modified: langchain-crash-course/5_agents_tools/agent_react_rag_context.py

        modified:   langchain-crash-course/5_agents_tools/logs/agent_tools.log

Author: SystemSolution21

langchain-crash-course/5_agents_tools/agent_react_rag_context.py

@@ -10,6 +10,8 @@
 from dotenv import load_dotenv
 
 # Import langchain modules
+from langchain import hub
+from langchain.agents import AgentExecutor, create_react_agent
 from langchain.chains import (
     create_history_aware_retriever,
     create_retrieval_chain,
@@ -20,6 +22,7 @@
 from langchain_core.messages.base import BaseMessage
 from langchain_core.prompts.chat import ChatPromptTemplate, MessagesPlaceholder
 from langchain_core.runnables.base import Runnable
+from langchain_core.tools import Tool
 from langchain_core.vectorstores.base import VectorStoreRetriever
 from langchain_ollama import ChatOllama
 from langchain_ollama.embeddings import OllamaEmbeddings
@@ -41,9 +44,9 @@
 logger.info(msg="Starting Agent ReAct RAG Context Application")
 logger.info(msg="=" * 50)
 
+# ===== Setup RAG =====
 # Define directories and paths
 rag_dir: Path = Path(__file__).parents[1] / "4_rag"
-books_dir: Path = rag_dir / "books"
 db_dir: Path = rag_dir / "db"
 store_name: str = "chroma_db_with_metadata"
 persistent_directory: Path = db_dir / store_name
@@ -102,7 +105,7 @@
 )
 
 # Create a history-aware retriever
-# this users the LLM to help reformulate the question based on chat history
+# this helps LLM to reformulate the question based on chat history
 history_aware_retriever: VectorStoreRetriever = create_history_aware_retriever(
     llm,
     retriever,
@@ -141,19 +144,46 @@
     history_aware_retriever, question_answering_chain
 )
 
+# ===== Setup ReAct Agent with RAG =====
+# load ReAct prompt template from hub
+react_prompt_template: Any = hub.pull(owner_repo_commit="hwchase17/react")
+
+# create a tool that uses the RAG chain
+tools: list[Tool] = [
+    Tool(
+        name="Answer Question",
+        func=lambda input, **kwargs: rag_chain.invoke(
+            input={"input": input, "chat_history": kwargs.get("chat_history", [])}
+        ),
+        description="Useful for answering questions based on the provided context.",
+    ),
+]
+
+# Create a ReAct agent with the RAG tool
+agent: Runnable[Any, Any] = create_react_agent(
+    tools=tools,
+    llm=llm,
+    prompt=react_prompt_template,
+)
+
+# Create agent executor
+agent_executor: AgentExecutor = AgentExecutor(
+    agent=agent, tools=tools, verbose=True, handle_parsing_errors=True
+)
 
-# Run RAG LLM conversation
+
+# ===== Run ReAct RAG conversation =====
 def main() -> None:
     """
-    Runs the main conversational loop for the RAG-based chat application.
+    Runs the main conversational loop for the RAG-based ReAct chat application.
 
     This function initializes the chat history and enters an infinite loop to
     continuously accept user input. It processes the user's query through the
     RAG chain, prints the AI's response, and updates the chat history.
     The loop can be exited by typing 'exit', or by sending a
     KeyboardInterrupt (Ctrl+C) or EOFError (Ctrl+D).
     """
-    print("\nStart chatting with AI! Type 'exit' to end the conversation.")
+    print("\nStart RAG-based ReAct chatting! Type 'exit' to end the conversation.")
 
     # Initialize chat history
     chat_history: list[BaseMessage] = []
@@ -164,27 +194,30 @@ def main() -> None:
             query: str = input("You: ").strip()
 
             if not query:
+                print("Please ask a question!.")
                 continue
 
             if query.lower() == "exit":
                 logger.info(msg="User exited conversation")
                 print("Exiting...")
                 break
 
-            # Process user query through RAG chain
-            logger.info(msg="Processing user query through RAG chain...")
-            result: Any = rag_chain.invoke(
+            # Process user query through agent executor
+            logger.info(
+                msg="Processing user query through ReAct Agent with RAG chain..."
+            )
+            response: Any = agent_executor.invoke(
                 input={"input": query, "chat_history": chat_history}
             )
 
             # Display AI response
-            if result:
+            if response:
                 logger.info(msg="AI response generated successfully")
-                print(f"AI: {result['answer']}")
+                print(f"AI: {response['output']}")
 
                 # Update chat history
                 chat_history.append(HumanMessage(content=query))
-                chat_history.append(AIMessage(content=result["answer"]))
+                chat_history.append(AIMessage(content=response["output"]))
                 logger.info(msg="Chat history updated successfully")
 
         except (KeyboardInterrupt, EOFError):

langchain-crash-course/5_agents_tools/logs/agent_tools.log

@@ -20,3 +20,18 @@
 2025-11-04 22:58:41,439 - INFO - agent_react_rag_context - AI response generated successfully
 2025-11-04 22:58:41,466 - INFO - agent_react_rag_context - Chat history updated successfully
 2025-11-04 22:59:12,397 - INFO - agent_react_rag_context - User exited conversation
+2025-11-06 18:12:21,988 - INFO - agent_react_rag_context - ==================================================
+2025-11-06 18:12:21,988 - INFO - agent_react_rag_context - Starting Agent ReAct RAG Context Application
+2025-11-06 18:12:21,988 - INFO - agent_react_rag_context - ==================================================
+2025-11-06 18:12:24,090 - INFO - agent_react_rag_context - Loading vector store 'chroma_db_with_metadata'...
+2025-11-06 18:12:24,255 - INFO - agent_react_rag_context - Created retriever from vector store 'chroma_db_with_metadata' successfully.
+2025-11-06 18:12:49,546 - INFO - agent_react_rag_context - Processing user query through ReAct Agent with RAG chain...
+2025-11-06 18:13:53,367 - INFO - agent_react_rag_context - AI response generated successfully
+2025-11-06 18:13:53,369 - INFO - agent_react_rag_context - Chat history updated successfully
+2025-11-06 18:15:03,021 - INFO - agent_react_rag_context - Processing user query through ReAct Agent with RAG chain...
+2025-11-06 18:16:15,185 - INFO - agent_react_rag_context - AI response generated successfully
+2025-11-06 18:16:15,186 - INFO - agent_react_rag_context - Chat history updated successfully
+2025-11-06 18:20:35,475 - INFO - agent_react_rag_context - Processing user query through ReAct Agent with RAG chain...
+2025-11-06 18:21:10,565 - INFO - agent_react_rag_context - AI response generated successfully
+2025-11-06 18:21:10,566 - INFO - agent_react_rag_context - Chat history updated successfully
+2025-11-06 18:27:06,733 - INFO - agent_react_rag_context - User exited conversation