batch_download_img.py

# -*- coding: utf-8 -*-
# requirements: requests
# conda/pip install requests
# pip install -i https://pypi.tuna.tsinghua.edu.cn/simple requests==1.2.3 # version for python2.6, but has problem in SSL
import os
from contextlib import closing
import threading
import time
import sys
import requests
from requests.adapters import HTTPAdapter

# lower python version does not support SNI used for SSL connection
# or throws exception: "hostname doesn't match", refer to https://docs.python-requests.org/zh_CN/latest/community/faq.html
assert sys.version_info >= (2, 7, 9)

s = requests.Session()
s.mount('http://', HTTPAdapter(max_retries=2))
s.mount('https://', HTTPAdapter(max_retries=2))

id_url_file = sys.argv[1]
out_dir = sys.argv[2] # './images'
thread_num = 20
if len(sys.argv) > 3: thread_num = int(sys.argv[3])
# connect + read timeout seconds
timeout = 5

headers = {
    'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36'
}

if not os.path.exists(out_dir):
    os.mkdir(out_dir)

def download(img_url, img_name):
    if os.path.isfile(os.path.join(out_dir, img_name)):
        return
    #with closing(requests.get(img_url, stream=True, headers=headers, timeout=timeout)) as r:
    with closing(s.get(img_url, stream=True, headers=headers, timeout=timeout)) as r:
        rc = r.status_code
        if 299 < rc or rc < 200:
            print('returnCode%s\t%s\t%s' % (rc, img_name, img_url))
            return
        content_length = int(r.headers.get('content-length', '0'))
        if content_length == 0:
            print('size0\t%s' % img_url)
            return
        try:
            with open(os.path.join(out_dir, img_name + ".jpg"), 'wb') as f:
                for data in r.iter_content(1024):
                    f.write(data)
        except:
            print('savefail\t%s' % img_url)

def get_imgurl_generate():
    with open(id_url_file, 'r') as f:
        index = 0
        for line in f:
            index += 1
            if index % 500 == 0:
                print('execute %s line at %s' % (index, time.time()))
            if not line:
                print('line %s is empty "\t"' % index)
                continue
            line = line.strip()
            try:
                imgs = line.split('\t')
                if len(imgs) != 2:
                    print('line %s splite error' % index)
                    continue
                if not imgs[0] or not imgs[1]:
                    print('line %s img is empty' % index)
                    continue
                yield imgs
            except:
                print('line %s can not split by "\t"' % index)


lock = threading.Lock()
def loop(imgs):
    print('thread %s is running...' % threading.current_thread().name)

    while True:
        try:
            with lock:
                #img_url, img_name = next(imgs)
                img_name, img_url = next(imgs)
        except StopIteration:
            break
        try:
            download(img_url, img_name)
        except Exception as e:
            print('exceptfail\t%s\t%s' % (img_url, e))
            time.sleep(1) # seconds
    print('thread %s is end...' % threading.current_thread().name)

img_gen = get_imgurl_generate()

for i in range(0, thread_num):
    t = threading.Thread(target=loop, name='LoopThread %s' % i, args=(img_gen,))
    t.start()