support Foundry Local on chat playground

Author: ks6088ts

.env.template

@@ -30,6 +30,9 @@ AZURE_AI_FOUNDRY_INFERENCE_MODEL_CHAT="gpt-5"
 ## Ollama Settings
 OLLAMA_MODEL_CHAT="gemma3:latest"
 
+## Foundry Local Settings
+FOUNDRY_LOCAL_MODEL_CHAT="phi-3-mini-4k"
+
 # ---------
 # Tools
 # ---------

template_langgraph/llms/foundry_locals.py

@@ -0,0 +1,38 @@
+from functools import lru_cache
+
+from foundry_local import FoundryLocalManager
+from langchain_openai import ChatOpenAI
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class Settings(BaseSettings):
+    foundry_local_model_chat: str = "phi-3-mini-4k"
+
+    model_config = SettingsConfigDict(
+        env_file=".env",
+        env_ignore_empty=True,
+        extra="ignore",
+    )
+
+
+@lru_cache
+def get_foundry_local_settings() -> Settings:
+    return Settings()
+
+
+class FoundryLocalWrapper:
+    def __init__(self, settings: Settings = None):
+        if settings is None:
+            settings = get_foundry_local_settings()
+
+        self.foundry_local_manager = FoundryLocalManager(
+            alias_or_model_id=settings.foundry_local_model_chat,
+        )
+
+        self.chat_model = ChatOpenAI(
+            model=self.foundry_local_manager.get_model_info(settings.foundry_local_model_chat).id,
+            base_url=self.foundry_local_manager.endpoint,
+            api_key=self.foundry_local_manager.api_key,
+            temperature=0.0,
+            streaming=True,
+        )

template_langgraph/services/streamlits/pages/playground_chat.py

@@ -8,6 +8,8 @@
 from langchain_openai import AzureChatOpenAI
 from openai import APIConnectionError, APIStatusError, APITimeoutError
 
+from template_langgraph.llms.foundry_locals import FoundryLocalWrapper
+from template_langgraph.llms.foundry_locals import Settings as FoundryLocalSettings
 from template_langgraph.loggers import get_logger
 
 load_dotenv(override=True)
@@ -24,7 +26,7 @@ def image_to_base64(image_bytes: bytes) -> str:
     "# Common Settings"
     stream_mode = st.checkbox(
         label="ストリーム出力を有効にする",
-        value=True,
+        value=False,
         key="STREAM_MODE",
     )
     "# Model"
@@ -33,6 +35,7 @@ def image_to_base64(image_bytes: bytes) -> str:
         options=[
             "azure",
             "ollama",
+            "foundry_local",
         ],
         index=0,
         key="model_choice",
@@ -66,7 +69,6 @@ def image_to_base64(image_bytes: bytes) -> str:
         "### Documents"
         "[Azure Portal](https://portal.azure.com/)"
         "[Azure OpenAI Studio](https://oai.azure.com/resource/overview)"
-        "[View the source code](https://github.com/ks6088ts-labs/template-streamlit)"
     elif model_choice == "ollama":
         ollama_model_chat = st.text_input(
             label="OLLAMA_MODEL_CHAT",
@@ -76,10 +78,18 @@ def image_to_base64(image_bytes: bytes) -> str:
         )
         "### Documents"
         "[Ollama Docs](https://github.com/ollama/ollama)"
-        "[View the source code](https://github.com/ks6088ts-labs/template-streamlit)"
+    elif model_choice == "foundry_local":
+        foundry_local_model_chat = st.text_input(
+            label="FOUNDRY_LOCAL_MODEL_CHAT",
+            value=getenv("FOUNDRY_LOCAL_MODEL_CHAT", "phi-3-mini-4k"),
+            key="FOUNDRY_LOCAL_MODEL_CHAT",
+            type="default",
+        )
+        "### Documents"
+        "[Get started with Foundry Local](https://learn.microsoft.com/en-us/azure/ai-foundry/foundry-local/get-started)"
     else:
-        st.error("Invalid model choice. Please select either 'azure' or 'ollama'.")
-        raise ValueError("Invalid model choice. Please select either 'azure' or 'ollama'.")
+        st.error("Invalid model choice. Please select either 'azure', 'ollama', or 'foundry_local'.")
+        raise ValueError("Invalid model choice. Please select either 'azure', 'ollama', or 'foundry_local'.")
 
 
 def is_azure_configured():
@@ -96,8 +106,12 @@ def is_ollama_configured():
     return st.session_state.get("OLLAMA_MODEL_CHAT") and st.session_state.get("model_choice") == "ollama"
 
 
+def is_foundry_local_configured():
+    return st.session_state.get("FOUNDRY_LOCAL_MODEL_CHAT") and st.session_state.get("model_choice") == "foundry_local"
+
+
 def is_configured():
-    return is_azure_configured() or is_ollama_configured()
+    return is_azure_configured() or is_ollama_configured() or is_foundry_local_configured()
 
 
 def get_model():
@@ -112,7 +126,13 @@ def get_model():
         return ChatOllama(
             model=st.session_state.get("OLLAMA_MODEL_CHAT", ""),
         )
-    raise ValueError("No model is configured. Please set up the Azure or Ollama model in the sidebar.")
+    elif is_foundry_local_configured():
+        return FoundryLocalWrapper(
+            settings=FoundryLocalSettings(
+                foundry_local_model_chat=st.session_state.get("FOUNDRY_LOCAL_MODEL_CHAT", "phi-3-mini-4k"),
+            )
+        ).chat_model
+    raise ValueError("No model is configured. Please set up the Azure, Ollama, or Foundry Local model in the sidebar.")
 
 
 st.title("Chat Playground")