Class: IndeedScraper2022

Inherits:
Object
  • Object
show all
Defined in:
lib/indeed_scraper2022.rb

Direct Known Subclasses

IS22Plus

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(url = 'https://uk.indeed.com/?r=us', q: '', location: '', headless: true, cookies: nil, debug: false) ⇒ IndeedScraper2022

Returns a new instance of IndeedScraper2022.



29
30
31
32
33
34
35
36
37
38
39
 
# File 'lib/indeed_scraper2022.rb', line 29

def initialize(url='https://uk.indeed.com/?r=us', q: '', location: '',
               headless: true, cookies: nil, debug: false)

  @debug = debug
  @url_base, @q, @location = url, q, location
  @headless, @cookies = headless, cookies

  fw = FerrumWizard.new( headless: @headless, cookies: @cookies, debug: @debug)
  @browser = fw.browser

end

Instance Attribute Details

#browser ⇒ Object (readonly)

Returns the value of attribute browser.



27
28
29
 
# File 'lib/indeed_scraper2022.rb', line 27

def browser
  @browser
end

Instance Method Details

#page(n) ⇒ Object 



166
167
168
169
170
171
172
173
174
 
# File 'lib/indeed_scraper2022.rb', line 166

def page(n)

  if n < 1 or n > @results.length then
    raise IndeedScraper2022Err, 'Invalid page no.'
  end

  url = @results[n-1][:link]
  fetchjob(url)
end

#results ⇒ Object

returns an array containing the job search result



43
44
45
 
# File 'lib/indeed_scraper2022.rb', line 43

def results()
  @results
end

#search(q: @q, location: @location, start: nil) ⇒ Object 



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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
 
# File 'lib/indeed_scraper2022.rb', line 47

def search(q: @q, location: @location, start: nil)
  puts 'inside search' if @debug
  url = @url_base
  url += 'start=' + start if start
  puts 'url: ' + url.inspect if @debug

  @browser.goto(url)
  #@browser.network.wait_for_idle
  puts 'sleeping for 4 seconds' if @debug
  sleep 4

  if q.length > 1 then

    input = @browser.at_xpath("//input[@name='q']")

    # select any existing text and overwrite it
    input.focus.type(:home); sleep 0.2
    input.focus.type(:shift, :end); sleep 0.2
    input.focus.type(q); sleep 0.2
  end

  if location.length > 1 then

    input2 = @browser.at_xpath("//input[@name='l']")

    # select any existing text and overwrite it
    input2.focus.type(:home); sleep 0.2
    input2.focus.type(:shift, :end); sleep 0.2
    input2.focus.type(location); sleep 0.2

  end

  button = @browser.at_xpath("//button[@type='submit']")
  button.click
  #@browser.network.wait_for_idle
  puts 'sleeping for 2 seconds' if @debug
  sleep 2

  doc2 = Nokogiri::XML(@browser.body)
  File.write '/tmp/body.txt', doc2.to_s if @debug

  a2 = doc2.root.xpath  "//li/div[div/div/div/div/table/tbody/tr/td/div/h2/a]"
  puts 'a2: ' + a2.length.inspect if @debug

  @a2 = a2.map {|x| Rexle.new x.to_s }

  @results = @a2.map do |doc|

    div = doc.element("div[@class='cardOutline']/div[@class='slider"  \
        "_container']/div[@class='slider_list']/div[@class='sl"  \
        "ider_item']/div[@class='job_seen_beacon']")

    td = div.element("table[@class='jobCard_mainContent']/tbo"  \
        "dy/tr/td[@class='resultContent']")

    # job title (e.g. Software Developer)
    job = td.element("div[@class='tapItem-gutter']/h2[@"  \
        "class='jobTitle-color-purple']/a")
    href = job.attributes[:href]
    jobtitle = job.element("span")&.text

    puts 'jobtitle: ' + jobtitle.inspect if @debug

    sal = td.element("div[@class='metadataContainer']/"  \
        "div[@class='salary-snippet-container']")

    salary = if sal then
      sal_e = sal.element("div[@class='attribute_snippet']")
      if sal_e then
        sal_e.texts[0]
      else
        sal_e2 = sal.element("div[@class='salary-snippet']/span")
        sal_e2 ? sal_e2.text : ''
      end
    else
      ''
    end

    puts 'salary: ' + salary.inspect if @debug
    div1 = td.element("div[@class='companyInfo']")

    # company name (e.g. Coda Octopus Products Ltd)
    coname = div1.element("span[@class='companyName']")
    puts 'coname: ' + coname.text.inspect if @debug
    company_name = coname.text.to_s.strip.length > 1 ? coname.text : coname.element('a').text

    # company location (e.g. Edinburgh)
    location = div1.element("div[@class='companyLocation']")&.text
    tbody = div.element("table[@class='jobCardShelfContainer']/tbody")

    div3 = tbody.element("tr[@class='underShelfFooter']/td/di"  \
        "v[@class='result-footer']")

    # job (e.g. Our products are primarily written in C#, using...)
    advert_items = div3.xpath("div[@class='job-snippet']/ul/li/text()")
    jobsnippet = if advert_items.any? then
      advert_items.join("\n")
    else
      div3.element("div[@class='job-snippet']").text
    end

    # visually (e.g. Posted 14 days ago)
    dateposted =  div3.element("span[@class='date']")&.texts
    date = (Date.today - dateposted.first.to_i).to_s if dateposted

    {
      link:  @url_base.sub(/\/[^\/]+$/,'') \
        + href.gsub(/&amp;/,'&'),
      title: jobtitle,
      salary: salary,
      company: company_name,
      location: location,
      jobsnippet: jobsnippet,
      date: date
    }

  end
end