fix: follow-up fixes for detached pydantic.BaseModel schemas

CHANGELOG.md

@@ -33,7 +33,8 @@ These are the section headers that we use:
 
 ### Fixed
 
-- Fixed saving of models trained with `ArgillaTrainer` with a `peft_config` parameter. ([#3795](https://github.com/argilla-io/argilla/pull/3795))
+- Fixed saving of models trained with `ArgillaTrainer` with a `peft_config` parameter ([#3795](https://github.com/argilla-io/argilla/pull/3795)).
+- Fixed backwards compatibility on `from_huggingface` when loading a `FeedbackDataset` from the Hugging Face Hub that was previously dumped using another version of Argilla, starting at 1.8.0, when it was first introduced ([#3829](https://github.com/argilla-io/argilla/pull/3829)).
 
 ## [1.16.0](https://github.com/argilla-io/argilla/compare/v1.15.1...v1.16.0)
 

src/argilla/client/feedback/config.py

@@ -12,6 +12,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
+import json
 import re
 import warnings
 from typing import List, Optional
@@ -43,15 +44,17 @@ def to_yaml(self) -> str:
         return dump(self.dict())
 
     @classmethod
-    def from_yaml(cls, yaml: str) -> "DatasetConfig":
-        yaml = re.sub(r"(\n\s*|)id: !!python/object:uuid\.UUID\s+int: \d+", "", yaml)
-        yaml = load(yaml, Loader=SafeLoader)
+    def from_yaml(cls, yaml_str: str) -> "DatasetConfig":
+        yaml_str = re.sub(r"(\n\s*|)id: !!python/object:uuid\.UUID\s+int: \d+", "", yaml_str)
+        yaml_dict = load(yaml_str, Loader=SafeLoader)
         # Here for backwards compatibility
-        for field in yaml["fields"]:
+        for field in yaml_dict["fields"]:
+            field.pop("id", None)
             field.pop("settings", None)
-        for question in yaml["questions"]:
+        for question in yaml_dict["questions"]:
+            question.pop("id", None)
             question.pop("settings", None)
-        return cls(**yaml)
+        return cls(**yaml_dict)
 
 
 # TODO(alvarobartt): here for backwards compatibility, remove in 1.14.0
@@ -70,11 +73,44 @@ def to_json(self) -> str:
         return self.json()
 
     @classmethod
-    def from_json(cls, json: str) -> "DeprecatedDatasetConfig":
+    def from_json(cls, json_str: str) -> "DeprecatedDatasetConfig":
         warnings.warn(
             "`DatasetConfig` can just be loaded from YAML, so make sure that you are"
             " loading a YAML file instead of a JSON file. `DatasetConfig` will be dumped"
             " as YAML from now on, instead of JSON.",
             DeprecationWarning,
         )
-        return cls.parse_raw(json)
+        parsed_json = json.loads(json_str)
+        # Here for backwards compatibility
+        for field in parsed_json["fields"]:
+            # for 1.10.0, 1.9.0, and 1.8.0
+            field.pop("id", None)
+            field.pop("inserted_at", None)
+            field.pop("updated_at", None)
+            if "settings" not in field:
+                continue
+            field["type"] = field["settings"]["type"]
+            if "use_markdown" in field["settings"]:
+                field["use_markdown"] = field["settings"]["use_markdown"]
+            # for 1.12.0 and 1.11.0
+            field.pop("settings", None)
+        for question in parsed_json["questions"]:
+            # for 1.10.0, 1.9.0, and 1.8.0
+            question.pop("id", None)
+            question.pop("inserted_at", None)
+            question.pop("updated_at", None)
+            if "settings" not in question:
+                continue
+            question.update({"type": question["settings"]["type"]})
+            if question["type"] in ["rating", "ranking"]:
+                question["values"] = [option["value"] for option in question["settings"]["options"]]
+            elif question["type"] in ["label_selection", "multi_label_selection"]:
+                if all(option["value"] == option["text"] for option in question["settings"]["options"]):
+                    question["labels"] = [option["value"] for option in question["settings"]["options"]]
+                else:
+                    question["labels"] = {option["value"]: option["text"] for option in question["settings"]["options"]}
+                if "visible_labels" in question["settings"]:
+                    question["visible_labels"] = question["settings"]["visible_labels"]
+            # for 1.12.0 and 1.11.0
+            question.pop("settings", None)
+        return cls(**parsed_json)

