Merge pull request #298 from edenai/pydantic-update

Update pydantic to 2.7.4

edenai_apis/features/image/explicit_content/category.py

@@ -19,7 +19,7 @@ class ACategoryType(str):
     pass
 
 
-class CategoryType(ACategoryType, Enum):
+class CategoryType(str,  Enum):
     """This enum are used to categorize the explicit content extracted from the text"""
 
     Toxic = "Toxic"
@@ -37,7 +37,7 @@ def list_available_type(cls):
         return [category for category in cls]
 
     @classmethod
-    def list_choices(cls) -> Dict["ACategoryType", SubCategoryBase]:
+    def list_choices(cls) -> Dict[str, SubCategoryBase]:
         return {
             cls.Toxic: ToxicSubCategoryType,
             cls.Content: ContentSubCategoryType,

edenai_apis/features/image/explicit_content/explicit_content_dataclass.py

@@ -22,8 +22,11 @@
     SafeSubCategoryType,
     OtherSubCategoryType,
 )
+from edenai_apis.utils.combine_enums import combine_enums
 
-SubCategoryType = Union[
+
+SubCategoryType = combine_enums(
+    "SubCategoryType",
     ToxicSubCategoryType,
     ContentSubCategoryType,
     SexualSubCategoryType,
@@ -33,7 +36,7 @@
     HateAndExtremismSubCategoryType,
     SafeSubCategoryType,
     OtherSubCategoryType,
-]
+)
 
 
 class ExplicitItem(BaseModel):

edenai_apis/features/text/anonymization/anonymization_dataclass.py

@@ -1,5 +1,13 @@
 from typing import Dict, Optional, Sequence, Union
-from pydantic import BaseModel, ConfigDict, Field, FieldSerializationInfo, field_serializer, field_validator, model_validator
+from pydantic import (
+    BaseModel,
+    ConfigDict,
+    Field,
+    FieldSerializationInfo,
+    field_serializer,
+    field_validator,
+    model_validator,
+)
 
 from edenai_apis.features.text.anonymization.category import (
     CategoryType,
@@ -16,8 +24,15 @@
 )
 from typing import Dict, Optional, Sequence, Union
 
-from pydantic import BaseModel, ConfigDict, Field, FieldSerializationInfo, field_serializer, field_validator, \
-    model_validator
+from pydantic import (
+    BaseModel,
+    ConfigDict,
+    Field,
+    FieldSerializationInfo,
+    field_serializer,
+    field_validator,
+    model_validator,
+)
 
 from edenai_apis.features.text.anonymization.category import (
     CategoryType,
@@ -32,8 +47,11 @@
     LocationInformationSubCategoryType,
     OtherSubCategoryType,
 )
+from edenai_apis.utils.combine_enums import combine_enums
 
-SubCategoryType = Union[
+
+SubCategoryType = combine_enums(
+    "SubCategoryType",
     FinancialInformationSubCategoryType,
     PersonalInformationSubCategoryType,
     IdentificationNumbersSubCategoryType,
@@ -42,7 +60,7 @@
     DateAndTimeSubCategoryType,
     LocationInformationSubCategoryType,
     OtherSubCategoryType,
-]
+)
 
 
 class AnonymizationEntity(BaseModel):
@@ -94,9 +112,9 @@ def round_confidence_score(cls, v):
             return round(v, 3)
         return v
 
-    @field_serializer('subcategory', mode="plain", when_used="always")
+    @field_serializer("subcategory", mode="plain", when_used="always")
     def serialize_subcategory(self, value: SubCategoryType, _: FieldSerializationInfo):
-        return getattr(value, 'value', None)
+        return getattr(value, "value", None)
 
 
 class AnonymizationDataClass(BaseModel):

edenai_apis/features/text/anonymization/category.py

@@ -14,11 +14,7 @@
 )
 
 
-class ACategoryType(str):
-    pass
-
-
-class CategoryType(ACategoryType, Enum):
+class CategoryType(str, Enum):
     """This enum are used to categorize the entities extracted from the text."""
 
