Enforce modern typing (#258)

* Allow UP

* Upgrade typing

* test smolagents

* Change to contextlib

---------

Co-authored-by: Marcus Valtonen Örnhag <marcus.valtonen.ornhag@ericsson.com>

demo/gemini_conversation/app.py

@@ -1,8 +1,8 @@
 import asyncio
 import base64
 import os
+from collections.abc import AsyncGenerator
 from pathlib import Path
-from typing import AsyncGenerator
 
 import librosa
 import numpy as np

demo/moonshine_live/app.py

@@ -1,5 +1,6 @@
-from functools import lru_cache
-from typing import Generator, Literal
+from collections.abc import Generator
+from functools import cache
+from typing import Literal
 
 import gradio as gr
 import numpy as np
@@ -17,7 +18,7 @@ from numpy.typing import NDArray
 load_dotenv()
 
 
-@lru_cache(maxsize=None)
+@cache
 def load_moonshine(
     model_name: Literal["moonshine/base", "moonshine/tiny"],
 ) -> MoonshineOnnxModel:

demo/talk_to_gemini/app.py

@@ -3,7 +3,8 @@ import base64
 import json
 import os
 import pathlib
-from typing import AsyncGenerator, Literal
+from collections.abc import AsyncGenerator
+from typing import Literal
 
 import gradio as gr
 import numpy as np

demo/talk_to_smolagents/app.py

@@ -1,5 +1,4 @@
 from pathlib import Path
-from typing import Dict, List
 
 from dotenv import load_dotenv
 from fastrtc import (
@@ -22,7 +21,7 @@ stt_model = get_stt_model()
 tts_model = get_tts_model()
 
 # Conversation state to maintain history
-conversation_state: List[Dict[str, str]] = []
+conversation_state: list[dict[str, str]] = []
 
 # System prompt for agent
 system_prompt = """You are a helpful assistant that can helps with finding places to 
@@ -78,9 +77,7 @@ def process_response(audio):
     response_content = agent.run(input_text)
 
     # Convert response to audio using TTS model
-    for audio_chunk in tts_model.stream_tts_sync(response_content or ""):
-        # Yield the audio chunk
-        yield audio_chunk
+    yield from tts_model.stream_tts_sync(response_content or "")
 
 
 stream = Stream(