Enable more ruff rules

pyproject.toml

@@ -93,7 +93,7 @@ target-version = "py311"
 [tool.ruff.lint]
 # https://docs.astral.sh/ruff/rules/
 # https://docs.astral.sh/ruff/formatter/#conflicting-lint-rules
-select = ["A", "B0", "C4", "D2", "D4", "E", "F", "I"]
+select = ["A", "B0", "C4", "D2", "D4", "E", "F", "FURB", "I", "PYI", "UP"]
 ignore = ["D203", "D206", "D213", "D400", "D401", "D413", "D415", "E1", "E501"]
 
 [tool.eva]

scripts/run_text_only.py

@@ -25,7 +25,7 @@
 import os
 import shutil
 import sys
-from datetime import datetime, timezone
+from datetime import UTC, datetime
 from pathlib import Path
 from typing import Any
 
@@ -514,7 +514,7 @@ async def run_record(
     user_message = record.user_goal["starting_utterance"]
     end_reason = "max_turns"
     turn_count = 0
-    started_at = datetime.now(timezone.utc)
+    started_at = datetime.now(UTC)
 
     for turn in range(max_turns):
         turn_count = turn + 1
@@ -546,7 +546,7 @@ async def run_record(
     else:
         end_reason = "max_turns"
 
-    ended_at = datetime.now(timezone.utc)
+    ended_at = datetime.now(UTC)
     duration = (ended_at - started_at).total_seconds()
     logger.info(f"Conversation ended: {end_reason} ({turn_count} turns, {duration:.1f}s)")
 

src/eva/assistant/agentic/audio_llm_system.py

@@ -5,8 +5,9 @@
 audio so the model has full conversational context across turns.
 """
 
+from collections.abc import AsyncGenerator
 from pathlib import Path
-from typing import Any, AsyncGenerator, Optional
+from typing import Any
 
 from eva.assistant.agentic.audit_log import AuditLog
 from eva.assistant.agentic.system import AgenticSystem
@@ -37,7 +38,7 @@ def __init__(
         tool_handler: ToolExecutor,
         audit_log: AuditLog,
         alm_client: ALMvLLMClient,
-        output_dir: Optional[Path] = None,
+        output_dir: Path | None = None,
     ):
         super().__init__(
             current_date_time=current_date_time,

src/eva/assistant/agentic/audit_log.py

@@ -4,7 +4,7 @@
 import time
 from enum import StrEnum
 from pathlib import Path
-from typing import Any, Optional
+from typing import Any
 
 from pydantic import BaseModel
 
@@ -32,11 +32,11 @@ class ConversationMessage(BaseModel):
 
     role: MessageRole
     content: str
-    tool_calls: Optional[list[dict[str, Any]]] = None
-    tool_call_id: Optional[str] = None
-    name: Optional[str] = None  # For tool messages
-    turn_id: Optional[int] = None  # For associating transcription updates
-    reasoning: Optional[str] = None  # For model reasoning (e.g., from OpenAI o1)
+    tool_calls: list[dict[str, Any]] | None = None
+    tool_call_id: str | None = None
+    name: str | None = None  # For tool messages
+    turn_id: int | None = None  # For associating transcription updates
+    reasoning: str | None = None  # For model reasoning (e.g., from OpenAI o1)
 
     def to_dict(self) -> dict[str, Any]:
         """Convert to a plain dict, excluding None fields and internal tracking fields."""
@@ -47,17 +47,17 @@ class LLMCall(BaseModel):
     """Record of an LLM call."""
 
     messages: list[dict]
-    tools: Optional[list[dict]] = None
-    response: Optional[ConversationMessage] = None
+    tools: list[dict] | None = None
+    response: ConversationMessage | None = None
     duration_seconds: float = 0.0
     start_time: str = ""
     end_time: str = ""
     status: str = "success"
-    model: Optional[str] = None
+    model: str | None = None
     # New fields for enhanced tracking (optional for backward compatibility)
-    latency_ms: Optional[float] = None
-    error_type: Optional[str] = None
-    error_source: Optional[str] = None
+    latency_ms: float | None = None
+    error_type: str | None = None
+    error_source: str | None = None
     retry_attempt: int = 0
 
 
@@ -75,13 +75,13 @@ def __init__(self):
         self.conversation_messages: list[ConversationMessage] = []  # Full message sequence for LLM context
         self._tool_calls_count = 0
         self._tools_called: list[str] = []
-        self._last_tool_call: Optional[str] = None  # Track last tool called for matching responses
+        self._last_tool_call: str | None = None  # Track last tool called for matching responses
 
     def append_user_input(
         self,
         content: str,
-        timestamp_ms: Optional[str] = None,
-        turn_id: Optional[int] = None,
+        timestamp_ms: str | None = None,
+        turn_id: int | None = None,
     ) -> None:
         """Record user input.
 