src/argilla/client/feedback/dataset/base.py

@@ -23,7 +23,7 @@
     FeedbackRecord,
     FieldSchema,
 )
-from argilla.client.feedback.schemas.types import AllowedQuestionTypes
+from argilla.client.feedback.schemas.types import AllowedFieldTypes, AllowedQuestionTypes
 from argilla.client.feedback.training.schemas import (
     TrainingTaskForChatCompletion,
     TrainingTaskForDPO,
@@ -43,7 +43,6 @@
     from datasets import Dataset
 
     from argilla.client.feedback.schemas.types import (
-        AllowedFieldTypes,
         AllowedRemoteFieldTypes,
         AllowedRemoteQuestionTypes,
     )
@@ -58,8 +57,8 @@ class FeedbackDatasetBase(ABC, HuggingFaceDatasetMixin):
     def __init__(
         self,
         *,
-        fields: Union[List["AllowedFieldTypes"], List["AllowedRemoteFieldTypes"]],
-        questions: Union[List["AllowedQuestionTypes"], List["AllowedRemoteQuestionTypes"]],
+        fields: Union[List[AllowedFieldTypes], List["AllowedRemoteFieldTypes"]],
+        questions: Union[List[AllowedQuestionTypes], List["AllowedRemoteQuestionTypes"]],
         guidelines: Optional[str] = None,
     ) -> None:
         """Initializes a `FeedbackDatasetBase` instance locally.
@@ -84,17 +83,21 @@ def __init__(
         any_required = False
         unique_names = set()
         for field in fields:
-            if not isinstance(field, FieldSchema):
-                raise TypeError(f"Expected `fields` to be a list of `FieldSchema`, got {type(field)} instead.")
+            if not isinstance(field, AllowedFieldTypes):
+                raise TypeError(
+                    f"Expected `fields` to be a list of `{AllowedFieldTypes.__name__}`, got {type(field)} instead."
+                )
             if field.name in unique_names:
                 raise ValueError(f"Expected `fields` to have unique names, got {field.name} twice instead.")
             unique_names.add(field.name)
             if not any_required and field.required:
                 any_required = True
+
         if not any_required:
-            raise ValueError("At least one `FieldSchema` in `fields` must be required (`required=True`).")
+            raise ValueError("At least one field in `fields` must be required (`required=True`).")
+
         self._fields = fields
-        self._fields_schema = None
+        self._fields_schema = generate_pydantic_schema(self.fields)
 
         if not isinstance(questions, list):
             raise TypeError(f"Expected `questions` to be a list, got {type(questions)} instead.")
@@ -113,8 +116,10 @@ def __init__(
             unique_names.add(question.name)
             if not any_required and question.required:
                 any_required = True
+
         if not any_required:
             raise ValueError("At least one question in `questions` must be required (`required=True`).")
+
         self._questions = questions
 
         if guidelines is not None:
@@ -126,6 +131,7 @@ def __init__(
                 raise ValueError(
                     "Expected `guidelines` to be either None (default) or a non-empty string, minimum length is 1."
                 )
+
         self._guidelines = guidelines
 
     @property
@@ -140,11 +146,11 @@ def guidelines(self) -> str:
         return self._guidelines
 
     @property
-    def fields(self) -> Union[List["AllowedFieldTypes"], List["AllowedRemoteFieldTypes"]]:
+    def fields(self) -> Union[List[AllowedFieldTypes], List["AllowedRemoteFieldTypes"]]:
         """Returns the fields that define the schema of the records in the dataset."""
         return self._fields
 
-    def field_by_name(self, name: str) -> Union["AllowedFieldTypes", "AllowedRemoteFieldTypes"]:
+    def field_by_name(self, name: str) -> Union[AllowedFieldTypes, "AllowedRemoteFieldTypes"]:
         """Returns the field by name if it exists. Othewise a `ValueError` is raised.
 
