def parse_all_content(self, response):
# 有阅读全文的情况，获取全文
tree_node = etree.HTML(response.body)
tweet_item = response.meta['item']
content_node = tree_node.xpath('//*[@id="M_"]/div[1]')[0]
tweet_html = etree.tostring(content_node, encoding='unicode')
tweet_item['content'] = extract_weibo_content(tweet_html)
yield tweet_item
def parse_all_content(self, response):
# 有阅读全文的情况，获取全文
tree_node = etree.HTML(response.body)
tweet_item = response.meta['item']
content_node = tree_node.xpath('//*[@id="M_"]/div[1]')[0]
tweet_html = etree.tostring(content_node, encoding='unicode')
tweet_item['content'] = extract_weibo_content(tweet_html)
if 'location' in tweet_item:
tweet_item['location'] = content_node.xpath(
'.//span[@class="ctt"]/a[last()]/text()')[0]
yield tweet_item
def parse_tweet(self, response):
if response.url.endswith('page=1'):
# 如果是第1页，一次性获取后面的所有页
all_page = re.search(r'/> 1/(\d+)页</div>', response.text)
if all_page:
all_page = all_page.group(1)
all_page = int(all_page)
for page_num in range(2, all_page + 1):
page_url = response.url.replace('page=1',
'page={}'.format(page_num))
yield Request(page_url,
self.parse_tweet,
dont_filter=True,
meta=response.meta)
"""
解析本页的数据
"""
tree_node = etree.HTML(response.body)
tweet_nodes = tree_node.xpath('//div[@class="c" and @id]')
for tweet_node in tweet_nodes:
try:
tweet_item = TweetsItem()
tweet_item['crawl_time'] = int(time.time())
tweet_repost_url = tweet_node.xpath(
'.//a[contains(text(),"转发[")]/@href')[0]
user_tweet_id = re.search(r'/repost/(.*?)\?uid=(\d+)',
tweet_repost_url)
tweet_item['weibo_url'] = 'https://weibo.com/{}/{}'.format(
user_tweet_id.group(2), user_tweet_id.group(1))
tweet_item['user_id'] = user_tweet_id.group(2)
tweet_item['_id'] = '{}_{}'.format(user_tweet_id.group(2),
user_tweet_id.group(1))
create_time_info_node = tweet_node.xpath(
'.//span[@class="ct"]')[-1]
create_time_info = create_time_info_node.xpath('string(.)')
if "来自" in create_time_info:
tweet_item['created_at'] = time_fix(
create_time_info.split('来自')[0].strip())
tweet_item['tool'] = create_time_info.split(
'来自')[1].strip()
else:
tweet_item['created_at'] = time_fix(
create_time_info.strip())
like_num = tweet_node.xpath(
'.//a[contains(text(),"赞[")]/text()')[-1]
tweet_item['like_num'] = int(
re.search('\d+', like_num).group())
repost_num = tweet_node.xpath(
'.//a[contains(text(),"转发[")]/text()')[-1]
tweet_item['repost_num'] = int(
re.search('\d+', repost_num).group())
comment_num = tweet_node.xpath(
'.//a[contains(text(),"评论[") and not(contains(text(),"原文"))]/text()'
)[-1]
tweet_item['comment_num'] = int(
re.search('\d+', comment_num).group())
images = tweet_node.xpath('.//img[@alt="图片"]/@src')
if images:
tweet_item['image_url'] = images[0]
videos = tweet_node.xpath(
'.//a[contains(@href,"https://m.weibo.cn/s/video/show?object_id=")]/@href'
)
if videos:
tweet_item['video_url'] = videos[0]
map_node = tweet_node.xpath('.//a[contains(text(),"显示地图")]')
if map_node:
map_node = map_node[0]
map_node_url = map_node.xpath('./@href')[0]
map_info = re.search(r'xy=(.*?)&', map_node_url).group(1)
tweet_item['location_map_info'] = map_info
repost_node = tweet_node.xpath(
'.//a[contains(text(),"原文评论[")]/@href')
if repost_node:
tweet_item['origin_weibo'] = repost_node[0]
# 检测由没有阅读全文:
all_content_link = tweet_node.xpath(
'.//a[text()="全文" and contains(@href,"ckAll=1")]')
if all_content_link:
all_content_url = self.base_url + all_content_link[
0].xpath('./@href')[0]
yield Request(all_content_url,
callback=self.parse_all_content,
meta={'item': tweet_item},
priority=1)
else:
tweet_html = etree.tostring(tweet_node, encoding='unicode')
tweet_item['content'] = extract_weibo_content(tweet_html)
yield tweet_item
# 抓取该微博的评论信息
comment_url = self.base_url + '/comment/' + tweet_item[
'weibo_url'].split('/')[-1] + '?page=1'
yield Request(url=comment_url,
callback=self.parse_comment,
meta={'weibo_url': tweet_item['weibo_url']})
except Exception as e:
self.logger.error(e)
def parse_tweet(self, response):
if response.url.endswith('page=1'):
# if page 1, get all page number
self.current_page = 1
all_page = re.search(r'/> 1/(\d+)页</div>', response.text)
if all_page:
all_page = all_page.group(1)
all_page = int(all_page)
self.all_page_num = all_page
print("[INFO] Crawling Tweets Page: " + str(self.current_page))
"""
解析本页的数据
"""
tree_node = etree.HTML(response.body)
tweet_nodes = tree_node.xpath('//div[@class="c" and @id]')
for tweet_node in tweet_nodes:
try:
tweet_item = TweetsItem()
tweet_item['crawl_time_utc'] = dt.utcnow(
) # insert datetime timestamp utc
tweet_repost_url = tweet_node.xpath(
'.//a[contains(text(),"转发[")]/@href')[0]
user_tweet_id = re.search(r'/repost/(.*?)\?uid=(\d+)',
tweet_repost_url)
tweet_item['weibo_url'] = 'https://weibo.com/{}/{}'.format(
user_tweet_id.group(2), user_tweet_id.group(1))
tweet_item['user_id'] = user_tweet_id.group(2)
# if tweet_item['user_id']:
# print("[DEBUG] user_id:" + str(tweet_item['user_id']))
# else:
# print("[DEBUG] user_id ERROR")
tweet_item['_id'] = '{}_{}'.format(user_tweet_id.group(2),
user_tweet_id.group(1))
create_time_info_node = tweet_node.xpath(
'.//span[@class="ct"]')[-1]
create_time_info = create_time_info_node.xpath('string(.)')
