def search_keyword(keyword, keyword_id):
crawler.info('We are searching keyword "{}"'.format(keyword))
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < LIMIT:
cur_url = URL.format(encode_keyword, cur_page)
# current only for login, maybe later crawling page one without login
search_page = get_page(cur_url, auth_level=2)
if not search_page:
crawler.warning('No search result for keyword {}, the source page is {}'.format(keyword, search_page))
return
search_list = parse_search.get_search_info(search_page)
# Because the search results are sorted by time, if any result has been stored in mysql,
# We need not crawl the same keyword in this turn
for wb_data in search_list:
rs = WbDataOper.get_wb_by_mid(wb_data.weibo_id)
KeywordsDataOper.insert_keyword_wbid(keyword_id, wb_data.weibo_id)
if rs:
crawler.info('Weibo {} has been crawled, skip it.'.format(wb_data.weibo_id))
continue
else:
WbDataOper.add_one(wb_data)
app.send_task('tasks.user.crawl_person_infos', args=(wb_data.uid,), queue='user_crawler',
routing_key='for_user_info')
if cur_page == 1:
cur_page += 1
elif 'noresult_tit' not in search_page:
cur_page += 1
else:
crawler.info('Keyword {} has been crawled in this turn'.format(keyword))
return
def determine(weibo_datum, timeafter):
weibo_time = time.mktime(
time.strptime(weibo_datum.create_time, '%Y-%m-%d %H:%M'))
if weibo_time < timeafter:
return False
if WbDataOper.get_wb_by_mid(weibo_datum.weibo_id):
return False
return True
def determine(weibo_datum, timeafter):
weibo_time = time.mktime(
time.strptime(weibo_datum.create_time, '%Y-%m-%d %H:%M'))
if weibo_time < timeafter:
return False
if WbDataOper.get_wb_by_mid(weibo_datum.weibo_id):
return False
return True
def search_items_v2(keyword, keyword_id, date_item):
search_time_list = [
"{}-{}:{}-{}".format(d, t, d, t + 2)
for d, t in itertools.product([date_item], TIME_LIIT)
]
for s_time in search_time_list:
crawler.info('We are searching keyword "{}", {}'.format(
keyword, s_time))
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < LIMIT:
cur_url = MAX_URL.format(encode_keyword, cur_page, s_time)
# current only for login, maybe later crawling page one without login
search_page = get_page(cur_url, auth_level=1, need_proxy=True)
if "您可以尝试更换关键词，再次搜索" in search_page:
break
if not search_page:
crawler.warning(
'No search result for keyword {}, the source page is {}'.
format(keyword, search_page))
cur_page += 1
continue
# return
search_list = parse_search.get_search_info(search_page)
if cur_page == 1:
cur_page += 1
elif 'noresult_tit' not in search_page:
cur_page += 1
else:
crawler.info(
'Keyword {} has been crawled in this turn'.format(keyword))
return
# Because the search results are sorted by time, if any result has been stored in mysql,
# We don't need to crawl the same keyword in this turn
for wb_data in search_list:
# print(wb_data)
rs = WbDataOper.get_wb_by_mid(wb_data.weibo_id)
KeywordsDataOper.insert_keyword_wbid(keyword_id,
wb_data.weibo_id)
# todo incremental crawling using time
if rs:
crawler.info('Weibo {} has been crawled, skip it.'.format(
wb_data.weibo_id))
continue
else:
WbDataOper.add_one(wb_data)
# todo: only add seed ids and remove this task
app.send_task('tasks.user.crawl_person_infos',
args=(wb_data.uid, ),
queue='user_crawler',
routing_key='for_user_info')
def test_weibodata_oper(self):
db_session.execute("insert into {} ({}.weibo_id) values ('".format(
weibo_data.name, weibo_data.name) + FAKE_ID + "')")
assert WbDataOper.get_wb_by_mid(FAKE_ID) is not None
assert len(WbDataOper.get_weibo_comment_not_crawled()) == 1
assert len(WbDataOper.get_weibo_repost_not_crawled()) == 1
WbDataOper.set_weibo_comment_crawled(FAKE_ID)
WbDataOper.set_weibo_repost_crawled(FAKE_ID)
assert len(WbDataOper.get_weibo_comment_not_crawled()) == 0
assert len(WbDataOper.get_weibo_repost_not_crawled()) == 0
def test_weibodata_oper(self):
db_session.execute("insert into {} ({}.weibo_id) values ('".format(weibo_data.name, weibo_data.name)
+ FAKE_ID + "')")
