Improve multi-turn capability for agent

[minmin-intel]

Description

1. Removed "with_store" and replaced with "memory_type". Two memory types: volatile, persistent. Volatile uses MemorySaver checkpointer, only knows thread_id, no assistant_id. Persistent uses RedisPersistence, knows assistant_id and thread_id. chat_completions APIs only supports volatile, assistants APIs support both types.
2. Improved react_llama prompt templates and code for multi-turn conversations.
3. Changed react_llama code to output thought process in AIMessage for UI to show to users.

Issues

#1184
#1183

Type of change

List the type of change like below. Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds new functionality)
• Breaking change (fix or feature that would break existing design and interface)
• Others (enhancement, documentation, validation, etc.)

Dependencies

No

Tests

Unit tests