yandex_xml 0.0.1 → 1.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 748231943ffb461959ebb4e682283c749a6c8b335b19f9931918a16b69b9b390
-  data.tar.gz: 34ef480451509f9c03ed3d7fb8dcf24ee6475d0cc8a741cf7eb2fdc605846698
+  metadata.gz: d3238a11e12c46cf6f59299e385f84a8cb2a62a468f1621582bcd0caa3f32fd7
+  data.tar.gz: 39dd8d86aac3b0898b087ec7c6fe46a44b8c0bf9613f475481f4db51a436c42a
 SHA512:
-  metadata.gz: 36973c51c195acb148341427845c5e67cf2ae5911ea7a7affbafdf995839bd03bffc353e4ee6f105a7c583d805ff7918574ebc6a693ade0788ea925e420e23eb
-  data.tar.gz: 540f102c72d968a60e8bef98855d2417e05adebc9a422f2e5bf82dc343a29fe439d7b1e1f0c92ff090ea6ed09e76bee8e58b2ce66d8447ba6792bac2a0018fac
+  metadata.gz: 229614f1f3268de67dc965c34304e8d3844def5c6ffe047bbdd96380577be15e9c3e423c7928460ebe2517fd5d272a45f9de68928ed639a965f86220fd297f84
+  data.tar.gz: '0990c6aa7dbb849aed11567219310299a39321b5add8aeb54f8650e12e08d1ab3fff4cba3608c99c847f7722b6700815fde5abd094fc054d7e9ef74e46a49164'

data/lib/yandex_xml.rb

@@ -1,33 +1,108 @@
-# Yandex.XML parser
+# Yandex.XML parser - version 1.0.1 (gem yandex_xml)
 # Author: https://github.com/krdprog/ - Alexey Tsaplin-Kupaysinov
 # License: MIT
 require 'net/http'
+require 'json'
+require 'nori'
 
 # Get data from Yandex.XML service by XML - https://xml.yandex.ru/settings/
+# Yandex.XML Doc: https://yandex.ru/dev/xml/doc/dg/concepts/response-docpage/
+# Yandex Regions: https://yandex.ru/dev/xml/doc/dg/reference/regions-docpage/
 class YandexXml
-  attr_accessor :keyword, :region
+  def initialize(args)
+    @user = args[:user]
+    @key = args[:key]
+    @region = args.fetch(:region, '213') # Moscow
+  end
+
+  # Get data from Yandex.XML and create Hash with result
+  def get(keyword)
+    # Create URL to Yandex.Xml
+    xml_url = create_url(keyword)
+    # Get data XML
+    xml = Net::HTTP.get_response(URI.parse(xml_url)).body
+    # Convert XML to Hash
+    converter = Nori.new(:convert_dashes_to_underscores => true)
+    hash = converter.parse(xml)
+    # Convert Hash to OpenStruct
+    data = JSON.parse(hash.to_json, object_class: OpenStruct)
+    # Create short route to data
+    groups = data.yandexsearch.response.results.grouping.group
+
+    result = {}
+
+    result[:meta] = {
+      keyword: data.yandexsearch.request.query,
+      timestamp: Time.now
+    }
+
+    data_array = []
+
+    groups.each_with_index do |i, index|
+      result_hash = {
+        position: index + 1,
+        url: i.doc.url,
+        domain: i.doc.domain,
+        title: i.doc.title,
+        modtime: i.doc.modtime,
+        size: i.doc.size,
+        charset: i.doc.charset,
+        # passage: i.doc.passages.passage, # TODO: not work
+        passages_type: i.doc.properties._PassagesType,
+        mime_type: i.doc.mime_type,
+        saved_copy_url: i.doc.saved_copy_url,
+        headline: i.doc.headline,
+        turbo_cgi_url: i.doc.properties.TurboCgiUrl,
+        turbo_fallback: i.doc.properties.TurboFallback,
+        turbo_link: i.doc.properties.TurboLink
+      }
+
+      data_array << result_hash
+    end
 
-  def initialize(attributes = {})
-    @user = attributes[:user]
-    @key = attributes[:key]
-    @region = attributes[:region]
+    result[:data] = data_array
+    result
   end
 
-  # Get data from Yandex.XML service to XML format
-  def to_xml
-    create_xml_url
+  # Get 100 urls (Yandex Top-100) from result
+  def get_top100_urls(keyword)
+    hash = get(keyword)
+    struct = JSON.parse(hash.to_json, object_class: OpenStruct)
+
+    urls = []
+    struct.data.each { |i| urls << i.url }
+    urls
+  end
+
+  # Get position in Yandex Top-100 for site by keyword
+  def get_position(keyword, my_domain)
+    hash = get(keyword)
+    struct = JSON.parse(hash.to_json, object_class: OpenStruct)
+
+    domains = []
+    struct.data.each { |i| domains << i.domain.downcase }
+
+    position = 0
+    domains.each_with_index do |domain, i|
+      position = i + 1 if domain == my_domain
+    end
+
+    position
+  end
 
-    @yandex_xml_to_xml = Net::HTTP.get_response(URI.parse(@xml_url)).body
+  # TODO: Get information about limits in Yandex.XML service.
+  def get_limits
+    # code
   end
 
   private
 
   # Create URL for query from Yandex.XML by XML
-  def create_xml_url
+  def create_url(keyword)
     base_url = 'https://yandex.ru/search/xml?'
     tail_url = '&groupby=attr%3Dd.mode%3Ddeep.groups-on-page%3D100.docs-in-group%3D1'
-    query_url = URI.encode_www_form([['user', @user], ['key', @key], ['query', @keyword], ['lr', @region]])
+    query_url = URI.encode_www_form([['user', @user], ['key', @key], ['query', keyword], ['lr', @region]])
 
-    @xml_url = base_url + query_url + tail_url
+    base_url + query_url + tail_url
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: yandex_xml
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 1.1.0
 platform: ruby
 authors:
 - Alexey Tsaplin-Kupaysinov
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-03-01 00:00:00.000000000 Z
+date: 2019-10-06 00:00:00.000000000 Z
 dependencies: []
 description: Get data from Yandex.XML service by XML
 email: info@krdprog.ru
@@ -20,8 +20,10 @@ files:
 homepage: https://github.com/krdprog/yandex_xml
 licenses:
 - MIT
-metadata: {}
-post_install_message: 
+metadata:
+  documentation_uri: https://rubydoc.info/github/krdprog/yandex_xml/
+  homepage_uri: https://github.com/krdprog/yandex_xml
+post_install_message: 'Thanks for installing! Author: @krdprog'
 rdoc_options: []
 require_paths:
 - lib
@@ -36,8 +38,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.8
+rubygems_version: 3.0.3
 signing_key: 
 specification_version: 4
 summary: Ruby application for Yandex.XML service