在 Python 中使用多处理或多线程来提高抓取速度

Question

我有一个爬虫代码如下：

import requests
import json
import pandas as pd
import numpy as np
from bs4 import BeautifulSoup
import re
from datetime import datetime

def crawl(id):
    try:
        url = 'https://www.china0001.com.cn/project/{0:06d}.html'.format(id)
        print(url)
        content = requests.get(url).text
        soup = BeautifulSoup(content, 'lxml')
        tbody = soup.find("table", attrs={"id":"mse_new"}).find("tbody", attrs={"class":"jg"})
        tr = tbody.find_all("tr")
        rows = []
        for i in tr[1:]:
            rows.append([j.text.strip() for j in i.findAll("td")])
        out = dict([map(str.strip, y.split('：')) for x in rows for y in x])
        return out

    except AttributeError:
        return False

data = list()
for id in range(699998, 700010):
    print(id)
    res = crawl(id)
    if res:
        data.append(res)

if len(data) > 0:
    df = pd.DataFrame(data)
    df.to_excel('test.xlsx', index = False)

有效，但是当我将 range 间隔变大时需要很长时间，所以我想也许我需要使用 multithreading 或 multiprocessing 或拆分 range() 分成多个块以提高抓取速度，但我不知道该怎么做。

有人可以帮忙吗？提前致谢。

更新：

MAX_WORKERS = 20 #play with it to get an optimal value
ids = list(range(699998, 700050))
workers = min(MAX_WORKERS, len(ids))

data = list()
for id in ids:
    print(id)
    res = crawl(id)
    if res:
        with futures.ThreadPoolExecutor(workers) as executor:
            res = executor.map(crawl, ids)
            data.append(res)

Answer 1

以下是我的处理方式：

from concurrent import futures

MAX_WORKERS = 20 #play with it to get an optimal value
ids = list(range(0, 10000))
workers = min(MAX_WORKERS, len(ids))
data = list()


with futures.ThreadPoolExecutor(workers) as executor:
    res = executor.map(crawl, ids)
    data.append(res)

不过我还没有测试过。但希望它能帮助你找到探索的方向。

在 Python 中使用多处理或多线程来提高抓取速度

Use multiprocessing or multithreading to improve scraping speed in Python

web-crawler

web-scraping

python-multithreading

python-3.x

python-multiprocessing