ruby 爬虫爬取拉钩网职位信息，产生词云报告

Posted 2020-11-20 wf0117

思路：1.获取拉勾网搜索到职位的页数

　　

2.调用接口获取职位id

　　

3.根据职位id访问页面，匹配出关键字

　　

 

url访问采用unirest，由于拉钩反爬虫，短时间内频繁访问会被限制访问，所以没有采用多线程，而且每个页面访问时间间隔设定为10s，通过nokogiri解析页面，正则匹配只获取技能要求中的英文单词，可能存在数据不准确的情况

数据持久化到excel中，采用ruby erb生成word_cloud报告

爬虫代码：

require ‘unirest‘
require ‘uri‘
require ‘nokogiri‘
require ‘json‘
require ‘win32ole‘

@position = ‘测试开发工程师‘
@city = ‘杭州‘

# 页面访问
def query_url(method, url, headers:{}, parameters:nil)
  case method
    when :get
      Unirest.get(url, headers:headers).body
    when :post
      Unirest.post(url, headers:headers, parameters:parameters).body
  end
end

# 获取页数
def get_page_num(url)
  html = query_url(:get, url).force_encoding(‘utf-8‘)
  html.scan(/<span class="span totalNum">(d+)</span>/).first.first
end

# 获取每页显示的所有职位的id
def get_positionsId(url, headers:{}, parameters:nil)
  response = query_url(:post, url, headers:headers, parameters:parameters)
  positions_id = Array.new
  response[‘content‘][‘positionResult‘][‘result‘].each{|i| positions_id << i[‘positionId‘]}
  positions_id
end

# 匹配职位英文关键字
def get_skills(url)
  puts "loading url: #{url}"
  html = query_url(:get, url).force_encoding(‘utf-8‘)
  doc = Nokogiri::HTML(html)
  data = doc.css(‘dd.job_bt‘)
  data.text.scan(/[a-zA-Z]+/)
end

# 计算词频
def word_count(arr)
  arr.map!(&:downcase)
  arr.select!{|i| i.length>1}
  counter = Hash.new(0)
  arr.each { |k| counter[k]+=1 }
  # 过滤num=1的数据
  counter.select!{|_,v| v > 1}
  counter2 = counter.sort_by{|_,v| -v}.to_h
  counter2
end

# 转换
def parse(hash)
  data = Array.new
  hash.each do |k,v|
    word = Hash.new
    word[‘name‘] = k
    word[‘value‘] = v
    data << word
  end
  JSON data
end

# 持久化数据
def save_excel(hash)
  excel = WIN32OLE.new(‘Excel.Application‘)
  excel.visible = false
  workbook = excel.Workbooks.Add()
  worksheet = workbook.Worksheets(1)
  # puts hash.size
  (1..hash.size+1).each do |i|
    if i == 1
      # puts "A#{i}:B#{i}"
      worksheet.Range("A#{i}:B#{i}").value = [‘关键词‘, ‘频次‘]
    else
      # puts i
      # puts hash.keys[i-2], hash.values[i-2]
      worksheet.Range("A#{i}:B#{i}").value = [hash.keys[i-2], hash.values[i-2]]
    end
  end
  excel.DisplayAlerts = false
  workbook.saveas(File.dirname(__FILE__)+‘lagouspider.xls‘)
  workbook.saved = true
  excel.ActiveWorkbook.Close(1)
  excel.Quit()
end

# 获取页数
url = URI.encode("https://www.lagou.com/jobs/list_#@position?city=#@city&cl=false&fromSearch=true&labelWords=&suginput=")
num = get_page_num(url).to_i
puts "存在 #{num} 个信息分页"

skills = Array.new
(1..num).each do |i|
  puts "定位在第#{i}页"
  # 获取positionsid
  url2 = URI.encode("https://www.lagou.com/jobs/positionAjax.json?city=#@city&needAddtionalResult=false")
  headers = {Referer:url, ‘User-Agent‘:i%2==1?‘Mozilla/5.0‘:‘Chrome/67.0.3396.87‘}
  parameters = {first:(i==1), pn:i, kd:@position}
  positions_id = get_positionsId(url2, headers:headers, parameters:parameters)
  positions_id.each do |id|
    # 访问具体职位页面,提取英文技能关键字
    url3 = "https://www.lagou.com/jobs/#{id}.html"
    skills.concat get_skills(url3)
    sleep 10
  end
end

count = word_count(skills)
save_excel(count)
@data = parse(count)

 

效果展示：