Support string slot values in CLUSTER SHARD response

[DanielKulesza]

It appears Redis 7.0.14 returns slot values in the CLUSTER SHARDS command response that are represented as strings.

With string slots, nebulex_redis_adapter fails to correctly set up redis cluster shards in the client. This results in consistent MOVED responses.

Redis 7.0.14 responses

<REDACTED_IP>:6379> INFO
# Server
redis_version:7.0.14
...

<REDACTED_IP:6379> CLUSTER SHARDS
 1) 1) "slots"
    2) 1) "4096"
       2) "5119"
    3) "nodes"
    4) 1)  1) "id"
           2) <REDACTED_ID>
           3) "port"
           4) (integer) 11041
           5) "ip"
           6) <REDACTED_IP>
           7) "endpoint"
           8) <REDACTED_IP>
           9) "role"
          10) "replica"
          11) "replication-offset"
          12) (integer) 1743874443
          13) "health"
          14) "online"
...

Redis 7.2.2 responses

localhost:6380> INFO
# Server
redis_version:7.2.2
...

localhost:6380> CLUSTER SHARDS
1) 1) "slots"
   2) 1) (integer) 10923
      2) (integer) 16383
   3) "nodes"
   4) 1)  1) "id"
          2) "fb359c5f566b29a0d3859fee3ca8c8c12f3ece79"
          3) "port"
          4) (integer) 6382
          5) "ip"
          6) "10.10.10.13"
          7) "endpoint"
          8) "10.10.10.13"
          9) "role"
         10) "master"
         11) "replication-offset"
         12) (integer) 266
         13) "health"
         14) "online"
...

Updating the handler for CLUSTER SHARDS response to support both string or integer slot values.

Note:
Possibly related Redis fix: redis/redis@4031a18

[cabol]

LGTM, but let's move the # coveralls-ignore-stop above to skip these new lines (or you can also add a test, but I think it is just a simple thing).

[cabol]

Thanks for the fix. Overall LGTM, I just added a comment. Also, you should rebase since I merged your other PR. Thanks.