@@ -189,7 +189,7 @@ def append_assistant_output(
         self,
         content: str,
         tool_calls: list[dict[str, Any]] | None = None,
-        timestamp_ms: Optional[str] = None,
+        timestamp_ms: str | None = None,
     ) -> None:
         """Record assistant output.
 
@@ -236,7 +236,7 @@ def append_tool_message(self, tool_call_id: str, content: str) -> None:
         )
         logger.debug(f"Audit: tool message for call_id {tool_call_id}")
 
-    def append_llm_call(self, llm_call: LLMCall, agent_name: Optional[str] = None) -> None:
+    def append_llm_call(self, llm_call: LLMCall, agent_name: str | None = None) -> None:
         """Record an LLM call."""
         response_content = llm_call.response.content if llm_call.response else ""
         response_dict = llm_call.response.to_dict() if llm_call.response else None
@@ -276,7 +276,7 @@ def append_tool_call(
         self,
         tool_name: str,
         parameters: dict[str, Any],
-        response: Optional[dict[str, Any]] = None,
+        response: dict[str, Any] | None = None,
     ) -> None:
         """Record a tool call and its response."""
         # Record tool call in transcript
@@ -342,7 +342,7 @@ def append_realtime_tool_call(
 
     def get_conversation_messages(
         self,
-        max_messages: Optional[int] = None,
+        max_messages: int | None = None,
     ) -> list[ConversationMessage]:
         """Get conversation messages for LLM context.
 

src/eva/assistant/agentic/system.py

@@ -5,8 +5,9 @@
 import json
 import time
 import warnings
+from collections.abc import AsyncGenerator
 from pathlib import Path
-from typing import Any, AsyncGenerator
+from typing import Any
 
 from eva.assistant.agentic.audit_log import (
     AuditLog,
@@ -220,7 +221,7 @@ async def _run_tool_loop(
                 llm_call_response = ConversationMessage(
                     role=MessageRole.ASSISTANT,
                     content=response_content,
-                    tool_calls=tool_calls_dicts if tool_calls_dicts else None,
+                    tool_calls=tool_calls_dicts or None,
                     reasoning=llm_stats.get("reasoning"),
                 )
 

src/eva/assistant/pipeline/alm_vllm.py

@@ -10,7 +10,7 @@
 import struct
 import time
 import wave
-from typing import Any, Optional
+from typing import Any
 
 from openai import AsyncOpenAI
 
@@ -163,7 +163,7 @@ def build_audio_user_message(
     async def complete(
         self,
         messages: list[dict[str, Any]],
-        tools: Optional[list[dict]] = None,
+        tools: list[dict] | None = None,
     ) -> tuple[Any, dict[str, Any]]:
         """Chat completion with audio and tool support.
 
@@ -188,7 +188,7 @@ async def complete(
             kwargs["tools"] = tools
             kwargs["tool_choice"] = "auto"
 
-        last_exception: Optional[Exception] = None
+        last_exception: Exception | None = None
         for attempt in range(self.max_retries + 1):
             try:
                 start_time = time.time()

src/eva/assistant/pipeline/audio_llm_processor.py

@@ -23,7 +23,7 @@
 import wave
 from collections.abc import Awaitable
 from pathlib import Path
-from typing import Any, Optional
+from typing import Any
 
 from openai import AsyncOpenAI
 from pipecat.frames.frames import (
@@ -81,7 +81,7 @@ def __init__(
         self,
         context,
         user_context_aggregator,
-        pre_speech_secs: Optional[float] = None,
+        pre_speech_secs: float | None = None,
         **kwargs,
     ):
         super().__init__(**kwargs)
@@ -176,7 +176,7 @@ def __init__(
         audit_log: AuditLog,
         alm_client: ALMvLLMClient,
         audio_collector: AudioLLMUserAudioCollector,
-        output_dir: Optional[Path] = None,
+        output_dir: Path | None = None,
         **kwargs,
     ) -> None:
         super().__init__(**kwargs)
@@ -195,11 +195,11 @@ def __init__(
         )
 
         # State tracking (mirrors BenchmarkAgentProcessor)
-        self._current_query_task: Optional[asyncio.Task] = None
+        self._current_query_task: asyncio.Task | None = None
         self._interrupted = asyncio.Event()
 
         # Optional callback for transcript saving (set by server.py)
-        self.on_assistant_response: Optional[Awaitable] = None
+        self.on_assistant_response: Awaitable | None = None
 
     async def process_frame(self, frame: Frame, direction: FrameDirection) -> None:
         if isinstance(frame, (EndFrame, CancelFrame)):
@@ -410,7 +410,7 @@ def __init__(
         audio_collector: AudioLLMUserAudioCollector,
         model: str = "",
         params: dict[str, Any] = None,
-        system_prompt: Optional[str] = None,
+        system_prompt: str | None = None,
         sample_rate: int = PIPELINE_SAMPLE_RATE,
         **kwargs,
     ):
@@ -426,7 +426,7 @@ def __init__(
         self._client: AsyncOpenAI = AsyncOpenAI(api_key=self._api_key, base_url=base_url)
 
