Add an endpoint for service metadata and use it to configure the app (#…

…116)

* Define types and lay the groundwork for service metadata

* Fix bug with settings active_keys being unordered

* Add /info endpoint for ServiceMetadata

* Add client support for /info endpoint

* Update app to use metadata endpoint for dynamic settings

* Introduce AgentClientError and clean up agent error handling

* Update app error handling

* Add an integration test with streamlit app

* Update README.md

README.md

@@ -57,20 +57,21 @@ docker compose watch
 1. **Multiple Agent Support**: Run multiple agents in the service and call by URL path
 1. **Asynchronous Design**: Utilizes async/await for efficient handling of concurrent requests.
 1. **Feedback Mechanism**: Includes a star-based feedback system integrated with LangSmith.
+1. **Dynamic Metadata**: `/info` endpoint provides dynamically configured metadata about the service and available agents and models.
 1. **Docker Support**: Includes Dockerfiles and a docker compose file for easy development and deployment.
+1. **Testing**: Includes robust unit and integration tests for the full repo.
 
 ### Key Files
 
 The repository is structured as follows:
 
-- `src/agents/research_assistant.py`: Defines the main LangGraph agent
-- `src/agents/llama_guard.py`: Defines the LlamaGuard content moderation
-- `src/agents/models.py`: Configures available models based on ENV
-- `src/agents/agents.py`: Mapping of all agents provided by the service
-- `src/schema/schema.py`: Defines the protocol schema
+- `src/agents/`: Defines several agents with different capabilities
+- `src/schema/`: Defines the protocol schema
+- `src/core/`: Core modules including LLM definition and settings
 - `src/service/service.py`: FastAPI service to serve the agents
 - `src/client/client.py`: Client to interact with the agent service
 - `src/streamlit_app.py`: Streamlit app providing a chat interface
+- `tests/`: Unit and integration tests
 
 ## Why LangGraph?
 
@@ -213,7 +214,7 @@ Contributions are welcome! Please feel free to submit a Pull Request.
 - [x] Add more sophisticated tools for the research assistant
 - [x] Increase test coverage and add CI pipeline
 - [x] Add support for multiple agents running on the same service, including non-chat agent
-- [ ] Deployment instructions and configuration for cloud providers
+- [x] Service metadata endpoint `/info` and dynamic app configuration
 - [ ] More ideas? File an issue or create a discussion!
 
 ## License

src/agents/__init__.py

@@ -1,3 +1,3 @@
-from agents.agents import DEFAULT_AGENT, agents
+from agents.agents import DEFAULT_AGENT, get_agent, get_all_agent_info
 
-__all__ = ["agents", "DEFAULT_AGENT"]
+__all__ = ["get_agent", "get_all_agent_info", "DEFAULT_AGENT"]

src/agents/agents.py

@@ -1,14 +1,35 @@
+from dataclasses import dataclass
+
 from langgraph.graph.state import CompiledStateGraph
 
 from agents.bg_task_agent.bg_task_agent import bg_task_agent
 from agents.chatbot import chatbot
 from agents.research_assistant import research_assistant
+from schema import AgentInfo
 
 DEFAULT_AGENT = "research-assistant"
 
 
-agents: dict[str, CompiledStateGraph] = {
-    "chatbot": chatbot,
-    "research-assistant": research_assistant,
-    "bg-task-agent": bg_task_agent,
+@dataclass
+class Agent:
+    description: str
+    graph: CompiledStateGraph
+
+
+agents: dict[str, Agent] = {
+    "chatbot": Agent(description="A simple chatbot.", graph=chatbot),
+    "research-assistant": Agent(
+        description="A research assistant with web search and calculator.", graph=research_assistant
+    ),
+    "bg-task-agent": Agent(description="A background task agent.", graph=bg_task_agent),
 }
+
+
+def get_agent(agent_id: str) -> CompiledStateGraph:
+    return agents[agent_id].graph
+
+
+def get_all_agent_info() -> list[AgentInfo]:
+    return [
+        AgentInfo(key=agent_id, description=agent.description) for agent_id, agent in agents.items()
+    ]

src/agents/research_assistant.py

@@ -2,6 +2,7 @@
 from typing import Literal
 
 from langchain_community.tools import DuckDuckGoSearchResults, OpenWeatherMapQueryRun
+from langchain_community.utilities import OpenWeatherMapAPIWrapper
 from langchain_core.language_models.chat_models import BaseChatModel
 from langchain_core.messages import AIMessage, SystemMessage
 from langchain_core.runnables import RunnableConfig, RunnableLambda, RunnableSerializable
@@ -31,7 +32,10 @@ class AgentState(MessagesState, total=False):
 # Add weather tool if API key is set
 # Register for an API key at https://openweathermap.org/api/
 if settings.OPENWEATHERMAP_API_KEY:
-    tools.append(OpenWeatherMapQueryRun(name="Weather"))
+    wrapper = OpenWeatherMapAPIWrapper(
+        openweathermap_api_key=settings.OPENWEATHERMAP_API_KEY.get_secret_value()
+    )
+    tools.append(OpenWeatherMapQueryRun(name="Weather", api_wrapper=wrapper))
 
 current_date = datetime.now().strftime("%B %d, %Y")
 instructions = f"""

src/client/__init__.py

@@ -1,3 +1,3 @@
-from client.client import AgentClient
+from client.client import AgentClient, AgentClientError
 
-__all__ = ["AgentClient"]
+__all__ = ["AgentClient", "AgentClientError"]