crawl companies, industries, locations

Gemfile

@@ -61,3 +61,7 @@ end
 
 # Windows does not include zoneinfo files, so bundle the tzinfo-data gem
 gem 'tzinfo-data', platforms: [:mingw, :mswin, :x64_mingw, :jruby]
+
+##
+gem "nokogiri"
+##

Gemfile.lock

@@ -203,6 +203,7 @@ DEPENDENCIES
   jbuilder (~> 2.5)
   listen (>= 3.0.5, < 3.2)
   mysql2 (>= 0.4.4, < 0.6.0)
+  nokogiri
   puma (~> 3.11)
   rails (~> 5.2.4, >= 5.2.4.3)
   sass-rails (~> 5.0)

db/migrate/20200720041935_add_index_to_company.rb

+class AddIndexToCompany < ActiveRecord::Migration[5.2]
+  def change
+    add_index :companies, :name, unique: true
+  end
+end

db/migrate/20200720075150_change_col_area_to_oversea_in_location.rb

+class ChangeColAreaToOverseaInLocation < ActiveRecord::Migration[5.2]
+  def change
+    remove_column :locations, :area, :string
+    add_column :locations, :oversea, :boolean
+  end
+end

db/schema.rb

@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema.define(version: 2020_07_17_014308) do
+ActiveRecord::Schema.define(version: 2020_07_20_075150) do
 
   create_table "applied_jobs", options: "ENGINE=InnoDB DEFAULT CHARSET=utf8", force: :cascade do |t|
     t.bigint "user_id"
@@ -30,6 +30,7 @@ ActiveRecord::Schema.define(version: 2020_07_17_014308) do
     t.string "address"
     t.datetime "created_at", null: false
     t.datetime "updated_at", null: false
+    t.index ["name"], name: "index_companies_on_name", unique: true
   end
 
   create_table "favorites", options: "ENGINE=InnoDB DEFAULT CHARSET=utf8", force: :cascade do |t|
@@ -77,10 +78,10 @@ ActiveRecord::Schema.define(version: 2020_07_17_014308) do
   end
 
   create_table "locations", options: "ENGINE=InnoDB DEFAULT CHARSET=utf8", force: :cascade do |t|
-    t.string "area"
     t.string "city"
     t.datetime "created_at", null: false
     t.datetime "updated_at", null: false
+    t.boolean "oversea"
   end
 
   create_table "locations_jobs", options: "ENGINE=InnoDB DEFAULT CHARSET=utf8", force: :cascade do |t|

lib/tasks/crawler.rake

+task crawl_companies_jobs: :environment do
+  require "open-uri"
+  crawl_companies_and_jobs(3)
+end
+
+task crawl_industries_locations: :environment do
+  require "open-uri"
+  crawl_industries_and_locations
+end
+
+def crawl_companies_and_jobs(page)
+  for i in 1..page
+    company_links, job_links = get_company_and_job_links("https://careerbuilder.vn/viec-lam/tat-ca-viec-lam-trang-#{i}-vi.html")
+    crawl_companies(company_links)
+  end
+end
+
+def get_company_and_job_links(base_link)
+  document = Nokogiri::HTML(open(base_link))
+  companies_xml = document.xpath('//div/a[@class="company-name"]/@href')
+  company_links = companies_xml.map(&:value)
+  jobs_xml = document.xpath('//div/a[@class="job_link"]/@href')
+  job_links = jobs_xml.map(&:value)
+
+  [company_links, job_links]
+end
+
+def crawl_companies(company_links)
+  company_links.each do |link|
+    crawl_company(link)
+  end
+end
+
+def crawl_company(company_link)
+  begin
+    document = Nokogiri::HTML(open(company_link))
+    company_name = document.css(".content .name").text
+    exist = Company.find_by(name: company_name).present?
+    return if exist || company_name.empty?
+    puts company_name
+    company_address = document.css(".content p")[1].text
+    company_description = document.css(".main-about-us").css('.content').text
+    company = Company.create!(name: company_name, address: company_address, description: company_description)
+  rescue => exception
+    return exception
+  end
+end
+
+def crawl_job(job_link)
+  begin
+    document = Nokogiri::HTML(open(job_link))
+    job_company = document.css('')
+    job_title = document.css('.job-desc p.title').text
+    job_salary = document.css('')
+    job_experience = document.css('')
+    job_level = document.css('')
+    job_expiration_date = document.css('')
+    job_description = document.css('')
+  rescue => exception
+    
+  end
+end
+
+def crawl_industries_and_locations
+  document = Nokogiri::HTML(open('https://careerbuilder.vn/viec-lam/tat-ca-viec-lam-vi.html'))
+  industries_xml = document.css('#industry option')
+  industries = industries_xml.map(&:text)
+  locations_xml = document.css('#location option')
+  locations = locations_xml.map(&:text)
+
+
+  industries.each do |industry|
+    exist = Industry.find_by(name: industry).present?
+    break if exist
+    puts industry
+    Industry.create!(name: industry)
+  end
+
+  locations.take(70).each do |location|
+    exist = Location.find_by(city: location).present?
+    break if exist
+    puts location
+    Location.create!(oversea: false, city: location )
+  end
+
+  locations.last(locations.count - 70).each do |location|
+    exist = Location.find_by(city: location).present?
+    break if exist
+    puts location
+    Location.create!(oversea: true, city: location )
+  end
+end
\ No newline at end of file