add reflection agent

Author: ks6088ts

README.md

@@ -41,6 +41,7 @@ Here are the preferred tools for development.
 | [9_streamlit_azure_document_intelligence](./apps/9_streamlit_azure_document_intelligence/README.md) | Call Azure AI Document Intelligence API with Streamlit  | ![9_streamlit_azure_document_intelligence](./docs/images/9_streamlit_azure_document_intelligence.main.png) |
 | [10_streamlit_batch_transcription](./apps/10_streamlit_batch_transcription/README.md)               | Call Batch Transcription API with Streamlit             | ![10_streamlit_batch_transcription](./docs/images/10_streamlit_batch_transcription.main.png)               |
 | [11_promptflow](./apps/11_promptflow/README.md)                                                     | Get started with Prompt flow                            | No Image                                                                                                   |
+| [12_langgraph_agent](./apps/12_langgraph_agent/README.md)                                           | Create agents with LangGraph                            | No Image                                                                                                   |
 | [99_streamlit_examples](./apps/99_streamlit_examples/README.md)                                     | Code samples for Streamlit                              | ![99_streamlit_examples](./docs/images/99_streamlit_examples.explaindata.png)                              |
 
 ## How to run

apps/12_langgraph_agent/README.md

@@ -0,0 +1,45 @@
+# Create agents with LangGraph
+
+This app demonstrates how to implement agents with LangGraph.
+
+## Prerequisites
+
+- Python 3.10 or later
+- Azure OpenAI Service
+
+## Overview
+
+**What is [LangGraph](https://langchain-ai.github.io/langgraph/)?**
+
+LangGraph is a library for building stateful, multi-actor applications with LLMs, used to create agent and multi-agent workflows.
+
+This chapter provides a practical example of how to use LangGraph to create an agent that can interact with users and external tools.
+
+## Usage
+
+1. Get Azure OpenAI Service API key
+1. Copy [.env.template](../../.env.template) to `.env` in the same directory
+1. Set credentials in `.env`
+1. Run main.py
+
+```shell
+# Create a virtual environment
+$ python -m venv .venv
+
+# Activate the virtual environment
+$ source .venv/bin/activate
+
+# Install dependencies
+$ pip install -r requirements.txt
+
+# Run the script
+$ python apps/12_langgraph_agent/reflection_agent/main.py
+```
+
+### Example
+
+## References
+
+- [LangGraph](https://langchain-ai.github.io/langgraph/)
+- [Udemy > LangGraph- Develop LLM powered agents with LangGraph](https://www.udemy.com/course/langgraph)
+- [Prompt flow > Tracing](https://microsoft.github.io/promptflow/how-to-guides/tracing/index.html)

apps/12_langgraph_agent/reflection_agent/chains.py

@@ -0,0 +1,39 @@
+from os import getenv
+
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain_openai import AzureChatOpenAI
+
+reflection_prompt = ChatPromptTemplate.from_messages(
+    [
+        (
+            "system",
+            "You are a viral twitter influencer grading a tweet. Generate critique and recommendations for the user's tweet."  # noqa
+            "Always provide detailed recommendations, including requests for length, virality, style, etc.",  # noqa
+        ),
+        MessagesPlaceholder(variable_name="messages"),
+    ]
+)
+
+generation_prompt = ChatPromptTemplate.from_messages(
+    [
+        (
+            "system",
+            "You are a twitter techie influencer assistant tasked with writing excellent twitter posts."
+            " Generate the best twitter post possible for the user's request."
+            " If the user provides critique, respond with a revised version of your previous attempts.",  # noqa
+        ),
+        MessagesPlaceholder(variable_name="messages"),
+    ]
+)
+
+
+llm = AzureChatOpenAI(
+    temperature=0,
+    api_key=getenv("AZURE_OPENAI_API_KEY"),
+    api_version=getenv("AZURE_OPENAI_API_VERSION"),
+    azure_endpoint=getenv("AZURE_OPENAI_ENDPOINT"),
+    model=getenv("AZURE_OPENAI_GPT_MODEL"),
+)
+
+generate_chain = generation_prompt | llm
+reflect_chain = reflection_prompt | llm

apps/12_langgraph_agent/reflection_agent/main.py

@@ -0,0 +1,50 @@
+from collections.abc import Sequence
+
+from chains import generate_chain, reflect_chain
+from dotenv import load_dotenv
+from langchain_core.messages import BaseMessage, HumanMessage
+from langgraph.graph import END, MessageGraph
+from promptflow.tracing import start_trace
+
+load_dotenv()
+start_trace()
+
+
+REFLECT = "reflect"
+GENERATE = "generate"
+
+
+def generation_node(state: Sequence[BaseMessage]):
+    return generate_chain.invoke({"messages": state})
+
+
+def reflection_node(messages: Sequence[BaseMessage]) -> list[BaseMessage]:
+    res = reflect_chain.invoke({"messages": messages})
+    return [HumanMessage(content=res.content)]
+
+
+builder = MessageGraph()
+builder.add_node(GENERATE, generation_node)
+builder.add_node(REFLECT, reflection_node)
+builder.set_entry_point(GENERATE)
+
+
+def should_continue(state: list[BaseMessage]):
+    if len(state) > 6:
+        return END
+    return REFLECT
+
+
+builder.add_conditional_edges(GENERATE, should_continue)
+builder.add_edge(REFLECT, GENERATE)
+
+graph = builder.compile()
+print(graph.get_graph().draw_mermaid())
+# graph.get_graph().print_ascii()
+
+if __name__ == "__main__":
+    inputs = HumanMessage(
+        content="""Make this tweet better:"
+        Microsoft's Azure OpenAI API is a game-changer for developers."""
+    )
+    response = graph.invoke(inputs)

requirements.txt

@@ -14,6 +14,7 @@ azure-storage-blob==12.22.0
 requests==2.32.3
 promptflow==1.15.0
 promptflow-evals==0.3.2
+langgraph==0.2.23
 
 # To run 99_streamlit_examples/pages/10_Object_Detection.py
 # ultralytics==8.2.89