     PersonalInformation = "PersonalInformation"
@@ -35,7 +31,7 @@ def list_available_type(cls):
         return [category for category in cls]
 
     @classmethod
-    def list_choices(cls) -> Dict["ACategoryType", SubCategoryBase]:
+    def list_choices(cls) -> Dict[str, SubCategoryBase]:
         return {
             cls.PersonalInformation: PersonalInformationSubCategoryType,
             cls.FinancialInformation: FinancialInformationSubCategoryType,

edenai_apis/features/text/moderation/category.py

@@ -15,11 +15,7 @@
 )
 
 
-class ACategoryType(str):
-    pass
-
-
-class CategoryType(ACategoryType, Enum):
+class CategoryType(str, Enum):
     """This enum are used to categorize the explicit content extracted from the text"""
 
     Toxic = "Toxic"
@@ -37,7 +33,7 @@ def list_available_type(cls):
         return [category for category in cls]
 
     @classmethod
-    def list_choices(cls) -> Dict["ACategoryType", SubCategoryBase]:
+    def list_choices(cls) -> Dict[str, SubCategoryBase]:
         return {
             cls.Toxic: ToxicSubCategoryType,
             cls.Content: ContentSubCategoryType,

edenai_apis/features/text/moderation/moderation_dataclass.py

@@ -1,7 +1,15 @@
 from enum import Enum
-from typing import Sequence, Union
+from typing import Sequence, Union, Type
 
-from pydantic import BaseModel, ConfigDict, Field, StrictStr, field_validator, FieldSerializationInfo, field_serializer
+from pydantic import (
+    BaseModel,
+    ConfigDict,
+    Field,
+    StrictStr,
+    field_validator,
+    FieldSerializationInfo,
+    field_serializer,
+)
 
 from edenai_apis.features.text.moderation.category import (
     CategoryType,
@@ -17,8 +25,10 @@
     SafeSubCategoryType,
     OtherSubCategoryType,
 )
+from edenai_apis.utils.combine_enums import combine_enums
 
-SubCategoryType = Union[
+SubCategoryType = combine_enums(
+    "SubCategoryType",
     ToxicSubCategoryType,
     ContentSubCategoryType,
     SexualSubCategoryType,
@@ -28,7 +38,8 @@
     HateAndExtremismSubCategoryType,
     SafeSubCategoryType,
     OtherSubCategoryType,
-]
+)
+
 class TextModerationCategoriesMicrosoftEnum(Enum):
     Category1 = "sexually explicit"
     Category2 = "sexually suggestive"
@@ -43,14 +54,16 @@ class TextModerationItem(BaseModel):
     likelihood_score: float
 
     model_config = ConfigDict(use_enum_values=True)
-    @field_serializer('subcategory', mode="plain", when_used="always")
+
+    @field_serializer("subcategory", mode="plain", when_used="always")
     def serialize_subcategory(self, value: SubCategoryType, _: FieldSerializationInfo):
-        return getattr(value, 'value', None)
+        return getattr(value, "value", None)
+
 
 class ModerationDataClass(BaseModel):
     nsfw_likelihood: int
     items: Sequence[TextModerationItem] = Field(default_factory=list)
-    nsfw_likelihood_score : float
+    nsfw_likelihood_score: float
 
     @field_validator("nsfw_likelihood")
     @classmethod
@@ -59,7 +72,6 @@ def check_min_max(cls, value):
             raise ValueError("Likelihood walue should be between 0 and 5")
         return value
 
-
     @staticmethod
     def calculate_nsfw_likelihood(items: Sequence[TextModerationItem]):
         if len(items) == 0:

edenai_apis/features/text/moderation/subcategory.py

@@ -38,6 +38,7 @@ def choose_label(cls, label: str) -> "SubCategoryBase":
             f"Unknown label {label}. Only {cls.list_choices().values()} are allowed."
         )
 
+
 class ToxicSubCategoryType(SubCategoryBase, Enum):
     Insult = "Insult"
     Obscene = "Obscene"
@@ -57,6 +58,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Toxic: SubCategoryPattern.Toxic.TOXIC,
         }
 
+
 class ContentSubCategoryType(SubCategoryBase, Enum):
     MiddleFinger = "MiddleFinger"
     PublicSafety = "PublicSafety"
@@ -80,6 +82,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Legal: SubCategoryPattern.Content.LEGAL,
         }
 
+
 class SexualSubCategoryType(SubCategoryBase, Enum):
     SexualActivity = "SexualActivity"
     SexualSituations = "SexualSituations"
@@ -103,6 +106,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Sexual: SubCategoryPattern.Sexual.SEXUAL,
         }
 
