from kafka import KafkaProducer
from kafka import KafkaConsumer
from kafka.structs import TopicPartition
import time

bootstrap_servers = []
class OperateKafka:
def init(self,bootstrap_servers,topic):
self.bootstrap_servers = bootstrap_servers
self.topic = topic

"""生產者"""
def produce(self):
    producer = KafkaProducer(bootstrap_servers=self.bootstrap_servers)
    for i in range(4):
        msg = "msg%d" %i
        producer.send(self.topic,key=str(i),value=msg)
    producer.close()

"""一個消費者消費一個topic"""
def consume(self):
    #consumer = KafkaConsumer(self.topic,auto_offset_reset=‘earliest‘,group_id="testgroup",bootstrap_servers=self.bootstrap_servers)
    consumer = KafkaConsumer(self.topic,bootstrap_servers=self.bootstrap_servers)
    print consumer.partitions_for_topic(self.topic)  #獲取test主題的分區信息
print consumer.topics()  #獲取主題列表
print consumer.subscription()  #獲取當前消費者訂閱的主題
print consumer.assignment()  #獲取當前消費者topic、分區信息
print consumer.beginning_offsets(consumer.assignment()) #獲取當前消費者可消費的偏移量
consumer.seek(TopicPartition(topic=self.topic, partition=0), 1)  #重置偏移量，從第1個偏移量消費
    for message in consumer:
        print ("%s:%d:%d: key=%s value=%s" 
        % (message.topic,message.partition,message.offset, message.key,message.value))

"""一個消費者訂閱多個topic """
def consume2(self):
    consumer = KafkaConsumer(bootstrap_servers=[‘192.168.124.201:9092‘])
consumer.subscribe(topics=(‘TEST‘,‘TEST2‘))  #訂閱要消費的主題
print consumer.topics()
print consumer.position(TopicPartition(topic=‘TEST‘, partition=0)) #獲取當前主題的最新偏移量
for message in consumer:
        print ("%s:%d:%d: key=%s value=%s" % (message.topic, message.partition,
                                      message.offset, message.key,
                                      message.value))
"""消費者(手動拉取消息)"""
def consume3(self):
    consumer = KafkaConsumer(group_id="mygroup",max_poll_records=3,bootstrap_servers=[‘192.168.124.201:9092‘])
consumer.subscribe(topics=(‘TEST‘,‘TEST2‘))
while True:
        message = consumer.poll(timeout_ms=5)   #從kafka獲取消息
        if message:
        print message
        time.sleep(1)
 

def main():
bootstrap_servers = [‘192.168.124.201:9092‘]
topic = "TEST"
operateKafka = OperateKafka(bootstrap_servers,topic)
operateKafka.produce()
#operateKafka.consume()
#operateKafka.consume2()
operateKafka.consume3()
main()

python對kafka的基本操作