Add MongoDB VectorDB Support and Optimize Existing MongoDB Implementations

[ArnoChenFx]

This Pull Request introduces MongoDB VectorDB support and optimizes existing MongoDB implementations for KV, Doc Status, and Graph storage within LightRAG.

• MongoVectorDBStorage (New): Implements vector data storage and provides similarity search capabilities using Atlas Vector Search. The MongoDB VectorDB implementation was contributed by @ankit1063 in issue Regarding Mongo DB vector storage integration. #672.
• MongoKVStorage (Optimized): This PR includes improvements for asynchronous operations.
• MongoDocStatusStorage (Optimized): This PR includes improvements for asynchronous operations.
• MongoGraphStorage (Optimized): This PR includes improvements for graph traversal and asynchronous operations. Adds the ability to retrieve all labels and query knowledge graphs with a specified label and maximum depth.

Specific Changes:

• Added MongoVectorDBStorage to the STORAGE_IMPLEMENTATIONS and STORAGES dictionaries in lightrag.py.
• Updated api/README.md to include MongoVectorDBStorage in the list of supported storage implementations.
• Modified the kg/mongo_impl.py file, containing the implementations and optimizations for existing features and the vector database functionality.
• Optimized existing KV, Doc Status, and Graph implementations to improve asynchronous operation compatibility.