[Feature][Connector-V2] The csv, text, json,xml,parquet and orc file types of the file type connector (localfile, hdfs, s3) as source support sharding

### Search before asking

- [x] I had searched in the [feature](https://github.com/apache/seatunnel/issues?q=is%3Aissue+label%3A%22Feature%22) and found no similar feature requirement.

### Description

When encountering large files or very large files to read, the file is divided according to a certain number of bytes, and parallel reading can improve reading efficiency.

### Usage Scenario

_No response_

### Implementation scope
| | text | csv | json | xml | parquet | orc |
| -------------- | ---- | ---- | ---- | ---- | -------- | ---- |
| local file | done | done | done | | done | doing |
| hdfs | done | done | done | | done | |
| s3 | | | | | | |
| obs | | | | | | |
| oss | | | | | | |
### Are you willing to submit a PR?

- [x] Yes I am willing to submit a PR!

### Code of Conduct

- [x] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)