b)
+      try!(*a, &b) if a.empty? || respond_to?(a.first)
+    end
+
+    # Same as #try, but will raise a NoMethodError exception if the receiver is not +nil+ and
+    # does not implement the tried method.
+
+    def try!(*a, &b)
+      if a.empty? && block_given?
+        if b.arity.zero?
+          instance_eval(&b)
+        else
+          yield self
+        end
+      else
+        public_send(*a, &b)
+      end
+    end
+  end
+
+  class NilClass
+    # Calling +try+ on +nil+ always returns +nil+.
+    # It becomes especially helpful when navigating through associations that may return +nil+.
+    #
+    #   nil.try(:name) # => nil
+    #
+    # Without +try+
+    #   @person && @person.children.any? && @person.children.first.name
+    #
+    # With +try+
+    #   @person.try(:children).try(:first).try(:name)
+    def try(*args)
+      nil
+    end
+
+    def try!(*args)
+      nil
+    end
+  end
+end

data/lib/yt_util/version.rb

@@ -1,3 +1,3 @@
 module YtUtil
-  VERSION = "0.0.1"
+  VERSION = "0.1.0"
 end

data/lib/yt_util.rb

@@ -5,6 +5,7 @@ require "mechanize"
 require "yt_util/version"
 require "yt_util/scrape"
 require "yt_util/url"
+require "yt_util/try"
 
 module YtUtil
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: yt_util
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.1.0
 platform: ruby
 authors:
 - Daniel P. Clark
@@ -96,6 +96,7 @@ files:
 - Rakefile
 - lib/yt_util.rb
 - lib/yt_util/scrape.rb
+- lib/yt_util/try.rb
 - lib/yt_util/url.rb
 - lib/yt_util/version.rb
 - spec/spec_helper.rb