From b5fd0b1ef8bb8fbf0c4dceae8e1e5e8d1ad68ee2 Mon Sep 17 00:00:00 2001
From: hnu202401010102 <2250946371@qq.com>
Date: Thu, 26 Dec 2024 14:42:38 +0800
Subject: [PATCH] ADD file via upload

---
 爬虫大作业.py | 65 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 爬虫大作业.py

diff --git a/爬虫大作业.py b/爬虫大作业.py
new file mode 100644
index 0000000..b9a6e40
--- /dev/null
+++ b/爬虫大作业.py
@@ -0,0 +1,65 @@
+import re
+from selenium import webdriver
+from bs4 import BeautifulSoup
+import requests
+import matplotlib.pyplot as plt
+import matplotlib
+headers = {"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/95.0.4638.69 Safari/537.36"}
+driver = webdriver.Edge()
+driver.switch_to.default_content()
+driver.get("https://top.chinaz.com/gongsitop/index_500top.html" )
+next_num =2
+temp_height=0
+money=[]
+name=[]
+while True:
+    driver.execute_script("window.scrollBy(0,10000)")
+    soup = BeautifulSoup(driver.page_source, 'xml')
+    title_nodes=soup.find_all("div",class_="CoListTxt")
+    if next_num ==2:
+        url = "https://top.chinaz.com/gongsitop/index_500top.html"
+    else:
+        url = "https://top.chinaz.com/gongsitop/index_500top_{}.html".format(next_num-1)
+    response = requests.get(url, headers=headers)
+    html = response.text
+    money_=re.findall('注册资本：</span>(.*?)</p>',html)
+    s=0 
+    for title_node in title_nodes:
+        s+=1
+        link_1=title_node.find("a")
+        name.append(link_1.get_text())
+        str_=''
+        for i in range(0,len(money_[s-1])):
+            if money_[s-1][i].isdigit() or money_[s-1][i]=='.':
+                str_+=money_[s-1][i]
+        m=0
+        if str_!='':
+            m=eval(str_)
+        if '亿' in money_[s-1]:
+            m=eval(str_)*10000
+        if "美元" in money_[s-1]:
+            m*=6.71
+        money.append(int(m))
+    if s==0:
+        break
+    check_height = driver.execute_script( "return document.documentElement.scrollTop || window.pageYOffset || document.body.scrollTop;")
+    if check_height == temp_height:
+       driver.get('https://top.chinaz.com/gongsitop/index_500top_'+str(next_num)+'.html')
+       next_num += 1
+    temp_height = check_height
+list_=[]
+for i in range(0,len(name)):
+    list_.append([name[i],money[i]])
+list_.sort(key=lambda x:x[1],reverse=True)
+y=[]
+x=[]
+for i in range(0,20):
+    y.append(list_[i][1]/10000)
+    x.append(list_[i][0])
+matplotlib.rc('font', family='SimHei', weight='bold')
+plt.rcParams['axes.unicode_minus'] = False
+plt.barh(x, y, color='skyblue')
+plt.xlabel('注册资本/万元')
+plt.ylabel('公司')
+plt.title('TOP10公司注册资本')
+plt.show()
\ No newline at end of file