Add Milvus vector database support

[hemanthrayuduu]

Description

This PR adds support for Milvus as a document engine option in RAGFlow, alongside the existing Elasticsearch and Infinity options.

Changes

• Created Milvus connector class in rag/utils/milvus_conn.py
• Added Milvus mapping configuration in conf/milvus_mapping.json
• Updated settings to include Milvus in api/settings.py
• Added Milvus services to Docker Compose configuration
• Added Milvus environment variables to .env file
• Updated service configuration template
• Added PyMilvus as a dependency in pyproject.toml
• Updated documentation to include Milvus configuration information

Testing

• Created unit tests for the Milvus implementation
• Verified all configuration files contain the necessary Milvus settings
• Tested the basic structure and functionality of the implementation

Related Issue

Closes #6359

[yingfeng]

Thanks for the contribution, however, Milvus can not satisfy the requirements of RAGFlow due to poor support for full text search. Actually, sparse embedding is not an alternative for full text search, there are many features can not being supported including:
Phrase search
Rank features
...,etc