assert WbDataOper.get_wb_by_mid(FAKE_ID) is not None
assert len(WbDataOper.get_weibo_comment_not_crawled()) == 1
assert len(WbDataOper.get_weibo_repost_not_crawled()) == 1
WbDataOper.set_weibo_comment_crawled(FAKE_ID)
WbDataOper.set_weibo_repost_crawled(FAKE_ID)
assert len(WbDataOper.get_weibo_comment_not_crawled()) == 0
assert len(WbDataOper.get_weibo_repost_not_crawled()) == 0
def search_keyword(keyword, keyword_id):
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < LIMIT:
cur_url = URL.format(encode_keyword, cur_page)
if cur_page == 1:
search_page = get_page(cur_url, auth_level=1)
else:
search_page = get_page(cur_url, auth_level=2)
if not search_page:
crawler.warning(
'No result for keyword {}, the source page is {}'.format(
keyword, search_page))
return
search_list = parse_search.get_search_info(search_page)
# Because the search results are sorted by time, if any result has been stored in mysql,
# we need not crawl the same keyword in this turn
for wb_data in search_list:
rs = WbDataOper.get_wb_by_mid(wb_data.weibo_id)
if rs:
crawler.info(
'Keyword {} has been crawled in this turn'.format(keyword))
return
else:
WbDataOper.add_one(wb_data)
KeywordsDataOper.insert_keyword_wbid(keyword_id,
wb_data.weibo_id)
# send task for crawling user info
app.send_task('tasks.user.crawl_person_infos',
args=(wb_data.uid, ),
queue='user_crawler',
routing_key='for_user_info')
if cur_page == 1:
cur_page += 1
elif 'page next S_txt1 S_line1' in search_page:
cur_page += 1
else:
crawler.info(
'Keyword {} has been crawled in this turn'.format(keyword))
return
def search_keyword_topic(keyword, keyword_id, start_time='', end_time=''):
crawler.info(
'We are crawling weibo topic content with keyword "{}"'.format(
keyword))
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < LIMIT:
cur_url = URL.format(encode_keyword, start_time, end_time, cur_page)
search_page = get_page(cur_url, auth_level=2)
if not search_page:
crawler.info(
'No such result for keyword {}, the source page is {}'.format(
keyword, search_page))
return
search_list = parse_topic.get_search_info(search_page)
if cur_page == 1:
cur_page += 1
elif '您可以尝试更换关键词' not in search_page:
cur_page += 1
else:
crawler.info(
'Keyword {} has been crawled in this turn'.format(keyword))
return
for wb_data in search_list:
rs = WbDataOper.get_wb_by_mid(wb_data.weibo_id)
KeywordsDataOper.insert_keyword_wbid(keyword_id, wb_data.weibo_id)
if rs:
crawler.info('Weibo {} has been crawled, skip it.'.format(
wb_data.weibo_id))
continue
else:
WbDataOper.add_one(wb_data)
app.send_task('tasks.user.crawl_person_infos',
args=(wb_data.uid, ),
queue='user_crawler',
routing_key='for_user_info')
def search_keyword(keyword, keyword_id):
crawler.info('We are searching keyword "{}"'.format(keyword))
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < LIMIT:
cur_url = URL.format(encode_keyword, cur_page)
# current only for login, maybe later crawling page one without login
search_page = get_page(cur_url, auth_level=2)
if not search_page:
crawler.warning('No search result for keyword {}, the source page is {}'.format(keyword, search_page))
return
search_list = parse_search.get_search_info(search_page)
if cur_page == 1:
cur_page += 1
elif 'noresult_tit' not in search_page:
cur_page += 1
else:
crawler.info('Keyword {} has been crawled in this turn'.format(keyword))
return
# Because the search results are sorted by time, if any result has been stored in mysql,
# We don't need to crawl the same keyword in this turn
for wb_data in search_list:
rs = WbDataOper.get_wb_by_mid(wb_data.weibo_id)
KeywordsDataOper.insert_keyword_wbid(keyword_id, wb_data.weibo_id)
# todo incremental crawling using time
if rs:
crawler.info('Weibo {} has been crawled, skip it.'.format(wb_data.weibo_id))
continue
else:
WbDataOper.add_one(wb_data)
# todo: only add seed ids and remove this task
app.send_task('tasks.user.crawl_person_infos', args=(wb_data.uid,), queue='user_crawler',
routing_key='for_user_info')
