[refactor](reader) refactor broker_file_reader to get _client in the constructor

[luozenglin]

Proposed changes

Issue Number: close #xxx

related pr: #15874

Problem summary

In #15874, @HappenLee's comment is valid. There is no problem getting the client in the constructor and the logic is much clearer.

Checklist(Required)

1. Does it affect the original behavior:
   • Yes
   • No
   • I don't know
2. Has unit tests been added:
   • Yes
   • No
   • No Need
3. Has document been added or modified:
   • Yes
   • No
   • No Need
4. Does it need to update dependencies:
   • Yes
   • No
5. Are there any changes that cannot be rolled back:
   • Yes (If Yes, please explain WHY)
   • No

Further comments

If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[hello-stephen]

TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 35.79 seconds
load time: 498 seconds
storage size: 17122578208 Bytes
https://doris-community-test-1308700295.cos.ap-hongkong.myqcloud.com/tmp/20230118032751_clickbench_pr_82748.html

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[luozenglin]

@HappenLee cc

[HappenLee]

LGTM

[github-actions]

PR approved by at least one committer and no changes requested.

[github-actions]

PR approved by anyone and no changes requested.

[Hastyshell]

The connection held in broker reader will be null(the internal client is set to nullptr) after reopening the connection to the broker failed.