add FastAPI server for running agents

.env.template

@@ -33,11 +33,11 @@ ELASTICSEARCH_URL="http://localhost:9200"
 DIFY_API_URL="https://api.dify.ai/v1"
 DIFY_API_KEY="xxx"
 
-## Cosmos DB Settings
+## Azure Cosmos DB Settings
 COSMOSDB_HOST="https://xxx.documents.azure.com:443/"
 COSMOSDB_KEY="xxx"
-COSMOSDB_DATABASE_NAME="langgraph"
-COSMOSDB_CONTAINER_NAME="docs_kabuto"
+COSMOSDB_DATABASE_NAME="template_langgraph"
+COSMOSDB_CONTAINER_NAME="kabuto"
 COSMOSDB_PARTITION_KEY="/id"
 
 # ---------

Makefile

@@ -115,3 +115,15 @@ ci-test-docs: install-deps-docs docs ## run CI test for documentation
 .PHONY: langgraph-studio
 langgraph-studio: ## run LangGraph Studio
 	uv run langgraph dev
+
+.PHONY: fastapi-dev
+fastapi-dev: ## run FastAPI
+	uv run fastapi dev ./template_langgraph/services/fastapis/main.py
+
+.PHONY: fastapi
+fastapi: ## run FastAPI in production mode
+	uv run fastapi run \
+		--host "0.0.0.0" \
+		--port 8000 \
+		--workers 4 \
+		template_langgraph/services/fastapis/main.py

docs/index.ja.md

@@ -35,6 +35,9 @@
 - [uv](https://docs.astral.sh/uv/getting-started/installation/) - モダンな Python パッケージマネージャー
 - [GNU Make](https://www.gnu.org/software/make/) - 一般的なタスクの実行用
 - [Docker](https://www.docker.com/) - ベクターデータベースの実行用（オプション）
+- Azure
+  - [Azure OpenAI](https://learn.microsoft.com/ja-jp/azure/ai-foundry/openai/overview) - LLM API
+  - [Azure Cosmos DB](https://learn.microsoft.com/ja-jp/azure/cosmos-db/) - データストレージ（オプション）
 
 ## クイックスタート
 
@@ -129,7 +132,7 @@ Pydantic モデルを使用して AI 応答から構造化データを取得す
 - **`template_langgraph/tools/`** - 検索、データ取得用ツール実装
 - **`template_langgraph/utilities/`** - ドキュメント読み込みと処理用ヘルパー関数
 
-## 例の実行
+## サンプルコードの実行
 
 ### オプション 1: LangGraph Studio（開発用推奨）
 
@@ -159,7 +162,7 @@ uv run jupyter lab
 
 ![jupyterlab.png](./images/jupyterlab.png)
 
-### オプション 3: コマンドライン（本番環境的）
+### オプション 3: コマンドライン（クイックな開発）
 
 ターミナルからエージェントを実行：
 
@@ -191,6 +194,23 @@ KABUTO起動時の紫画面点滅は「忍者プロトコル」初期化エラ
 ')]}}
 ```
 
+### オプション 4: FastAPI サーバー（プロダクションレディ）
+
+FastAPI サーバーを実行して、エージェントを API として公開：
+
+```shell
+uv run fastapi run \
+  --host "0.0.0.0" \
+  --port 8000 \
+  --workers 4 \
+  template_langgraph/services/fastapis/main.py
+# http://localhost:8000/docs を開いて Swagger UI 経由で API にアクセス
+```
+
+これにより、HTTP リクエストを介してプログラム的にエージェントと対話できます。
+
+![fastapi.png](./images/fastapi.png)
+
 ## 実演されている主要概念
 
 ### 1. **ReAct パターン**（推論 + 行動）
@@ -237,8 +257,8 @@ LangGraph が複数のインタラクションステップにわたってコン
 
 ## 学習リソース
 
-- [LangGraph 文書](https://langchain-ai.github.io/langgraph/)
-- [LangChain 文書](https://python.langchain.com/)
+- [LangGraph 公式ドキュメント](https://langchain-ai.github.io/langgraph/)
+- [LangChain 公式ドキュメント](https://python.langchain.com/)
 
 ## アーキテクチャの例
 

docs/index.md

@@ -35,6 +35,9 @@ This template provides working examples of all these patterns using LangGraph.
 - [uv](https://docs.astral.sh/uv/getting-started/installation/) - Modern Python package manager
 - [GNU Make](https://www.gnu.org/software/make/) - For running common tasks
 - [Docker](https://www.docker.com/) - For running vector databases (optional)
+- Azure
+  - [Azure OpenAI](https://learn.microsoft.com/ja-jp/azure/ai-foundry/openai/overview) - LLM API
+  - [Azure Cosmos DB](https://learn.microsoft.com/ja-jp/azure/cosmos-db/) - Data storage (optional)
 
 ## Quick Start
 
@@ -159,7 +162,7 @@ uv run jupyter lab
 
 ![jupyterlab.png](./images/jupyterlab.png)
 
-### Option 3: Command Line (Production-like)
+### Option 3: Command Line (Quick development)
 
 Run agents from the terminal:
 
@@ -191,6 +194,23 @@ The purple screen flashing during KABUTO startup indicates a "Shinobi Protocol"
 ')]}}
 ```
 