+
 class ViolenceSubCategoryType(SubCategoryBase, Enum):
     GraphicViolenceOrGore = "GraphicViolenceOrGore"
     PhysicalViolence = "PhysicalViolence"
@@ -118,6 +122,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Violence: SubCategoryPattern.Violence.VIOLENCE,
         }
 
+
 class DrugAndAlcoholSubCategoryType(SubCategoryBase, Enum):
     DrugProducts = "DrugProducts"
     DrugUse = "DrugUse"
@@ -137,6 +142,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Drinking: SubCategoryPattern.DrugAndAlcohol.DRINKING,
         }
 
+
 class FinanceSubCategoryType(SubCategoryBase, Enum):
     Gambling = "Gambling"
     Finance = "Finance"
@@ -150,6 +156,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.MoneyContent: SubCategoryPattern.Finance.MONEY_CONTENT,
         }
 
+
 class HateAndExtremismSubCategoryType(SubCategoryBase, Enum):
     Hate = "Hate"
     Harassment = "Harassment"
@@ -167,6 +174,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Racy: SubCategoryPattern.HateAndExtremism.RACY,
         }
 
+
 class SafeSubCategoryType(SubCategoryBase, Enum):
     Safe = "Safe"
     NotSafe = "NotSafe"
@@ -178,6 +186,7 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.NotSafe: SubCategoryPattern.Safe.NOT_SAFE,
         }
 
+
 class OtherSubCategoryType(SubCategoryBase, Enum):
     Spoof = "Spoof"
     Religion = "Religion"
@@ -191,4 +200,4 @@ def list_choices(cls) -> Dict["SubCategoryBase", List[str]]:
             cls.Religion: SubCategoryPattern.Other.RELIGION,
             cls.Offensive: SubCategoryPattern.Other.OFFENSIVE,
             cls.Other: SubCategoryPattern.Other.OTHER,
-        }
+        }

edenai_apis/tests/features/dataclasses/test_text_moderation_dataclass.py

@@ -43,8 +43,30 @@ class TestTextModeration:
             ),
         ],
     )
-    def test_valid_value_check_min_mac_nsfw(self, nsfw_likelihood, nsfw_likelihood_score):
+    def test_valid_value_check_min_mac_nsfw(
+        self, nsfw_likelihood, nsfw_likelihood_score
+    ):
         try:
-            ModerationDataClass(nsfw_likelihood=nsfw_likelihood, items=[], nsfw_likelihood_score=nsfw_likelihood_score)
+            ModerationDataClass(
+                nsfw_likelihood=nsfw_likelihood,
+                items=[],
+                nsfw_likelihood_score=nsfw_likelihood_score,
+            )
         except ValueError:
             pytest.fail(f"{nsfw_likelihood} value doesn't raises a ValueError")
+
+    def test_text_moderation_items(self):
+
+        ModerationDataClass(
+            nsfw_likelihood=0,
+            items=[
+                {
+                    "label": "hi",
+                    "category": "Toxic",
+                    "likelihood": 1,
+                    "subcategory": "Toxic",
+                    "likelihood_score": 0.5,
+                }
+            ],
+            nsfw_likelihood_score=0.0,
+        )

edenai_apis/tests/test_providers_errors.py

@@ -6,7 +6,7 @@
 import pytest
 from apis.amazon.errors import ERRORS as amazon_errors
 from apis.google.errors import ERRORS as google_errors
-from apis.ibm.errors import ERRORS as ibm_errors
+
 from apis.microsoft.errors import ERRORS as microsoft_errors
 from features.audio.speech_to_text_async.speech_to_text_async_args import (
     data_path as audio_data_path,

edenai_apis/utils/combine_enums.py

@@ -0,0 +1,8 @@
+from typing import Type
+from enum import Enum
+
+def combine_enums(name: str, *enum_classes: Type[Enum]) -> Enum:
+    combined = {}
+    for enum_class in enum_classes:
+        combined.update(enum_class.__members__)
+    return Enum(name, combined)

requirements.in

@@ -5,7 +5,7 @@ pandas
 pdf2image
 pillow
 pycountry
-pydantic
+pydantic==2.7.4
 pylint
 mypy
 pydub
@@ -63,7 +63,7 @@ azure-core
 aleph_alpha_client==6.0.0
 
 # openai
-openai==1.46.0
+openai==1.55.3
 
 python-dotenv