rumors-api-client 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e01c7b9c0baec0166aa6dbd36e44d1c9598a7dac1c51bc34eb8c3128d06bae65
-  data.tar.gz: 6fa1296b339fcb393c978796e15fdcbf6e4e115ba012fa3b05a2324b37bc752b
+  metadata.gz: 83f729f071e745c9af334ed82e1f36920dd4cf87a1e6372298aa240d25d178da
+  data.tar.gz: ffe2a61891125cfaece0f80c3bd84f0e151a129dfb48e1f8e7a73854c5f39f19
 SHA512:
-  metadata.gz: '090d5f3963c7e08b457678a3d5e14d8af9d6e232794bbf83838d72530bf6d317f8cad9cb0752a302ed3d1e2972db65925505723a1e86655af134f036b709e6a2'
-  data.tar.gz: 5b4d41c3bea11c67f7da88d377a0998f5222355f2140e0ec08ee9573e982ce035a5012552c5525063c3f47ad80507bade8c1252892792b9bd925e8e1b44b0eaa
+  metadata.gz: 8a934bdb885d196c884792db0778a1382e770b083f21394aebb93ac6c82add6651caf591ea975827cbd6081fe875de36effd8e2b2a20f2a2b93664f81f15f4ed
+  data.tar.gz: 4235fd57b928d7fb535d88d754b37a6e8b7d21aebece396ac4655e8ce79bc9119f99c70d9c8ae64b38079926f2ac6b42468f151e35f2be414c729eb90937d67e

data/lib/rumors/api/client/base.rb

@@ -6,7 +6,8 @@ module Rumors
         SIMILARITY = 0.8
 
         def initialize(text)
-          @text = text.split.join("")
+          @text = text.strip
+          @urls = URI.extract(@text).map { |url| URI.parse(URI.escape(url)) }
         end
 
         def search
@@ -25,10 +26,18 @@ module Rumors
 
         def return_article
           contents = parse_content
-          most_like = calculate_similarity(contents)
-          return unless most_like[:score] > SIMILARITY
+          return if contents.nil? || contents.empty?
+          article_id = nil
+
+          if @urls.any?
+            article_id = compare_urls(contents)
+          else
+            most_like = calculate_similarity(contents)
+            return unless most_like[:score] > SIMILARITY
+            article_id = most_like[:article_id]
+          end
 
-          find_article(most_like[:article_id])
+          find_article(article_id) if article_id
         end
 
         def find_article(article_id)
@@ -36,13 +45,36 @@ module Rumors
         end
 
         def parse_content
+          # [{ 'article_id' => TfIdfSimilarity::Document(text), 'urls' => ["url"] }]
           parsed_articles = JSON.parse(@articles.body)
           parsed_articles['data']['ListArticles']['edges'].map do |article|
             node = article['node']
-            Hash[node['id'], TfIdfSimilarity::Document.new(node['text'])]
+            content = Hash[node['id'], TfIdfSimilarity::Document.new(node['text'])]
+            content['urls'] = node['hyperlinks'].nil? ? nil : node["hyperlinks"].map { |link| URI.parse(URI.escape(link["url"])) }
+            content
           end
         end
 
+        def compare_urls(contents)
+          contents.each do |content|
+            return content.keys.first if exist_same_url?(content['urls'])
+          end
+        end
+
+        def exist_same_url?(response_urls)
+          response_urls.each do |response_url|
+            @urls.each do |url|
+              next unless response_url.host == url.host
+
+              response_uris = response_url.path.split("/").reject { |path| path.empty? }
+              uris = url.path.split("/").reject { |path| path.empty? }
+              return true if (response_uris & uris) == response_uris
+            end
+          end
+
+          false
+        end
+
         def calculate_similarity(contents)
           # NOTE: https://github.com/jpmckinney/tf-idf-similarity
           most_like = {

data/lib/rumors/api/client/utils/list_articles.rb

@@ -18,7 +18,7 @@ module Rumors
           private
 
           def gql_query
-            <<-GQL
+            <<~GQL
             query($text: String) {
               ListArticles(
                 filter: { moreLikeThis: { like: $text } }
@@ -29,6 +29,9 @@ module Rumors
                   node {
                     id
                     text
+                    hyperlinks {
+                      url
+                    }
                     articleReplies {
                       reply {
                         id

data/lib/rumors/api/client/version.rb

@@ -1,7 +1,7 @@
 module Rumors
   module Api
     module Client
-      VERSION = "0.2.0"
+      VERSION = "0.3.0"
     end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rumors-api-client
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Carol H
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-01-02 00:00:00.000000000 Z
+date: 2019-01-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: httparty