Ruby 使用 mechanize 模拟登陆 ruby-china 成功，但是使用 curl 失败了。

flowerwrong · 2014年08月19日 · 最后由 mok 回复于 2015年01月23日 · 11108 次阅读

一.mechanize 代码

require "mechanize"

agent = Mechanize.new
login_page = agent.get "https://ruby-china.org/account/sign_in"
login_form = login_page.forms[1]

username_field = login_form.field_with(:name => "user[login]")
username_field.value = "[email protected]"
password_field = login_form.field_with(:name => "user[password]")
password_field.value = "xxx"

step2_page = agent.submit login_form
tmp_cookie = agent.cookie_jar

agent.cookie_jar = tmp_cookie
index_page = agent.get("https://ruby-china.org/")
re = /^flowerwrong/
puts index_page.body.to_s[0..5000]

请问使用 curl 怎么模拟登陆，我使用的是 curb 库，麻烦看那下问题所在，谢谢。

require 'curb'

USER_NAME = "[email protected]"
PASS = "xxx"

user = {"user\[login\]" => USER_NAME, "user\[password\]" => PASS}
login_url = "https://ruby-china.org/account/sign_in"
c = Curl::Easy.new
c.url = login_url
c.enable_cookies = true
c.cookiejar = "cookies.txt"
c.http_post(c.url, user)

puts c.header_str

http header 打印结果

flowerwrong@flowerwrong:~/dev/ruby/mechanize$ ruby ruby_china.rb 
HTTP/1.1 200 OK
Server: nginx/1.6.0
Date: Tue, 19 Aug 2014 09:13:31 GMT
Content-Type: text/html; charset=utf-8
Transfer-Encoding: chunked
Connection: keep-alive
Strict-Transport-Security: max-age=31536000
X-Frame-Options: SAMEORIGIN
X-XSS-Protection: 1; mode=block
X-Content-Type-Options: nosniff
ETag: "51a9034a6f409ef3e8944e80113415bb"
Cache-Control: max-age=0, private, must-revalidate
X-Request-Id: 4f6f03ef-b598-43dd-9fc2-49d388bb3d5c
X-Runtime: 0.039903

8 个赞

saiga #1 2014年08月19日

mechanize 是根据页面表单来提交的，里面有一些隐藏表单. 你 curl 模拟登录缺少了了 authenticity_token，要先抓页面的 meta 才能登录

flowerwrong #2 2014年08月19日

#1 楼 @saiga 我改了下，先 get 取得 token，然后 post，可是还是不行

require 'curb'
require 'nokogiri'

USER_NAME = "[email protected]"
PASS = "xxx"
login_url = "https://ruby-china.org/account/sign_in"

c = Curl::Easy.new
c.url = login_url
c.http_get

doc = Nokogiri::HTML(c.body_str)
csrf_token = doc.xpath("//meta[@name='csrf-token']/@content")[0].value.strip
c.url = login_url
c.enable_cookies = true
c.cookiejar = "cookies.txt"

user = {
  "user\[login\]" => USER_NAME,
  "user\[password\]" => PASS,
  "authenticity_token" => csrf_token,
  "user\[remember_me\]" => 0,
  "utf8" => "✓",
  "commit" => "登陆"
}
c.http_post(c.url,user)

puts csrf_token
puts "---" * 20
puts c.header_str

打印结果

flowerwrong@flowerwrong:~/dev/ruby/mechanize$ ruby ruby_china.rb 
2O85a3ZBPvdTyo3XQNvvvVlg57VLQ/+vhm3ktyIXXbE=
------------------------------------------------------------
HTTP/1.1 200 OK
Server: nginx/1.6.0
Date: Tue, 19 Aug 2014 10:09:45 GMT
Content-Type: text/html; charset=utf-8
Transfer-Encoding: chunked
Connection: keep-alive
Strict-Transport-Security: max-age=31536000
X-Frame-Options: SAMEORIGIN
X-XSS-Protection: 1; mode=block
X-Content-Type-Options: nosniff
ETag: "c889b4d41d8ddd454f2d022cdc8ef457"
Cache-Control: max-age=0, private, must-revalidate
X-Request-Id: ffce7524-a903-4844-943e-0b54b872ccfc
X-Runtime: 0.042324

saiga #3 2014年08月19日

帮你改了一下可以了...