if "来自" in create_time_info:
tweet_item['created_at'] = time_fix(
create_time_info.split('来自')[0].strip())
self.time_stop_flag = self.time_flag_compare(
tweet_item['created_at']
) # time compare to trigger stop flag
tweet_item['tool'] = create_time_info.split(
'来自')[1].strip()
else:
tweet_item['created_at'] = time_fix(
create_time_info.strip())
self.time_stop_flag = self.time_flag_compare(
tweet_item['created_at']
) # time compare to trigger stop flag
tweet_item['tool'] = ""
like_num = tweet_node.xpath(
'.//a[contains(text(),"赞[")]/text()')[-1]
tweet_item['like_num'] = int(
re.search('\d+', like_num).group())
#print("[DEBUG] like_num:" + str(tweet_item['like_num']))
repost_num = tweet_node.xpath(
'.//a[contains(text(),"转发[")]/text()')[-1]
tweet_item['repost_num'] = int(
re.search('\d+', repost_num).group())
#print("[DEBUG] repost_num:" + str(tweet_item['repost_num']))
comment_num = tweet_node.xpath(
'.//a[contains(text(),"评论[") and not(contains(text(),"原文"))]/text()'
)[-1]
tweet_item['comment_num'] = int(
re.search('\d+', comment_num).group())
#print("[DEBUG] comment_num:" + str(tweet_item['comment_num']))
# Add to grab all images 1) test if multi images link exists 2) if not use the
multi_img_link = tweet_node.xpath(
'.//a[contains(text(),"组图")]/@href')
if multi_img_link:
#print("[DEBUG] multi_img_link:" + multi_img_link[-1])
tweet_item['multi_imgs'] = True
yield Request(url=multi_img_link[-1],
callback=self.parse_multi_images,
meta={'_id': tweet_item['_id']},
priority=1)
else:
tweet_item['multi_imgs'] = False
images = tweet_node.xpath('.//img[@alt="图片"]/@src')
if images:
tweet_item['image_url'] = images[0]
else:
tweet_item['image_url'] = "NA"
videos = tweet_node.xpath(
'.//a[contains(@href,"https://m.weibo.cn/s/video/show?object_id=")]/@href'
)
if videos:
tweet_item['video_url'] = videos[0]
else:
tweet_item['video_url'] = "NA"
map_node = tweet_node.xpath('.//a[contains(text(),"显示地图")]')
if map_node:
map_node = map_node[0]
map_node_url = map_node.xpath('./@href')[0]
map_info = re.search(r'xy=(.*?)&', map_node_url).group(1)
tweet_item['location_map_info'] = map_info
else:
tweet_item['location_map_info'] = "NA"
repost_node = tweet_node.xpath(
'.//a[contains(text(),"原文评论[")]/@href')
if repost_node:
tweet_item['retweet'] = True
tweet_item['origin_weibo'] = repost_node[0]
# crawl original weibo
# origin_weibo_url = self.base_url + '/repost/' + tweet_item['weibo_url'].split('/')[-1] + '?page=1'
# yield Request(url=repost_url, callback=self.parse_repost, meta={'weibo_url': tweet_item['weibo_url']},priority = 2)
else:
tweet_item['retweet'] = False
tweet_item['origin_weibo'] = "NA"
# 检测由没有阅读全文:
all_content_link = tweet_node.xpath(
'.//a[text()="全文" and contains(@href,"ckAll=1")]')
if all_content_link:
all_content_url = self.base_url + all_content_link[
0].xpath('./@href')[0]
yield Request(all_content_url,
callback=self.parse_all_content,
meta={'item': tweet_item},
priority=1)
else:
tweet_html = etree.tostring(tweet_node, encoding='unicode')
tweet_item['content'] = extract_weibo_content(tweet_html)
yield tweet_item
# 抓取该微博的评论信息
comment_url = self.base_url + '/comment/' + tweet_item[
'weibo_url'].split('/')[-1] + '?page=1'
yield Request(url=comment_url,
callback=self.parse_comment,
meta={'weibo_url': tweet_item['weibo_url']},
priority=2)
# Crawl tweet repost
repost_url = self.base_url + '/repost/' + tweet_item[
'weibo_url'].split('/')[-1] + '?page=1'
yield Request(url=repost_url,
callback=self.parse_repost,
meta={'weibo_url': tweet_item['weibo_url']},
priority=2)
except Exception as e:
self.logger.error(e)
# keep looping until hit page with time range limit
self.current_page = self.current_page + 1
if self.time_stop_flag == 0 and self.current_page < (
self.all_page_num + 1) and self.current_page >= 2:
next_page = self.current_page
current_page_str = "page=" + str(next_page - 1)
page_url = response.url.replace(current_page_str,
'page={}'.format(next_page))
yield Request(page_url,
self.parse_tweet,
dont_filter=True,
meta=response.meta,
priority=1)
