在 Java 中压缩，在 Python 中解压缩 - snappy/redis-py-cluster

Question

我正在 python 中为 redis 集群 编写 cron 脚本，并使用 redis-py-cluster 仅从 prod 服务器读取数据。一个单独的 Java 应用程序正在使用 快速压缩和 java 字符串编解码器 utf-8 写入 redis 集群。

我可以读取数据但无法解码。

from rediscluster import RedisCluster
import snappy


host, port ="127.0.0.1", "30001"
startup_nodes = [{"host": host, "port": port}]
print("Trying connecting to redis cluster host=" + host +  ", port=" + str(port))

rc = RedisCluster(startup_nodes=startup_nodes, max_connections=32, decode_responses=True)
print("Connected",  rc)

print("Reading all keys, value ...\n\n")
for key in rc.scan_iter("uidx:*"):
   value = rc.get(key)
   #uncompress = snappy.uncompress(value, decoding="utf-8")
   print(key, value)
   print('\n')

print("Done. exit()")
exit()

decode_responses=False 可以很好地处理评论。但是更改 decode_responses=True 会引发错误。我的猜测是它无法获得正确的解码器。

Traceback (most recent call last):
File "splooks_cron.py", line 22, in <module>
print(key, rc.get(key))
File "/Library/Python/2.7/site-packages/redis/client.py", line 1207, in get
return self.execute_command('GET', name)
File "/Library/Python/2.7/site-packages/rediscluster/utils.py", line 101, in inner
return func(*args, **kwargs)
File "/Library/Python/2.7/site-packages/rediscluster/client.py", line 410, in execute_command
return self.parse_response(r, command, **kwargs)
File "/Library/Python/2.7/site-packages/redis/client.py", line 768, in parse_response
response = connection.read_response()
File "/Library/Python/2.7/site-packages/redis/connection.py", line 636, in read_response
raise e
: 'utf8' codec can't decode byte 0x82 in position 0: invalid start byte

PS：取消注释这一行 uncompress = snappy.uncompress(value, decoding="utf-8") 会出现错误

Traceback (most recent call last):
File "splooks_cron.py", line 27, in <module>
uncompress = snappy.uncompress(value, decoding="utf-8")
File "/Library/Python/2.7/site-packages/snappy/snappy.py", line 91, in uncompress
return _uncompress(data).decode(decoding)
snappy.UncompressError: Error while decompressing: invalid input 

Answer 1

经过几个小时的调试，我终于能够解决这个问题。

我在写入 redis 集群的 Java 代码中使用了xerial/snappy-java 压缩器。有趣的是，在压缩过程中 xerial SnappyOutputStream 在压缩数据的开头添加了一些 offset。就我而言，这看起来像这样

"\x82SNAPPY\x00\x00\x00\x00\x01\x00\x00\x00\x01\x00\x00\x01\xb6\x8b\x06\\******actual data here*****

因此，解压器无法弄清楚。我修改了如下代码，并从值中删除了 offset。现在一切正常。

for key in rc.scan_iter("uidx:*"):
   value = rc.get(key) 
   #in my case offset was 20 and utf-8 is default ecoder/decoder for snappy 
   # https://github.com/andrix/python-snappy/blob/master/snappy/snappy.py
   uncompress_value = snappy.decompress(value[20:])
   print(key, uncompress_value)
   print('\n')