# -*- coding: utf-8 -*-
require 'curb'
require 'nokogiri'

USER_NAME = "xx"
PASS = "xxx"
login_url = "https://ruby-china.org/account/sign_in"

c = Curl::Easy.new
# cookie 一开始就要开启，auth_token 会记录在cookie的
c.enable_cookies = true
c.cookiejar = "cookies.txt"
c.url = login_url
c.http_get

doc = Nokogiri::HTML(c.body_str)
csrf_token = doc.xpath("//meta[@name='csrf-token']/@content")[0].value.strip


user = {
  "user\[login\]" => USER_NAME,
  "user\[password\]" => PASS,
  "authenticity_token" => csrf_token,
  "user\[remember_me\]" => 0,
  "utf8" => "✓",
  "commit" => "登陆"
}

# 看api......
c.http_post c.url, user.map { |k, v| Curl::PostField.content k, v }

puts csrf_token
puts "---" * 20
puts c.header_str

2 个赞

flowerwrong #4 2014年08月19日

#3 楼 @saiga thx

ericguo #5 2014年08月19日

#3 楼 @saiga 多谢，我也学了一招，顺便分享一下如何在 Windows 环境下安装 curb

flowerwrong #6 2014年08月19日

补充 watir 版本

# -*- coding: utf-8 -*-
require 'watir-webdriver'
b = Watir::Browser.new

USER_NAME = "[email protected]"
PASS = "xxx"
login_url = "https://ruby-china.org/account/sign_in"



b.goto login_url
b.text_field(:name => 'user[login]').set USER_NAME
b.text_field(:name => 'user[password]').set PASS

b.button(:name => 'commit').click
puts b.text

tim_lang #7 2014年08月22日

curl -X POST 'https://ruby-china.org/account/sign_in.json' -d 'username=Tim_Lang&password=password'

flowerwrong #8 2014年08月29日

#7 楼 @Tim_Lang 如果是 GBK 网页单纯的 crul 怎么处理乱码呢？

mok #9 2015年01月22日

==========method=get========== HTTP/1.1 200 OK Date: Thu, 22 Jan 2015 09:55:51 GMT Server: IBM_HTTP_Server Pragma: no-cache Expires: Thu, 01 Jan 1970 00:00:00 GMT Set-Cookie: JSESSIONID=0000nnKUCci1wVHUkTPEdmCSNR7:-1; Path=/ Cache-Control: no-store, no-cache=set-cookie Keep-Alive: timeout=10, max=100 Connection: Keep-Alive Transfer-Encoding: chunked Content-Type: text/html; charset=UTF-8 Content-Language: zh-CN ==========method=post========== HTTP/1.1 500 Internal Server Error Date: Thu, 22 Jan 2015 09:55:51 GMT Server: IBM_HTTP_Server Pragma: no-cache Expires: Thu, 01 Jan 1970 00:00:00 GMT $WSEP: Content-Length: 13 Set-Cookie: JSESSIONID=0000ugRJQAmTQ9D1dC56Jswkoei:-1; Path=/ Cache-Control: no-store, no-cache=set-cookie Connection: close Content-Type: text/html;charset=UTF-8 Content-Language: zh-CN

我 get 拿到的 cookies 和 post 拿到的 cookie 不一致导致 500 错误如何才能让他们保持一致

mok #10 2015年01月22日

如果我直接用 login_url = "*********" c = Curl::Easy.new c.url = login_url c.http_get puts c.header_str

拿到 response header HTTP/1.1 200 OK Date: Thu, 22 Jan 2015 10:03:14 GMT Server: IBM_HTTP_Server Pragma: no-cache Expires: Thu, 01 Jan 1970 00:00:00 GMT Set-Cookie: JSESSIONID=0000OiEy8TiTGJ0t_ikYEP4ExC6:-1; Path=/ Cache-Control: no-store, no-cache=set-cookie Transfer-Encoding: chunked Content-Type: text/html; charset=UTF-8 Content-Language: zh-CN

显然比直接 Curl.get(login_url) 少了下面两个键值 Keep-Alive: timeout=10, max=100 Connection: Keep-Alive

麻烦大神们帮忙解答下，谢谢！！

mok #11 2015年01月22日

@saiga

mok #12 2015年01月22日

以上问题解决了接下来还有一个问题我请求的页面中又验证码，然后通过 uri = URI('http://xxx/image.jsp') open("logo.jpg","wb"){|f|f.write(Net::HTTP.get(uri))} 下载到本地并 RTesseract 识别出图片中的字符串然后提交的时候老是提示验证码错误我肯定的是识别没有问题问题是不是就是 cookie 问题请问如何解决

mok #13 2015年01月22日

已经解决了谢谢

mok #14 2015年01月23日

form data 的内容如下 origins:CTU dests:HKG departs:2015-01-30 origins:HKG dests:BKK departs:2015-01-31 origins:BKK dests:HKG departs:2015-02-01 origins:HKG dests:CTU departs:2015-02-02 origins: dests: departs: 我通过如下方式来处理表单内容 search_form_data = { 'origins' => 'CTU', 'dests' => 'HKG', 'departs' => '2015-01-30',

'origins' => 'HKG', 'dests' => 'BKK', 'departs' => '2015-01-31',

'origins' => 'BKK', 'dests' => 'HKG', 'departs' => '2015-02-01',

'origins' => 'HKG', 'dests' => 'CTU', 'departs' => '2015-02-02' } puts search_form_data # => {"origins"=>"HKG", "dests"=>"CTU", "departs"=>"2015-02-02"} 请问这种情况下如何提交？ @saiga

需要登录后方可回复, 如果你还没有账号请注册新账号

8 个赞

共收到 14 条回复

收到新回复，点击立即加载