Better support for working with XREADGROUP and XREAD replies

`XREADGROUP` and `XREAD` replies have the same structure. if we use `std::vector< aedis::resp3::node< std::string > >` for their response it will be very difficult to work with their replies.
It would be great if you could add some library support to make working with Redis streams easier.
```BASH
> XREAD COUNT 2 STREAMS stream_1 stream_2 0
1) 1) "stream_1"
2) 1) 1) 1519073278252-0
2) 1) "key_1"
2) "value_1"
3) "key_2"
4) "value_2"
2) 1) 1519073279157-0
2) 1) "key_1"
2) "value_1"
3) "key_2"
4) "value_2"
3) 1) 1519073279157-0
2) 1) "key_1"
2) "value_1"
3) "key_2"
4) "value_2"
2) 1) "stream_2"
2) 1) 1) 1519073278252-0
2) 1) "key_1"
2) "value_1"
2) 1) 1519073279157-0
2) 1) "key_1"
2) "value_1"
```

Start by examining the existing handling of XREADGROUP and XREAD replies, especially the std::vector> representation. Compare the nested reply structures shown in the issue and determine the library support and completion criteria needed for easier Redis stream access. No source files or tests are named, so repository exploration is required first.

Written by the indexing model from the issue text.