Class: Hertools::WebsiteParser

Inherits:
Object
  • Object
show all
Defined in:
lib/hertools/website_parser.rb

Overview

Summary: help to get the website info by a url.

Instance Method Summary collapse

Instance Method Details

#crawl_title_and_favicon_file(url, options = {}) ⇒ Object

Summary: get the title and favicon file by the url of one webpage. Arguments url: the url of a webpage options:

html_parser: %w[httparty nokogiri net_http]
root_path: existing file directory



19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
 
# File 'lib/hertools/website_parser.rb', line 19

def crawl_title_and_favicon_file(url, options = {})
  puts '>>> Parsing the arguments <<<'
  unless parse_url(url)
    puts 'Failed because the bad url!'
    return false
  end
  unless parse_options(options)
    puts 'Failed because the bad options!'
    return false
  end

  puts '>>> Analysing the http response <<<'
  case @html_parser
  when 'nokogiri'
    response = HTTParty.head(@url)
    res = begin
            Nokogiri::HTML(URI.open(url), nil, 'UTF-8')
          rescue StandardError => e
            puts e
            nil
          end
  when 'httparty'
    response = HTTParty.get(@url)
    res = response.body
  else
    response = Net::HTTP.get_response(URI(@url))
    res = response.body.force_encoding("utf-8")
  end
  puts "HttpCode: #{response.code}"

  if res.nil? || res.to_s.empty?
    puts 'No content!'
    @title = @domain_name
    @favicon_url = "#{@index_url}/favicon.ico"
    puts "Use the default favicon url: #{@favicon_url}."
  else
    @title = if nokogiri?
               res.xpath('//head/title')[0]&.content.to_s
             else
               res[%r{<title>\n*(.*)\n*</title>}, 1].to_s
             end
    if @title.empty?
      puts 'Not found the title!'
      puts 'Use the domain name as the title.'
      @title = @domain_name
    end
    unless nokogiri?
      coder = HTMLEntities.new
      @title = coder.decode(@title)
    end
    puts "Title: #{@title}"

    @favicon_url = if nokogiri?
                     favicon_links = res.xpath('//head/link[@rel="icon"]')
                     favicon_links.empty? ? '' : favicon_links[0][:href].to_s
                   else
                     res[/<link rel="icon".*href="([^"]+)/, 1].to_s
                   end
    if @favicon_url.empty?
      puts 'Not found the favicon url!'
      @favicon_url = "#{@index_url}/favicon.ico"
      puts "Use the default favicon url: #{@favicon_url}."
    else
      puts "FaviconUrl: #{@favicon_url}"
      unless @favicon_url.include?('http')
        if @favicon_url.include?('//')
          @favicon_url = "#{@protocol}:#{@favicon_url}"
          puts "Fixed favicon url: #{@favicon_url}"
        elsif @favicon_url.include?('/')
          @favicon_url = "#{@index_url}#{@favicon_url}"
          puts "Fixed favicon url: #{@favicon_url}"
        else
          @favicon_url = "#{@index_url}/favicon.ico"
          puts "Use the default favicon url: #{@favicon_url}."
        end
      end
    end
  end

  if @title.empty? && @favicon_url.empty?
    puts 'Failed because not found the title and favicon url!'
    return false
  end

  identifier = Digest::MD5.hexdigest(@url)
  file_directory_path = "#{@root_path}/#{@domain_name}_#{identifier}"
  puts "FileDirectory: #{file_directory_path}"
  Dir.mkdir(file_directory_path) unless File.directory?(file_directory_path)
  if File.directory?(file_directory_path)
    unless @title.empty?
      info_file_path = "#{file_directory_path}/website_info.txt"
      puts "InfoFilePath: #{info_file_path}"
      open(info_file_path, 'wb') { |f| f << "Title: #{@title}" }
    end

    unless @favicon_url.empty?
      favicon_file_suffix = @favicon_url.split('.').last
      favicon_file_name = Digest::MD5.hexdigest(@favicon_url) + '.' + favicon_file_suffix
      favicon_file_path = "#{file_directory_path}/#{favicon_file_name}"
      puts "FaviconFilePath: #{favicon_file_path}"
      open(favicon_file_path, 'wb') { |f| f << URI.open(@favicon_url).read }
    end
    puts 'Finished!'
    true
  else
    puts 'Failed to create the directory!'
    false
  end
rescue StandardError => e
  puts e
  puts 'Failed because the unexpected exception!'
  false
end