python - 无法使用发布请求进入下一页

Question

我用 python 编写了一个脚本，用于从网页中获取指向不同文章的不同链接。运行我的脚本后，我可以完美地获取它们。但是，我面临的问题是文章链接遍历多个页面，因为它们的数量很大以适合单个页面。如果我单击下一页按钮，我可以在开发人员工具中看到附加信息，这些信息实际上通过发布请求生成 ajax 调用。由于没有链接到下一页按钮，我无法找到任何方法进入下一页并从那里解析链接。我已经尝试使用 post 请求 和那个 formdata 但它似乎不起作用。我哪里错了？

Link to the landing page containing articles

这是我在单击下一页按钮时使用 chrome 开发工具获得的信息:

GENERAL
=======================================================
Request URL: https://www.ncbi.nlm.nih.gov/pubmed/
Request Method: POST
Status Code: 200 OK
Remote Address: 130.14.29.110:443
Referrer Policy: origin-when-cross-origin

RESPONSE HEADERS
=======================================================
Cache-Control: private
Connection: Keep-Alive
Content-Encoding: gzip
Content-Security-Policy: upgrade-insecure-requests
Content-Type: text/html; charset=UTF-8
Date: Fri, 29 Jun 2018 10:27:42 GMT
Keep-Alive: timeout=1, max=9
NCBI-PHID: 396E3400B36089610000000000C6005E.m_12.03.m_8
NCBI-SID: CE8C479DB3510951_0083SID
Referrer-Policy: origin-when-cross-origin
Server: Apache
Set-Cookie: ncbi_sid=CE8C479DB3510951_0083SID; domain=.nih.gov; path=/; expires=Sat, 29 Jun 2019 10:27:42 GMT
Set-Cookie: WebEnv=1Jqk9ZOlyZSMGjHikFxNDsJ_ObuK0OxHkidgMrx8vWy2g9zqu8wopb8_D9qXGsLJQ9mdylAaDMA_T-tvHJ40Sq_FODOo33__T-tAH%40CE8C479DB3510951_0083SID; domain=.nlm.nih.gov; path=/; expires=Fri, 29 Jun 2018 18:27:42 GMT
Strict-Transport-Security: max-age=31536000; includeSubDomains; preload
Transfer-Encoding: chunked
Vary: Accept-Encoding
X-UA-Compatible: IE=Edge
X-XSS-Protection: 1; mode=block

REQUEST HEADERS
========================================================
Accept: text/html, */*; q=0.01
Accept-Encoding: gzip, deflate, br
Accept-Language: en-US,en;q=0.9
Connection: keep-alive
Content-Length: 395
Content-Type: application/x-www-form-urlencoded; charset=UTF-8
Cookie: ncbi_sid=CE8C479DB3510951_0083SID; _ga=GA1.2.1222765292.1530204312; _gid=GA1.2.739858891.1530204312; _gat=1; WebEnv=18Kcapkr72VVldfGaODQIbB2bzuU50uUwU7wrUi-x-bNDgwH73vW0M9dVXA_JOyukBSscTE8Qmd1BmLAi2nDUz7DRBZpKj1wuA_QB%40CE8C479DB3510951_0083SID; starnext=MYGwlsDWB2CmAeAXAXAbgA4CdYDcDOsAhpsABZoCu0IA9oQCZxLJA===
Host: www.ncbi.nlm.nih.gov
NCBI-PHID: 396E3400B36089610000000000C6005E.m_12.03
Origin: https://www.ncbi.nlm.nih.gov
Referer: https://www.ncbi.nlm.nih.gov/pubmed
User-Agent: Mozilla/5.0 (Windows NT 6.1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36
X-Requested-With: XMLHttpRequest

FORM DATA
========================================================
p$l: AjaxServer
portlets: id=relevancesortad:sort=;id=timelinead:blobid=NCID_1_120519284_130.14.22.215_9001_1530267709_1070655576_0MetA0_S_MegaStore_F_1:yr=:term=%222015%22%5BDate%20-%20Publication%5D%20%3A%20%223000%22%5BDate%20-%20Publication%5D;id=reldata:db=pubmed:querykey=1;id=searchdetails;id=recentactivity
load: yes

到目前为止，这是我的脚本(如果未注释，get 请求将完美运行，但对于第一页):

import requests
from urllib.parse import urljoin
from bs4 import BeautifulSoup

geturl = "https://www.ncbi.nlm.nih.gov/pubmed/?term=%222015%22%5BDate+-+Publication%5D+%3A+%223000%22%5BDate+-+Publication%5D"
posturl = "https://www.ncbi.nlm.nih.gov/pubmed/"

