我想处理180个请求/15分钟的搜索api速率限制。我想到的第一个解决方案是检查报头中剩余的请求并等待900秒。请看下面的代码片段:
results = search_interface.cursor(search_interface.search, q=k, lang=lang, result_type=result_mode)
while True:
try:
tweet = next(results)
if limit_reached(search_interface):
sleep(900)
self.writer(tweet)
def limit_reached(search_interface):
remaining_rate = int(search_interface.get_lastfunction_header('X-Rate-Limit-Remaining'))
return remaining_rate <= 2
但是看起来,头信息在到达剩下的两个请求后并没有重置为180。
我想到的第二个解决方案是处理速率限制的python异常,并等待剩余的时间:
results = search_interface.cursor(search_interface.search, q=k, lang=lang, result_type=result_mode)
while True:
try:
tweet = next(results)
self.writer(tweet)
except TwythonError as inst:
logger.error(inst.msg)
wait_for_reset(search_interface)
continue
except StopIteration:
break
def wait_for_reset(search_interface):
reset_timestamp = int(search_interface.get_lastfunction_header('X-Rate-Limit-Reset'))
now_timestamp = datetime.now().timestamp()
seconds_offset = 10
t = reset_timestamp - now_timestamp + seconds_offset
logger.info('Waiting {0} seconds for Twitter rate limit reset.'.format(t))
sleep(t)
但是有了这个解决方案,我收到这个消息INFO: reset dropped connection: api.twitter.com",循环将不会继续与生成器的最后一个元素。有人遇到过同样的问题吗?
致意。
我的建议是限制自己的速度(假设你经常达到极限…)
QUERY_PER_SEC = 15*60/180.0 #180 per 15 minutes
#~5 seconds per query
class TwitterBot:
last_update=0
def doQuery(self,*args,**kwargs):
tdiff = time.time()-self.last_update
if tdiff < QUERY_PER_SEC:
time.sleep(QUERY_PER_SEC-tdiff)
self.last_update = time.time()
return search_interface.cursor(*args,**kwargs)