Add LLM configuration and update URL routing

- Introduced LLM configuration in rag_config.yaml and corresponding LLMConfig class in config.py.
- Updated load_rag_config function to parse LLM settings from the configuration file.
- Added new API route for RAG in urls.py to facilitate access to the chat model.
- Modified QdrantVectorStore to use query_points method for improved functionality.

rag/__init__.py

@@ -3,6 +3,7 @@
 فاز یک: Qdrant به‌عنوان vector store
 """
 
+from .chat import chat_rag_stream
 from .chunker import chunk_text, chunk_texts
 from .client import get_qdrant_client
 from .config import load_rag_config
@@ -12,6 +13,7 @@ from .retrieve import search_with_query
 from .vector_store import QdrantVectorStore
 
 __all__ = [
+    "chat_rag_stream",
     "chunk_text",
     "chunk_texts",
     "embed_single",