+### Option 4: FastAPI (Production-ready)
+
+Run the FastAPI server to expose the agent as an API:
+
+```shell
+uv run fastapi run \
+  --host "0.0.0.0" \
+  --port 8000 \
+  --workers 4 \
+  template_langgraph/services/fastapis/main.py
+# Access the API at http://localhost:8000/docs via Swagger UI
+```
+
+This allows you to interact with the agent programmatically via HTTP requests.
+
+![fastapi.png](./images/fastapi.png)
+
 ## Key Concepts Demonstrated
 
 ### 1. **ReAct Pattern** (Reasoning + Acting)

docs/references.ja.md

@@ -20,3 +20,7 @@
 - [CSVLoader](https://python.langchain.com/docs/how_to/document_loader_csv/)
 - [Qdrant](https://github.com/qdrant/qdrant)
 - [Azure Cosmos DB No SQL](https://python.langchain.com/docs/integrations/vectorstores/azure_cosmos_db_no_sql/)
+
+### Services
+
+- [FastAPI](https://fastapi.tiangolo.com/)

docs/references.md

@@ -20,3 +20,7 @@
 - [CSVLoader](https://python.langchain.com/docs/how_to/document_loader_csv/)
 - [Qdrant](https://github.com/qdrant/qdrant)
 - [Azure Cosmos DB No SQL](https://python.langchain.com/docs/integrations/vectorstores/azure_cosmos_db_no_sql/)
+
+### Services
+
+- [FastAPI](https://fastapi.tiangolo.com/)

pyproject.toml

@@ -7,6 +7,7 @@ requires-python = ">=3.10"
 dependencies = [
     "azure-cosmos>=4.9.0",
     "elasticsearch>=9.1.0",
+    "fastapi[standard]>=0.116.1",
     "httpx>=0.28.1",
     "langchain-azure-ai>=0.1.4",
     "langchain-community>=0.3.27",

template_langgraph/loggers.py

@@ -1,16 +1,21 @@
 import logging
 
 
-def get_logger(name: str = "default") -> logging.Logger:
+def get_logger(
+    name: str = "default",
+    verbosity: int = logging.INFO,
+) -> logging.Logger:
     """
     Get a logger with the specified name.
 
     Args:
         name (str): The name of the logger.
+        verbosity (int): The logging level (default: logging.INFO).
     Returns:
         logging.Logger: Configured logger instance.
     """
     logger = logging.getLogger(name)
+    logger.setLevel(verbosity)
     formatter = logging.Formatter("%(asctime)s [%(levelname)8s] %(message)s (%(filename)s:%(lineno)s)")
     handler = logging.StreamHandler()
     handler.setFormatter(formatter)

template_langgraph/services/fastapis/main.py

@@ -0,0 +1,12 @@
+from fastapi import FastAPI
+
+from template_langgraph.services.fastapis.routers import agents as agents_router
+
+app = FastAPI()
+
+
+app.include_router(
+    agents_router.router,
+    prefix="/agents",
+    tags=["agents"],
+)

template_langgraph/services/fastapis/routers/agents.py

@@ -0,0 +1,56 @@
+import logging
+
+from fastapi import APIRouter
+from pydantic import BaseModel, ConfigDict
+
+from template_langgraph.agents.chat_with_tools_agent.agent import AgentState
+from template_langgraph.agents.chat_with_tools_agent.agent import graph as chat_with_tools_agent
+from template_langgraph.loggers import get_logger
+
+router = APIRouter()
+logger = get_logger(
+    name=__name__,
+    verbosity=logging.DEBUG,
+)
+
+
+class RunChatWithToolsAgentRequest(BaseModel):
+    model_config = ConfigDict(extra="ignore")
+    question: str
+
+
+class RunChatWithToolsAgentResponse(BaseModel):
+    model_config = ConfigDict(extra="ignore")
+    response: str
+
+
+@router.post(
+    "/chat_with_tools_agent/",
+    response_model=RunChatWithToolsAgentResponse,
+    operation_id="create_transcription_job",
+)
+async def run_chat_with_tools_agent(
+    request: RunChatWithToolsAgentRequest,
+) -> RunChatWithToolsAgentResponse:
+    try:
+        async for event in chat_with_tools_agent.astream(
+            input=AgentState(
+                messages=[
+                    {
+                        "role": "user",
+                        "content": request.question,
+                    },
+                ],
+            ),
+            config={
+                "recursion_limit": 30,
+            },
+        ):
+            logger.debug(f"Event received: {event}")
+        response = event["chat_with_tools"]["messages"][0].content
+        return RunChatWithToolsAgentResponse(response=response)
+    except Exception as e:
+        logger.error(f"Error processing event: {e}")
+        return RunChatWithToolsAgentResponse(
+            response=f"An error occurred while processing your request with {e}",
+        )