         Args:

src/argilla/client/feedback/dataset/local.py

@@ -17,18 +17,20 @@
 from argilla.client.feedback.constants import FETCHING_BATCH_SIZE
 from argilla.client.feedback.dataset.base import FeedbackDatasetBase
 from argilla.client.feedback.dataset.mixins import ArgillaMixin, UnificationMixin
+from argilla.client.feedback.schemas.fields import TextField
+from argilla.client.feedback.schemas.types import AllowedQuestionTypes
 
 if TYPE_CHECKING:
     from argilla.client.feedback.schemas.records import FeedbackRecord
-    from argilla.client.feedback.schemas.types import AllowedFieldTypes, AllowedQuestionTypes
+    from argilla.client.feedback.schemas.types import AllowedFieldTypes
 
 
 class FeedbackDataset(FeedbackDatasetBase, ArgillaMixin, UnificationMixin):
     def __init__(
         self,
         *,
         fields: List["AllowedFieldTypes"],
-        questions: List["AllowedQuestionTypes"],
+        questions: List[AllowedQuestionTypes],
         guidelines: Optional[str] = None,
     ) -> None:
         """Initializes a `FeedbackDataset` instance locally.

src/argilla/client/feedback/dataset/mixins.py

@@ -175,13 +175,15 @@ def push_to_argilla(
         except Exception as e:
             raise Exception(f"Failed while creating the `FeedbackDataset` in Argilla with exception: {e}") from e
 
-        fields = self.__add_fields(client=httpx_client, id=argilla_id)
+        # TODO(alvarobartt): re-use ArgillaMixin components when applicable
+        self.__add_fields(client=httpx_client, id=argilla_id)
+        fields = self.__get_fields(client=httpx_client, id=argilla_id)
 
-        questions = self.__add_questions(client=httpx_client, id=argilla_id)
+        self.__add_questions(client=httpx_client, id=argilla_id)
+        questions = self.__get_questions(client=httpx_client, id=argilla_id)
         question_name_to_id = {question.name: question.id for question in questions}
 
         self.__publish_dataset(client=httpx_client, id=argilla_id)
-
         self.__push_records(
             client=httpx_client, id=argilla_id, show_progress=show_progress, question_name_to_id=question_name_to_id
         )

src/argilla/client/feedback/dataset/remote/base.py

@@ -134,9 +134,7 @@ def __init__(
             TypeError: if `guidelines` is not None and not a string.
             ValueError: if `guidelines` is an empty string.
         """
-        self._fields = fields
-        self._questions = questions
-        self._guidelines = guidelines
+        super().__init__(fields=fields, questions=questions, guidelines=guidelines)
 
         self._client = client  # Required to be able to use `allowed_for_roles` decorator
         self._id = id

src/argilla/client/feedback/dataset/remote/dataset.py

@@ -24,14 +24,14 @@
 from argilla.client.feedback.schemas.remote.records import RemoteFeedbackRecord
 from argilla.client.sdk.users.models import UserRole
 from argilla.client.sdk.v1.datasets import api as datasets_api_v1
-from argilla.client.sdk.v1.datasets.models import FeedbackResponseStatusFilter
 from argilla.client.utils import allowed_for_roles
 
 if TYPE_CHECKING:
     from uuid import UUID
 
     import httpx
 
+    from argilla.client.feedback.schemas.enums import ResponseStatusFilter
     from argilla.client.feedback.schemas.types import AllowedRemoteFieldTypes, AllowedRemoteQuestionTypes
     from argilla.client.sdk.v1.datasets.models import FeedbackRecordsModel
     from argilla.client.workspaces import Workspace
@@ -145,7 +145,7 @@ def __init__(
         )
 
     def filter_by(
-        self, response_status: Union[FeedbackResponseStatusFilter, List[FeedbackResponseStatusFilter]]
+        self, response_status: Union["ResponseStatusFilter", List["ResponseStatusFilter"]]
     ) -> FilteredRemoteFeedbackDataset:
         """Filters the current `RemoteFeedbackDataset` based on the `response_status` of
         the responses of the records in Argilla. This method creates a new class instance

src/argilla/client/feedback/schemas/enums.py

@@ -25,3 +25,16 @@ class QuestionTypes(str, Enum):
     label_selection = "label_selection"
     multi_label_selection = "multi_label_selection"
     ranking = "ranking"
+
+
+class ResponseStatus(str, Enum):
+    draft = "draft"
+    submitted = "submitted"
+    discarded = "discarded"
+
+
+class ResponseStatusFilter(str, Enum):
+    draft = "draft"
+    submitted = "submitted"
+    discarded = "discarded"
+    missing = "missing"

src/argilla/client/feedback/schemas/fields.py

@@ -12,7 +12,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-from abc import abstractproperty
+from abc import ABC, abstractmethod
 from typing import Any, Dict, Literal, Optional
 
 from pydantic import BaseModel, Extra, Field, validator
@@ -21,7 +21,7 @@
 from argilla.client.feedback.schemas.validators import title_must_have_value
 
 
-class FieldSchema(BaseModel):
+class FieldSchema(BaseModel, ABC):
     """Base schema for the `FeedbackDataset` fields.
 
