[ENHANCEMENT / BUGFIX] argilla: manage responses for deleted users

argilla/docs/how_to_guides/record.md

@@ -453,7 +453,7 @@ dataset.records.log(records=updated_data)
     ```
 
 === "Update vectors"
-    When a new vector field is added to the dataset settings, or some value for the existing record vectors must updated, you can iterate over the records and update the vectors in the same way as the metadata. 
+    When a new vector field is added to the dataset settings, or some value for the existing record vectors must updated, you can iterate over the records and update the vectors in the same way as the metadata.
 
     ```python
     updated_records = []

argilla/src/argilla/_models/_record/_record.py

@@ -26,6 +26,7 @@
 
 FieldValue = Union[str, None]
 
+
 class RecordModel(ResourceModel):
     """Schema for the records of a `Dataset`"""
 

argilla/src/argilla/_models/_record/_response.py

@@ -12,12 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import warnings
 from enum import Enum
 from typing import Dict, Optional, Union, Any
 from uuid import UUID
 
-from pydantic import BaseModel, field_serializer, field_validator, Field
+from pydantic import BaseModel, field_serializer, Field, ConfigDict
 
 
 class ResponseStatus(str, Enum):
@@ -33,20 +32,10 @@ class UserResponseModel(BaseModel):
     status: ResponseStatus
     user_id: Optional[UUID] = Field(None, validate_default=True)
 
-    class Config:
-        validate_assignment = True
-
-    @field_validator("user_id")
-    @classmethod
-    def user_id_must_have_value(cls, user_id: Optional[UUID]):
-        if not user_id:
-            warnings.warn(
-                "`user_id` not provided, so it will be set to `None`. Which is not an"
-                " issue, unless you're planning to log the response in Argilla, as"
-                " it will be automatically set to the active `user_id`.",
-            )
-        return user_id
+    model_config = ConfigDict(
+        validate_assignment=True,
+    )
 
     @field_serializer("user_id", when_used="always")
-    def serialize_user_id(value: UUID) -> str:
+    def serialize_user_id(self, value: UUID) -> str:
         return str(value)

argilla/src/argilla/records/_dataset_records.py

@@ -31,7 +31,6 @@
 from argilla.settings._metadata import MetadataPropertyBase
 from argilla.settings._question import QuestionPropertyBase
 from argilla.suggestions import Suggestion
-from argilla.vectors import Vector
 
 if TYPE_CHECKING:
     from argilla.datasets import Dataset

argilla/src/argilla/responses.py

@@ -19,6 +19,7 @@
 from argilla._models import UserResponseModel, ResponseStatus as ResponseStatusModel
 from argilla._resource import Resource
 from argilla.settings import RankingQuestion
+from argilla.users import DELETED_USER
 
 if TYPE_CHECKING:
     from argilla import Argilla, Dataset, Record
@@ -143,7 +144,7 @@ def status(self, status: ResponseStatus) -> None:
         self._model.status = status
 
     @property
-    def user_id(self) -> UUID:
+    def user_id(self) -> Optional[UUID]:
         """Returns the user_id of the UserResponse"""
         return self._model.user_id
 
@@ -160,6 +161,10 @@ def answers(self) -> List[Response]:
     @classmethod
     def from_model(cls, model: UserResponseModel, dataset: "Dataset") -> "UserResponse":
         """Creates a UserResponse from a ResponseModel"""
+
+        if model.user_id is None:
+            model.user_id = DELETED_USER.id
+
         answers = cls.__model_as_response_list(model)
         for answer in answers:
             question = dataset.settings.question_by_name(answer.question_name)

argilla/src/argilla/users/__init__.py

@@ -11,7 +11,11 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+from uuid import UUID
 
 from argilla.users._resource import User
 
-__all__ = ["User"]
+__all__ = ["User", "DELETED_USER"]
+
+# This is the user id for the deleted user. Used when records contains responses from a user that has been deleted.
+DELETED_USER = User(id=UUID("00000000-0000-0000-0000-000000000000"), username="deleted")

argilla/src/argilla/users/_resource.py

@@ -15,7 +15,7 @@
 from typing import Optional
 from uuid import UUID
 
-from argilla import Workspace
+from argilla.workspaces import Workspace
 from argilla._api import UsersAPI
 from argilla._models import UserModel, Role
 from argilla._resource import Resource

argilla/tests/unit/export/test_record_export_import_compatibillity.py

@@ -23,7 +23,6 @@
 
 @pytest.fixture
 def record():
-
     return rg.Record(
         id=uuid.uuid4(),
         fields={"text": "Hello World, how are you?"},
@@ -34,7 +33,6 @@ def record():
         responses=[rg.Response("label", "positive", user_id=uuid.uuid4())],
         metadata={"source": "twitter", "language": "en"},
         vectors={"text": [0, 0, 0]},
-
     )
 
 

argilla/tests/unit/test_resources/test_records.py

@@ -30,11 +30,11 @@ def test_record_repr(self):
             responses=[Response(question_name="question", value="answer", user_id=user_id)],
         )
         assert (
-                record.__repr__() == f"Record(id={record_id},"
-                                     "fields={'name': 'John', 'age': '30'},"
-                                     "metadata={'key': 'value'},"
-                                     "suggestions={'question': {'value': 'answer', 'score': None, 'agent': None}},"
-                                     f"responses={{'question': [{{'value': 'answer'}}]}})"
+            record.__repr__() == f"Record(id={record_id},"
+            "fields={'name': 'John', 'age': '30'},"
+            "metadata={'key': 'value'},"
+            "suggestions={'question': {'value': 'answer', 'score': None, 'agent': None}},"
+            f"responses={{'question': [{{'value': 'answer'}}]}})"
         )
 
     def test_update_record_metadata_by_key(self):
@@ -62,4 +62,3 @@ def test_update_record_vectors(self):
 
         record.vectors["new-vector"] = [1.0, 2.0, 3.0]
         assert record.vectors == {"vector": [1.0, 2.0, 3.0], "new-vector": [1.0, 2.0, 3.0]}
-

argilla/tests/unit/test_resources/test_responses.py

@@ -16,7 +16,10 @@
 
 import pytest
 
-from argilla import UserResponse, Response
+from argilla import Response, Dataset, Settings, TextQuestion, Workspace
+from argilla._models import UserResponseModel
+from argilla.responses import ResponseStatus, UserResponse
+from argilla.users import DELETED_USER
 
 
 class TestResponses:
@@ -85,3 +88,13 @@ def test_create_user_response_with_multiple_user_id(self):
                     Response(question_name="other-question", value="answer", user_id=other_user_id),
                 ],
             )
+
+    def test_create_user_response_from_model_without_user_id(self):
+        mock_workspace = Workspace(id=uuid.uuid4(), name="workspace")
+        dataset = Dataset(settings=Settings(questions=[TextQuestion(name="question")]), workspace=mock_workspace)
+        response = UserResponse.from_model(
+            UserResponseModel(values={"question": {"value": "answer"}}, user_id=None, status=ResponseStatus.draft),
+            dataset=dataset,
+        )
+
+        assert response.user_id == DELETED_USER.id