artvee_scraper 0.0.1 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7b6975645d66e1f093ffb1a32bbc516039a25b3baeab5cc842c136bbc6a97911
-  data.tar.gz: a5a112f811c7117970718d956020707927f8b611ef3c659fc2e19ed59507c893
+  metadata.gz: ad2436d422de3a69964170b72a3296a2f27062d05d395c217178ccac22c7fbed
+  data.tar.gz: d24d8a2d9533b58f9ccdfd8a92334f38950a5a49793cd837de42faa4a7fc0acb
 SHA512:
-  metadata.gz: 331227e2dce6602d4518bb364f70e5e273e7185f1f4b4c408e5e8f4ac4e60b6596e38b3e445e2bdf86730bf6aa411b2c4bc518f974c5bc4832176e4ca726fa56
-  data.tar.gz: 48936c5e55f8a103e056820469e5935424d7061fd9cfeb7b8ad96ed1c0eb3d2a562288bb43b10e45293f4c97e9ab2a28261c6547e0f6756585a7b6caefe41661
+  metadata.gz: b43f433cf1462e3dd90cab43d4c53e80190199bc8d8beba89c1f049c5a4f7ff52ca487e19680dea54427c5dfd6959bd8db8c7cc00a0309afbdee80681712a370
+  data.tar.gz: 1870b93e70763b2b1992e623a891a3f5bff5e4d187c5dc6934cb10d005047000533de2937f08a7b1903a7c7ae3eb4a25beb8a668021156b587461f8d33bde9b6

data/lib/artvee_scraper.rb

@@ -1,13 +1,13 @@
-# rubocop:disable Lint/MixedRegexpCaptureTypes
 # frozen_string_literal: true
 
-require 'open-uri'
-require 'nokogiri'
+require_relative 'card'
+require_relative 'http_fetcher'
 
+# Scrapes art data from artvee.com
 class ArtveeScraper
   BASE_URL = 'https://artvee.com/'
   @arts = []
-  @doc = Nokogiri::HTML(URI.open(BASE_URL))
+  @doc = ::Nokogiri::HTML(HttpFetcher.call(BASE_URL))
 
   class << self
     def scrape
@@ -18,55 +18,20 @@ class ArtveeScraper
     private
 
     def populate_arts
-      @doc.search('.product-grid-item.product.woodmart-hover-tiled').each do |card|
-        @arts << {
-          img_url: big_pic_url(card.at('img').attributes['src'].value),
-          title: title(card.at('h3').text),
-          date: date(card.at('h3').text),
-          artist: card.at('.woodmart-product-brands-links a')&.text,
-          artist_details: artist_details(card.at('.woodmart-product-brands-links').text),
-          tag: card.at('.woodmart-product-cats a')&.text
-        }
+      @doc.search('.product-grid-item.product.woodmart-hover-tiled').each do |obj|
+        @arts << art_hash(Card.new(obj))
       end
     end
 
-    def big_pic_url(original_url)
-      original_url.sub(/ftmp/, 'sftb')
-    end
-
-    def title(h3_text)
-      h3_text[..-2].match(/^(?<title>.+?)\s*(\((?<date>[^)]+)\))?$/)[:title]
-    end
-
-    def date(h3_text)
-      h3_text[..-2].match(/^(?<title>.+?)\s*(\((?<date>[^)]+)\))?$/)[:date]
-    end
-
-    def artist_details(div_text)
-      return {} if div_text.split('(').count < 2
-
-      @details = div_text.split('(')[1][0..-2].split(', ')
-      author_life_cycle.merge(nationality)
-    end
-
-    def author_life_cycle
-      return { birth_date: @details.first } if @details.count == 1
-      return { birth_date: @details.last } if @details.last.delete(' ').split(/-|–/).count == 1
-
-      life_cycle_hash(@details.last.delete(' ').split(/-|–/))
-    end
-
-    def life_cycle_hash(life_cycle)
+    def art_hash(card)
       {
-        birth_date: life_cycle.first,
-        passing_date: life_cycle.last
+        img_url: card.img_url,
+        title: card.title,
+        date: card.date,
+        artist: card.artist,
+        artist_details: card.artist_details,
+        tag: card.tag
       }
     end
-
-    def nationality
-      return {} if @details.count == 1
-
-      { nationality: @details.first }
-    end
   end
 end

data/lib/card.rb

@@ -0,0 +1,64 @@
+# frozen_string_literal: true
+
+# Helps the ArtveeScraper class to populate the @arts array
+class Card
+  attr_reader :img_url, :title, :date, :artist, :artist_details, :tag
+
+  def initialize(html_obj)
+    @html_obj = html_obj
+    set_img_url
+    set_date_and_title
+    set_artist
+    set_artist_details
+    set_tag
+  end
+
+  private
+
+  def set_img_url
+    img_src = @html_obj.at('img').attributes['src'].value
+    @img_url = img_src.sub(/ftmp/, 'sftb')
+  end
+
+  def set_date_and_title
+    h3_text = @html_obj.at('h3').text[..-2]
+    date_title = h3_text.match(/^(?<title>.+?)\s*(\((?<date>[^)]+)\))?$/)
+    @title = date_title[:title]
+    @date = date_title[:date]
+  end
+
+  def set_artist
+    @artist = @html_obj.at('.woodmart-product-brands-links a')&.text
+  end
+
+  def set_artist_details
+    div_text = @html_obj.at('.woodmart-product-brands-links').text
+    return {} if div_text.split('(').count < 2
+
+    @details = div_text.split('(')[1][0..-2].split(', ')
+    @artist_details = author_life_cycle.merge(nationality)
+  end
+
+  def set_tag
+    @tag = @html_obj.at('.woodmart-product-cats a')&.text
+  end
+
+  def author_life_cycle
+    return {} unless @details.count >= 1 && @details.last.match?(/\d+/)
+
+    life_cycle_hash(@details.last.delete(' ').split(/-|–/))
+  end
+
+  def life_cycle_hash(life_cycle)
+    {
+      birth_date: life_cycle[0],
+      passing_date: life_cycle[1]
+    }
+  end
+
+  def nationality
+    return {} unless @details.first.match?(/^[a-zA-Z]{2,}/)
+
+    { nationality: @details.first }
+  end
+end

data/lib/http_fetcher.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+# Fetches the content of a given URL and returns its body only if the HTTP request receives a 200 response.
+class HttpFetcher
+  def self.call(url)
+    uri = ::URI.parse(url)
+
+    response = Net::HTTP.start(uri.host, uri.port, use_ssl: uri.scheme == 'https') do |http|
+      http.request(Net::HTTP::Get.new(uri.request_uri))
+    end
+
+    response.body if response.code == '200'
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: artvee_scraper
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.2.0
 platform: ruby
 authors:
 - Leon Siqueira
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-05-25 00:00:00.000000000 Z
+date: 2023-07-14 00:00:00.000000000 Z
 dependencies: []
 description: A gem that gets titles, dates, artist, image URLs, etc. and returns as
   a Hash
@@ -18,6 +18,8 @@ extensions: []
 extra_rdoc_files: []
 files:
 - lib/artvee_scraper.rb
+- lib/card.rb
+- lib/http_fetcher.rb
 homepage: https://github.com/leon-siqueira/artvee-scraper
 licenses:
 - MIT
@@ -30,7 +32,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '0'
+      version: '3.1'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="