Class: LinkedinData

Inherits:
Object
  • Object
show all
Includes:
GetRelated, Linkedin, ParseProfile
Defined in:
lib/linkedindata.rb

Instance Method Summary collapse

Methods included from ParseProfile

#addPersonFields, #deleteDuplicatePics, #parseResume

Methods included from GetRelated

#addPointsToProfile, #downloadRelated, #fullProfileList, #getList, #getRelatedProfiles, #relScore

Constructor Details

#initialize(todegree, proxylist) ⇒ LinkedinData

Returns a new instance of LinkedinData.



16
17
18
19
20
21
22
23
 
# File 'lib/linkedindata.rb', line 16

def initialize(todegree, proxylist)
  @proxylist = IO.readlines(proxylist)
  @proxy_list_path = proxylist
  @usedproxies = Hash.new
  @output = Array.new
  @startindex = 10
  @numhops = todegree
end

Instance Method Details

#getByKeywords(search_term) ⇒ Object

Gets all profiles in search results and returns in JSON



83
84
85
86
 
# File 'lib/linkedindata.rb', line 83

def getByKeywords(search_term)
  search(search_term)
  return prepareResults
end

#getSingleProfile(url) ⇒ Object

Gets one profile and the related profiles



77
78
79
80
 
# File 'lib/linkedindata.rb', line 77

def getSingleProfile(url)
  scrape(url, 0)
  return prepareResults
end

#prepareResults ⇒ Object

Gets related profiles then adds relevance scores and any missing keys



70
71
72
73
74
 
# File 'lib/linkedindata.rb', line 70

def prepareResults
  getRelatedProfiles
  deleteDuplicatePics
  return JSON.pretty_generate(relScore(showAllKeys(@output)))
end

#scrape(url, curhops) ⇒ Object

Scrapes and parses individual profile



34
35
36
37
38
39
40
41
42
43
44
 
# File 'lib/linkedindata.rb', line 34

def scrape(url, curhops)
  # Download profile and rescue on error
  begin
    url.gsub!("https", "http")
    profile = Linkedin::Profile.get_profile(url, curhops, @proxylist, @usedproxies)

    # Parse profile if returned and add to output
    @output.concat(parseResume(profile)) if profile
  rescue
  end
end

#search(search_terms) ⇒ Object

Searches for profiles on Google



26
27
28
29
30
31
 
# File 'lib/linkedindata.rb', line 26

def search(search_terms)
  g = GeneralScraper.new("site:linkedin.com/pub", search_terms, @proxy_list_path)
  JSON.parse(g.getURLs).each do |profile|
    scrape(profile, 0)
  end
end

#showAllKeys(data) ⇒ Object

Make sure all keys that occur occur in each item (even if nil)



47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
 
# File 'lib/linkedindata.rb', line 47

def showAllKeys(data)
  # Get all keys
  fields = Set.new
  data.map { |o| fields.merge(o.keys) }

  # Make sure all items have all keys
  datarr = Array.new
  data.each do |d|
    temphash = Hash.new
    fields.each do |f|
      if !d[f]
        temphash[f] = nil
      else
        temphash[f] = d[f]
      end
    end
    datarr.push(temphash)
  end

  return datarr
end