fix: corrige autenticação e ciclo de streaming do DeepAgents

src/voice_assistant/agent/deep_agent.py

@@ -16,7 +16,10 @@ Does not manage:
 
 from __future__ import annotations
 
-from collections.abc import AsyncGenerator
+import asyncio
+import math
+from collections.abc import AsyncGenerator, Iterator
+from typing import Any
 from contextlib import aclosing
 
 from deepagents import create_deep_agent
@@ -41,6 +44,23 @@ de ferramentas que não foram executadas.
 """
 
 
+def _text_fragments(content: str | list[str | dict[str, Any]]) -> Iterator[str]:
+    """Extract text without exposing reasoning, tools or other content blocks."""
+    if isinstance(content, str):
+        if content:
+            yield content
+        return
+
+    for block in content:
+        if isinstance(block, str):
+            if block:
+                yield block
+        elif isinstance(block, dict) and block.get("type") == "text":
+            text = block.get("text")
+            if isinstance(text, str) and text:
+                yield text
+
+
 class DeepAgentsAdapter(StreamingAgentAdapter):
     def __init__(
         self,
@@ -51,6 +71,10 @@ class DeepAgentsAdapter(StreamingAgentAdapter):
         max_completion_tokens: int,
         timeout_seconds: float,
         system_prompt: str = DEFAULT_SYSTEM_PROMPT,
+        *,
+        auth_type: str = "INSTANCE_PRINCIPAL",
+        auth_profile: str = "DEFAULT",
+        auth_file_location: str = "~/.oci/config",
     ) -> None:
         if not region:
             raise ValueError("OCI region is required.")
@@ -61,8 +85,15 @@ class DeepAgentsAdapter(StreamingAgentAdapter):
         if not model_id:
             raise ValueError("OCI model_id is required.")
 
-        if timeout_seconds <= 0:
-            raise ValueError("timeout_seconds must be positive.")
+        if (
+            isinstance(timeout_seconds, bool)
+            or not isinstance(timeout_seconds, (int, float))
+            or not math.isfinite(timeout_seconds)
+            or timeout_seconds <= 0
+        ):
+            raise ValueError("timeout_seconds must be finite and positive.")
+
+        self._stream_event_timeout = float(timeout_seconds)
 
         self.model = ChatOCIGenAI(
             model_id=model_id,
@@ -71,7 +102,9 @@ class DeepAgentsAdapter(StreamingAgentAdapter):
                 f"https://inference.generativeai."
                 f"{region}.oci.oraclecloud.com"
             ),
-            auth_type="INSTANCE_PRINCIPAL",
+            auth_type=auth_type,
+            auth_profile=auth_profile,
+            auth_file_location=auth_file_location,
             is_stream=True,
             model_kwargs={
                 "temperature": temperature,
@@ -142,27 +175,30 @@ class DeepAgentsAdapter(StreamingAgentAdapter):
 
         config = self._config(session_id)
 
-        async for message, metadata in self.agent.astream(
-            {
-                "messages": [
-                    HumanMessage(content=text)
-                ]
-            },
+        events = self.agent.astream(
+            {"messages": [HumanMessage(content=text)]},
             config=config,
             stream_mode="messages",
-        ):
-            # Ignore tools and subagent outputs.
-            if metadata.get("langgraph_node") != "model":
-                continue
+        )
+        async with aclosing(events):
+            while True:
+                try:
+                    async with asyncio.timeout(self._stream_event_timeout):
+                        message, metadata = await anext(events)
+                except StopAsyncIteration:
+                    break
+                except TimeoutError as exc:
+                    raise TimeoutError(
+                        "Deep Agent timed out waiting for the next stream event."
+                    ) from exc
 
-            # Ignore tool-call fragments.
-            if getattr(message, "tool_call_chunks", None):
-                continue
+                if metadata.get("langgraph_node") != "model":
+                    continue
+                if getattr(message, "tool_call_chunks", None):
+                    continue
 
-            content = message.content
-
-            if isinstance(content, str) and content:
-                yield content
+                for fragment in _text_fragments(message.content):
+                    yield fragment
 
 
     async def aclose(self) -> None: