feat: Cria o objeto GPTModel a partir do modelo do fine tuning

openiziai/fine_tuning.py

@@ -1,6 +1,6 @@
 from enum import Enum
 from pathlib import Path
-from typing import Any
+from typing import Any, Optional
 
 from openai import OpenAI
 from pydantic import (
@@ -11,6 +11,7 @@
     field_validator,
 )
 
+from openiziai.schemas import GPTModel
 from openiziai.task import Task
 
 
@@ -31,6 +32,7 @@ class FineTuning(BaseModel):
     _file_id: str = PrivateAttr(default=None)
     _job_id: str = PrivateAttr(default=None)
     _job_status: JobStatus = PrivateAttr(default=None)
+    _model: GPTModel = PrivateAttr(default=None)
 
     model_config = ConfigDict(arbitrary_types_allowed=True)
 
@@ -94,3 +96,20 @@ def status(self) -> str:
             self._job_status = JobStatus(_job_status)
 
         return self._job_status.name
+
+    # TODO: Consertar mocks para testar a propriedade
+    @property
+    def model(self) -> Optional[GPTModel]:
+        if self._model:
+            return self._model
+
+        model_name = self.client.fine_tuning.jobs.retrieve(
+            self.job_id
+        ).fine_tuned_model
+        if not model_name:
+            print(f'Modelo não disponível. Status: {self.status}')
+            return None
+
+        self._model = GPTModel(
+            name=model_name, base_model=self.base_model, task=self.task
+        )

openiziai/schemas.py

@@ -1,6 +1,12 @@
 import sys
+from datetime import datetime
 from typing import Any, Protocol
 
+from pydantic import Field
+from pydantic.dataclasses import dataclass
+
+from openiziai.task import Task
+
 if sys.version_info < (3, 12):
     from typing_extensions import TypedDict
 else:
@@ -14,3 +20,10 @@ class DataDict(TypedDict):
 class Pipeline(Protocol):
     def run(self): ...
 
+
+@dataclass
+class GPTModel:
+    name: str
+    task: Task
+    base_model: str
+    created_at: datetime = Field(init=False, default_factory=datetime.now)