### What problem does this PR solve?

This PR solve the problems metioned in the
pr(https://github.com/infiniflow/ragflow/pull/7140) which is also
submitted by me

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
- [ ] New Feature (non-breaking change which adds functionality)
- [ ] Documentation Update
- [ ] Refactoring
- [ ] Performance Improvement
- [ ] Other (please describe):


### Introduction
I fixed the problems when using OpenSearch as the DOC_ENGINE, the
failures of pytest and the wrong API's return.
Mainly about delete chunk, list chunks, update chunk, retrieval chunk.
The pytest comand "cd sdk/python && uv sync --python 3.10 --group test
--frozen && source .venv/bin/activate && cd test/test_http_api &&
DOC_ENGINE=opensearch pytest test_chunk_management_within_dataset -s
--tb=short " is finally successful.

###Others
As some changes between Elasticsearch And Opensearch differ, some pytest
results about OpenSearch are correct and resonable. However, some pytest
params (skipif params) are incompatible. So I changed some pytest params
about skipif.

As a search engine programmer, I will still focus on the usage of vector
databases (especially OpenSearch) for the RAG stuff.
Thanks for your review