     Args:
@@ -52,12 +52,13 @@ class Config:
         extra = Extra.forbid
         exclude = {"type"}
 
-    @abstractproperty
+    @property
+    @abstractmethod
     def server_settings(self) -> Dict[str, Any]:
         """Abstract property that should be implemented by the classes that inherit from
         this one, and that will be used to create the `FeedbackDataset` in Argilla.
         """
-        raise NotImplementedError
+        ...
 
     def to_server_payload(self) -> Dict[str, Any]:
         """Method that will be used to create the payload that will be sent to Argilla

src/argilla/client/feedback/schemas/questions.py

@@ -13,7 +13,7 @@
 #  limitations under the License.
 
 import warnings
-from abc import abstractproperty
+from abc import ABC, abstractmethod
 from typing import Any, Dict, List, Literal, Optional, Union
 
 from pydantic import BaseModel, Extra, Field, conint, conlist, root_validator, validator
@@ -23,7 +23,7 @@
 from argilla.client.feedback.schemas.validators import title_must_have_value
 
 
-class QuestionSchema(BaseModel):
+class QuestionSchema(BaseModel, ABC):
     """Base schema for the `FeedbackDataset` questions. Which means that all the questions
     in the dataset will have at least these fields.
 
@@ -58,12 +58,13 @@ class Config:
         extra = Extra.forbid
         exclude = {"type"}
 
-    @abstractproperty
+    @property
+    @abstractmethod
     def server_settings(self) -> Dict[str, Any]:
         """Abstract property that should be implemented by the classes that inherit from
         this one, and that will be used to create the `FeedbackDataset` in Argilla.
         """
-        raise NotImplementedError
+        ...
 
     def to_server_payload(self) -> Dict[str, Any]:
         """Method that will be used to create the payload that will be sent to Argilla

src/argilla/client/feedback/schemas/records.py

@@ -13,12 +13,13 @@
 #  limitations under the License.
 
 import warnings
-from enum import Enum
 from typing import TYPE_CHECKING, Any, Dict, List, Literal, Optional, Tuple, Union
 from uuid import UUID
 
 from pydantic import BaseModel, Extra, Field, PrivateAttr, StrictInt, StrictStr, conint, validator
 
+from argilla.client.feedback.schemas.enums import ResponseStatus
+
 if TYPE_CHECKING:
     from argilla.client.feedback.unification import UnifiedValueSchema
 
@@ -46,12 +47,6 @@ class ValueSchema(BaseModel):
     value: Union[StrictStr, StrictInt, List[str], List[RankingValueSchema]]
 
 
-class ResponseStatus(str, Enum):
-    draft = "draft"
-    submitted = "submitted"
-    discarded = "discarded"
-
-
 class ResponseSchema(BaseModel):
     """Schema for the `FeedbackRecord` response.
 
@@ -103,9 +98,7 @@ class SuggestionSchema(BaseModel):
     """Schema for the suggestions for the questions related to the record.
 
     Args:
-        question_id: ID of the question in Argilla. Defaults to None, and is automatically
-           fulfilled internally once the question is pushed to Argilla.
-        question_name: name of the question.
+        question_name: name of the question in the `FeedbackDataset`.
         type: type of the question. Defaults to None. Possible values are `model` or `human`.
         score: score of the suggestion. Defaults to None.
         value: value of the suggestion, which should match the type of the question.

src/argilla/client/feedback/schemas/remote/shared.py

@@ -12,15 +12,15 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-from abc import abstractmethod
+from abc import ABC, abstractmethod
 from typing import Optional, Type
 from uuid import UUID
 
 import httpx
 from pydantic import BaseModel
 
 
-class RemoteSchema(BaseModel):
+class RemoteSchema(BaseModel, ABC):
     id: Optional[UUID] = None
     client: Optional[httpx.Client] = None