         # Callback for when transcription is ready (set by server.py)
-        self.on_transcription: Optional[Any] = None
+        self.on_transcription: Any | None = None
 
         # Track background transcription tasks so they can complete even during interruptions
         self._transcription_tasks: list[asyncio.Task] = []
@@ -463,7 +463,7 @@ async def process_frame(self, frame: Frame, direction: FrameDirection) -> None:
         # Clean up completed tasks
         self._transcription_tasks = [t for t in self._transcription_tasks if not t.done()]
 
-    async def transcribe(self, timestamp: str, turn_id: Optional[int] = None) -> Optional[str]:
+    async def transcribe(self, timestamp: str, turn_id: int | None = None) -> str | None:
         """Transcribe audio from the collector using chat completions.
 
         This method can be called directly from event handlers or via frame processing.
@@ -479,9 +479,7 @@ async def transcribe(self, timestamp: str, turn_id: Optional[int] = None) -> Opt
         audio_data = self._audio_collector.peek_buffered_audio()
         return await self._transcribe_audio(audio_data, timestamp, turn_id)
 
-    async def _transcribe_audio(
-        self, audio_data: bytes, timestamp: str, turn_id: Optional[int] = None
-    ) -> Optional[str]:
+    async def _transcribe_audio(self, audio_data: bytes, timestamp: str, turn_id: int | None = None) -> str | None:
         """Transcribe pre-captured audio data using chat completions.
 
         This method takes audio data directly instead of reading from the collector,

src/eva/assistant/pipeline/nvidia_stt.py

@@ -10,7 +10,7 @@
 import json
 import ssl
 import time
-from typing import AsyncGenerator, Optional
+from collections.abc import AsyncGenerator
 
 import websockets
 from loguru import logger
@@ -48,7 +48,7 @@ def __init__(
         self,
         *,
         url: str = "ws://localhost:8080",
-        api_key: Optional[str] = None,
+        api_key: str | None = None,
         sample_rate: int = 16000,
         verify: bool = True,
         **kwargs,
@@ -58,7 +58,7 @@ def __init__(
         self._api_key = api_key
         self._verify = verify
         self._websocket = None
-        self._receive_task: Optional[asyncio.Task] = None
+        self._receive_task: asyncio.Task | None = None
         self._ready = False
 
     def can_generate_metrics(self) -> bool:
@@ -135,7 +135,7 @@ async def _connect_websocket(self):
             self._websocket = await websockets.connect(
                 self._url,
                 ssl=ssl_context,
-                additional_headers=extra_headers if extra_headers else None,
+                additional_headers=extra_headers or None,
             )
             self._ready = False
 
@@ -149,7 +149,7 @@ async def _connect_websocket(self):
                 else:
                     logger.warning(f"{self} unexpected initial message: {data}")
                     self._ready = True
-            except asyncio.TimeoutError:
+            except TimeoutError:
                 logger.warning(f"{self} timeout waiting for ready, proceeding")
                 self._ready = True
 

src/eva/assistant/pipeline/realtime_llm.py

@@ -14,7 +14,7 @@
 import struct
 import time
 from dataclasses import dataclass
-from typing import Any, Optional
+from typing import Any
 
 from pipecat.frames.frames import Frame, InputAudioRawFrame, VADUserStartedSpeakingFrame, VADUserStoppedSpeakingFrame
 from pipecat.processors.frame_processor import FrameDirection
@@ -75,14 +75,14 @@ def __init__(self, *, audit_log: AuditLog, **kwargs: Any) -> None:
 
         # Assistant response accumulation (across audio_transcript_delta events)
         self._current_assistant_transcript_parts: list[str] = []
-        self._assistant_response_start_wall_ms: Optional[str] = None
+        self._assistant_response_start_wall_ms: str | None = None
 
         # Track whether we're mid-assistant-response (for interruption flushing)
         self._assistant_responding: bool = False
 
         # Track audio frame timing for VAD delay calculation
-        self._last_audio_frame_time: Optional[float] = None
-        self._vad_delay_ms: Optional[int] = None
+        self._last_audio_frame_time: float | None = None
+        self._vad_delay_ms: int | None = None
 
     async def process_frame(self, frame: Frame, direction: FrameDirection) -> None:
         """Track audio frame timing before passing to parent.
@@ -295,7 +295,7 @@ def _reset_assistant_state(self) -> None:
         self._assistant_responding = False
 
     @property
-    def last_vad_delay_ms(self) -> Optional[int]:
+    def last_vad_delay_ms(self) -> int | None:
         """Return the most recent VAD delay in milliseconds.
 
         This is the time between when audio frames stopped arriving and when