🚀 feat(ai): atualizar Gemini para 2.5 Pro e suportar respostas multipart do Vertex

- troca o modelo padrao do projeto para gemini-2.5-pro no settings e no .env.example\n- ajusta a cadeia de fallback para priorizar a linha 2.5 antes do 2.0 flash\n- trata respostas do Vertex com multiplos content.parts, combinando texto e function_call sem depender de response.text\n- adiciona cobertura dedicada para garantir o parse correto de respostas mistas do SDK

.env.example

@@ -4,7 +4,7 @@
 
 GOOGLE_PROJECT_ID=id_do_seu_projeto
 GOOGLE_LOCATION=loc_do_seu_projeto
-VERTEX_MODEL_NAME=gemini-2.5-flash
+VERTEX_MODEL_NAME=gemini-2.5-pro
 
 # ============================================
 # CONFIGURACOES DO BANCO DE DADOS (MYSQL - TOOLS)

app/core/settings.py

@@ -4,7 +4,7 @@ from pydantic_settings import BaseSettings
 class Settings(BaseSettings):
     google_project_id: str
     google_location: str = "us-central1"
-    vertex_model_name: str = "gemini-2.5-flash"
+    vertex_model_name: str = "gemini-2.5-pro"
 
     # Tools database (MySQL)
     db_host: str = "127.0.0.1"

app/services/ai/llm_service.py

@@ -27,7 +27,7 @@ class LLMService:
             LLMService._vertex_initialized = True
 
         configured = settings.vertex_model_name.strip()
-        fallback_models = ["gemini-2.5-flash", "gemini-2.0-flash-001", "gemini-1.5-pro"]
+        fallback_models = ["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.0-flash-001"]
         self.model_names = [configured] + [m for m in fallback_models if m != configured]
 
     def build_vertex_tools(self, tools: List[ToolDefinition]) -> Optional[List[Tool]]:
@@ -73,6 +73,30 @@ class LLMService:
             LLMService._models[model_name] = model
         return model
 
+    def _extract_response_payload(self, response) -> Dict[str, Any]:
+        candidate = response.candidates[0] if getattr(response, "candidates", None) else None
+        content = getattr(candidate, "content", None)
+        parts = list(getattr(content, "parts", None) or [])
+
+        tool_call = None
+        text_parts: list[str] = []
+        for part in parts:
+            function_call = getattr(part, "function_call", None)
+            if function_call is not None and tool_call is None:
+                tool_call = {
+                    "name": function_call.name,
+                    "arguments": dict(function_call.args),
+                }
+            text_value = getattr(part, "text", None)
+            if isinstance(text_value, str) and text_value.strip():
+                text_parts.append(text_value)
+
+        response_text = "\n".join(text_parts).strip() or None
+        return {
+            "response": response_text,
+            "tool_call": tool_call,
+        }
+
     async def generate_response(
         self,
         message: str,
@@ -106,21 +130,7 @@ class LLMService:
                 ) from last_error
             raise RuntimeError("Falha ao gerar resposta no Vertex AI.")
 
-        part = response.candidates[0].content.parts[0]
+        return self._extract_response_payload(response)
-
-        if part.function_call:
-            return {
-                "response": None,
-                "tool_call": {
-                    "name": part.function_call.name,
-                    "arguments": dict(part.function_call.args),
-                },
-            }
-
-        return {
-            "response": response.text,
-            "tool_call": None,
-        }
 
     async def warmup(self) -> None:
         """Preaquece conexao/modelo para reduzir latencia da primeira requisicao real."""

tests/test_llm_service.py

@@ -0,0 +1,59 @@
+import os
+import unittest
+from types import SimpleNamespace
+
+os.environ.setdefault("DEBUG", "false")
+
+from app.services.ai.llm_service import LLMService
+
+
+class LLMServiceResponseParsingTests(unittest.TestCase):
+    def test_extract_response_payload_supports_text_and_function_call_in_same_candidate(self):
+        service = LLMService.__new__(LLMService)
+        response = SimpleNamespace(
+            candidates=[
+                SimpleNamespace(
+                    content=SimpleNamespace(
+                        parts=[
+                            SimpleNamespace(text="Legal! Buscando carros de ate 70 mil para voce.", function_call=None),
+                            SimpleNamespace(
+                                text=None,
+                                function_call=SimpleNamespace(
+                                    name="consultar_estoque",
+                                    args={"preco_max": 70000.0},
+                                ),
+                            ),
+                        ]
+                    )
+                )
+            ]
+        )
+
+        payload = service._extract_response_payload(response)
+
+        self.assertEqual(payload["response"], "Legal! Buscando carros de ate 70 mil para voce.")
+        self.assertEqual(
+            payload["tool_call"],
+            {
+                "name": "consultar_estoque",
+                "arguments": {"preco_max": 70000.0},
+            },
+        )
+
+    def test_extract_response_payload_handles_text_only_candidate_without_response_text_accessor(self):
+        service = LLMService.__new__(LLMService)
+        response = SimpleNamespace(
+            candidates=[
+                SimpleNamespace(
+                    content=SimpleNamespace(
+                        parts=[
+                            SimpleNamespace(text="Resposta simples", function_call=None),
+                        ]
+                    )
+                )
+            ]
+        )
+
+        payload = service._extract_response_payload(response)
+
+        self.assertEqual(payload, {"response": "Resposta simples", "tool_call": None})