emengweb
用python下载文件的若干种方法汇总。
0. requests标准模板
import requests
url="******"
try:
r=requests.get(url)
r.raise_for_status() #如果不是200，产生异常requests.HTTPError
r.encoding=r.apparent_encoding
print(r.text)
except:
print("爬取失败...")1. 下载图片
import requests
url = 'https://www.python.org/static/img/python-logo@2x.png'
myfile = requests.get(url)
open('PythonImage.png', 'wb').write(myfile.content)用wget
import wget
url = "https://www.python.org/static/img/python-logo@2x.png"
wget.download(url, 'pythonLogo.png')2. 下载重定向的文件
import requests
url = 'https://readthedocs.org/projects/python-guide/downloads/pdf/latest/'
myfile = requests.get(url, allow_redirects=True)
open('hello.pdf', 'wb').write(myfile.content)3. 分块下载大文件
import requests
url = 'https://buildmedia.readthedocs.org/media/pdf/python-guide/latest/python-guide.pdf'
r = requests.get(url, stream = True)
with open("PythonBook.pdf", "wb") as Pypdf:
for chunk in r.iter_content(chunk_size = 1024): # 1024 bytes
if chunk:
Pypdf.write(chunk)4. 并行下载多文件
不并行版本：
import os
import requests
from time import time
from multiprocessing.pool import ThreadPool
def url_response(url):
path, url = url
r = requests.get(url, stream=True)
with open(path, 'wb') as f:
for ch in r:
f.write(ch)
urls = [("Event1", "https://www.python.org/events/python-events/805/"),
("Event2", "https://www.python.org/events/python-events/801/"),
("Event3", "https://www.python.org/events/python-events/790/"),
("Event4", "https://www.python.org/events/python-events/798/"),
("Event5", "https://www.python.org/events/python-events/807/"),
("Event6", "https://www.python.org/events/python-events/807/"),
("Event7", "https://www.python.org/events/python-events/757/"),
("Event8", "https://www.python.org/events/python-user-group/816/")]
start = time()
for x in urls:
url_response(x)
print(f"Time to download: {time() - start}")
# Time to download: 7.306085824966431并行版本，只需改动一行代码ThreadPool(9).imap_unordered(url_response, urls)，时间会大幅度减少：
import os
import requests
from time import time
from multiprocessing.pool import ThreadPool
def url_response(url):
path, url = url
r = requests.get(url, stream=True)
with open(path, 'wb') as f:
for ch in r:
f.write(ch)
urls = [("Event1", "https://www.python.org/events/python-events/805/"),
("Event2", "https://www.python.org/events/python-events/801/"),
("Event3", "https://www.python.org/events/python-events/790/"),
("Event4", "https://www.python.org/events/python-events/798/"),
("Event5", "https://www.python.org/events/python-events/807/"),
("Event6", "https://www.python.org/events/python-events/807/"),
("Event7", "https://www.python.org/events/python-events/757/"),
("Event8", "https://www.python.org/events/python-user-group/816/")]
start = time()
ThreadPool(9).imap_unordered(url_response, urls)
print(f"Time to download: {time() - start}")
# Time to download: 0.00649619102478027345. 使用urllib获取html页面
import urllib.request
# urllib.request.urlretrieve('url', 'path')
urllib.request.urlretrieve('https://www.python.org/', 'PythonOrganization.html')6. python下载视频的神器
you-get，目前you-get所支持的网站包含国内外几十个网站(youtube、twitter、腾讯、爱奇艺、优酷、bilibili等)。
pip install you-get测试一下：
you-get https://www.bilibili.com/video/av52694584/?spm_id_from=333.334.b_686f6d655f706f70756c6172697a65.3youtube-dl也是一个类似的工具。
7. 举个例子
批量下载： NOAA-CIRES 20th Century 2m气温再分析资料。一个个点手会点残，这时候可以借助Python来批量化下载数据。
首先打开页面，按F12查看网页源码：
可以看出，对应下载文件的链接都在div标签下的a标签中，需要将这些链接一一获取然后就可以进行批量化下载了。
# -*- coding: utf-8 -*-
import urllib
from bs4 import BeautifulSoup
rawurl='https://www.esrl.noaa.gov/psd/cgi-bin/db_search/DBListFiles.pl?did=118&tid=40290&vid=2227'
content = urllib.request.urlopen(rawurl).read().decode('ascii') # 获取页面的HTML
soup = BeautifulSoup(content, 'lxml')
url_cand_html=soup.find_all(id='content') # 定位到存放url的标号为content的div标签
list_urls=url_cand_html[0].find_all("a") # 定位到a标签，其中存放着文件的url
urls=[]
for i in list_urls[1:]:
urls.append(i.get('href')) # 取出链接
for i,url in enumerate(urls):
print("This is file"+str(i+1)+" downloading! You still have "+str(142-i-1)+" files waiting for downloading!!")
file_name = "./ncfile/"+url.split('/')[-1] # 文件保存位置+文件名
urllib.request.urlretrieve(url, file_name)并行下载的版本，大家自己先试试，欢迎留言写下你的方案。