# res = requests.get(geturl,headers={"User-Agent":"Mozilla/5.0"})
# soup = BeautifulSoup(res.text,"lxml")
# for items in soup.select("div.rslt p.title a"):
#     print(items.get("href"))

FormData={
    'p$l': 'AjaxServer',
    'portlets': 'id=relevancesortad:sort=;id=timelinead:blobid=NCID_1_120519284_130.14.22.215_9001_1530267709_1070655576_0MetA0_S_MegaStore_F_1:yr=:term=%222015%22%5BDate%20-%20Publication%5D%20%3A%20%223000%22%5BDate%20-%20Publication%5D;id=reldata:db=pubmed:querykey=1;id=searchdetails;id=recentactivity',
    'load': 'yes'
    }

req = requests.post(posturl,data=FormData,headers={"User-Agent":"Mozilla/5.0"})
soup = BeautifulSoup(req.text,"lxml")
for items in soup.select("div.rslt p.title a"):
    print(items.get("href"))

顺便说一句，当我点击下一页链接时，浏览器中的 url 变为“https://www.ncbi.nlm.nih.gov/pubmed”。

我不想寻求与任何浏览器模拟器相关的任何解决方案。提前致谢。

Answer 1

内容是高度动态的，因此最好使用 selenium 或类似的客户端，但我意识到这不切实际，因为结果数量如此之大。因此，我们必须分析浏览器提交的 HTTP 请求，并用 requests 模拟它们。

下一页的内容通过POST请求加载到/pubmed，post数据是EntrezForm表单的输入字段。表单提交由 js 控制(点击“下一页”按钮时触发)，并通过 .submit() 方法执行。

经过一些检查，我发现了一些有趣的字段:

• EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_Pager.CurrPage 和
  EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_Pager.cPage 指示当前页和下一页。

• EntrezSystem2.PEntrez.DbConnector.Cmd 似乎执行数据库查询。如果我们不提交该字段，结果将不会改变。

• EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_DisplayBar.PageSize 和EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_DisplayBar.PrevPageSize 表示每页的结果数。

根据这些信息，我可以使用下面的脚本获取多个页面。

import requests
from urllib.parse import urljoin
from bs4 import BeautifulSoup

geturl = "https://www.ncbi.nlm.nih.gov/pubmed/?term=%222015%22%5BDate+-+Publication%5D+%3A+%223000%22%5BDate+-+Publication%5D"
posturl = "https://www.ncbi.nlm.nih.gov/pubmed/"

s = requests.session()
s.headers["User-Agent"] = "Mozilla/5.0"

soup = BeautifulSoup(s.get(geturl).text,"lxml")
inputs = {i['name']: i.get('value', '') for i in soup.select('form#EntrezForm input[name]')}

results = int(inputs['EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_ResultsController.ResultCount'])
items_per_page = 100
pages = results // items_per_page + int(bool(results % items_per_page))

inputs['EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_DisplayBar.PageSize'] = items_per_page
inputs['EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_DisplayBar.PrevPageSize'] = items_per_page
inputs['EntrezSystem2.PEntrez.DbConnector.Cmd'] = 'PageChanged'

links = []

for page in range(pages):
    inputs['EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_Pager.CurrPage'] = page + 1
    inputs['EntrezSystem2.PEntrez.PubMed.Pubmed_ResultsPanel.Pubmed_Pager.cPage'] = page

    res = s.post(posturl, inputs)
    soup = BeautifulSoup(res.text, "lxml")

    items = [i['href'] for i in soup.select("div.rslt p.title a[href]")]
    links += items

    for i in items:
        print(i)

我请求每页 100 个项目，因为更高的数字似乎会“破坏”服务器，但您应该能够通过一些错误检查来调整该数字。

最后，链接按降序显示(/29960282，/29960281，...)，所以我认为我们可以在不执行任何 POST 的情况下计算链接要求:

geturl = "https://www.ncbi.nlm.nih.gov/pubmed/?term=%222015%22%5BDate+-+Publication%5D+%3A+%223000%22%5BDate+-+Publication%5D"
posturl = "https://www.ncbi.nlm.nih.gov/pubmed/"

s = requests.session()
s.headers["User-Agent"] = "Mozilla/5.0"
soup = BeautifulSoup(s.get(geturl).text,"lxml")

results = int(soup.select_one('[name$=ResultCount]')['value'])
first_link = int(soup.select_one("div.rslt p.title a[href]")['href'].split('/')[-1])
last_link = first_link - results

links = [posturl + str(i) for i in range(first_link, last_link, -1)]

但不幸的是，结果并不准确。