bump to pydantic 2.0

Signed-off-by: ThibaultFy <50656860+ThibaultFy@users.noreply.github.com>

substra/sdk/backends/local/backend.py

@@ -135,7 +135,7 @@ def _check_metadata(metadata: Optional[Dict[str, str]]):
         if metadata is not None:
             if any([len(key) > _MAX_LEN_KEY_METADATA for key in metadata]):
                 raise exceptions.InvalidRequest("The key in metadata cannot be more than 50 characters", 400)
-            if any([len(value) > _MAX_LEN_VALUE_METADATA or len(value) == 0 for value in metadata.values()]):
+            if any([len(str(value)) > _MAX_LEN_VALUE_METADATA or len(str(value)) == 0 for value in metadata.values()]):
                 raise exceptions.InvalidRequest("Values in metadata cannot be empty or more than 100 characters", 400)
             if any("__" in key for key in metadata):
                 raise exceptions.InvalidRequest(
@@ -327,7 +327,7 @@ def _add_dataset(self, key, spec, spec_options=None):
                 "storage_address": dataset_description_path,
             },
             metadata=spec.metadata if spec.metadata else dict(),
-            logs_permission=logs_permission.dict(),
+            logs_permission=logs_permission.model_dump(),
         )
         return self._db.add(asset)
 
@@ -555,7 +555,7 @@ def update(self, key, spec, spec_options=None):
         asset_type = spec.__class__.type_
         asset = self.get(asset_type, key)
         data = asset.dict()
-        data.update(spec.dict())
+        data.update(spec.model_dump())
         updated_asset = models.SCHEMA_TO_MODEL[asset_type](**data)
         self._db.update(updated_asset)
         return
@@ -588,7 +588,7 @@ def _output_from_spec(outputs: Dict[str, schemas.ComputeTaskOutputSpec]) -> Dict
     """Convert a list of schemas.ComputeTaskOuput to a list of models.ComputeTaskOutput"""
     return {
         identifier: models.ComputeTaskOutput(
-            permissions=models.Permissions(process=output.permissions), transient=output.is_transient, value=None
+            permissions=models.Permissions(process=dict(output.permissions)), transient=output.is_transient, value=None
         )
         # default isNone (= outputs are not computed yet)
         for identifier, output in outputs.items()

substra/sdk/backends/local/compute/worker.py

@@ -84,7 +84,7 @@ def _save_cp_performances_as_json(self, compute_plan_key: str, path: Path):
         performances = self._db.get_performances(compute_plan_key)
 
         with (path).open("w", encoding="UTF-8") as json_file:
-            json.dump(performances.dict(), json_file, default=str)
+            json.dump(performances.model_dump(), json_file, default=str)
 
     def _get_asset_unknown_type(self, asset_key, possible_types: List[schemas.Type]) -> Tuple[Any, schemas.Type]:
         for asset_type in possible_types:

substra/sdk/models.py

@@ -173,8 +173,8 @@ class Function(_Model):
     permissions: Permissions
     metadata: Dict[str, str]
     creation_date: datetime
-    inputs: List[FunctionInput]
-    outputs: List[FunctionOutput]
+    inputs: List[schemas.FunctionInputSpec]
+    outputs: List[schemas.FunctionOutputSpec]
 
     description: _File
     function: _File
@@ -190,7 +190,7 @@ def dict_input_to_list(cls, v):
         if isinstance(v, dict):
             # Transform the inputs dict to a list
             return [
-                FunctionInput(
+                schemas.FunctionInputSpec(
                     identifier=identifier,
                     kind=function_input["kind"],
                     optional=function_input["optional"],
@@ -246,7 +246,7 @@ class InputRef(schemas._PydanticConfig):
     parent_task_output_identifier: Optional[str] = None
 
     # either (asset_key) or (parent_task_key, parent_task_output_identifier) must be specified
-    _check_asset_key_or_parent_ref = pydantic.root_validator(allow_reuse=True)(schemas.check_asset_key_or_parent_ref)
+    _check_asset_key_or_parent_ref = pydantic.model_validator(mode="before")(schemas.check_asset_key_or_parent_ref)
 
 
 class ComputeTaskOutput(schemas._PydanticConfig):
@@ -262,7 +262,7 @@ class Task(_Model):
     function: Function
     owner: str
     compute_plan_key: str
-    metadata: Dict[str, str]
+    metadata: Union[Dict[str, str], Dict[str, int]]
     status: Status
     worker: str
     rank: Optional[int] = None
@@ -361,7 +361,8 @@ class Organization(schemas._PydanticConfig):
     type_: ClassVar[str] = schemas.Type.Organization
 
 
-class OrganizationInfoConfig(schemas._PydanticConfig):
+class OrganizationInfoConfig(schemas._PydanticConfig, extra="allow"):
+    model_config = ConfigDict(protected_namespaces=())
     model_export_enabled: bool
 
 

substra/sdk/schemas.py

@@ -7,6 +7,7 @@
 from typing import Dict
 from typing import List
 from typing import Optional
+from typing import Union
 
 import pydantic
 from pydantic.fields import Field
@@ -76,6 +77,7 @@ def __str__(self):
 
 class _PydanticConfig(pydantic.BaseModel):
     """Shared configuration for all schemas here"""
+
     model_config = ConfigDict(extra="ignore")
 
 
@@ -143,7 +145,7 @@ class DataSampleSpec(_Spec):
     def is_many(self):
         return self.paths and len(self.paths) > 0
 
-    @pydantic.root_validator(pre=True)
+    @pydantic.model_validator(mode="before")
     def exclusive_paths(cls, values):
         """Check that one and only one path(s) field is defined."""
         if "paths" in values and "path" in values:
@@ -167,8 +169,8 @@ def build_request_kwargs(self, local):
 def check_asset_key_or_parent_ref(cls, values):
     """Check that either (asset key) or (parent_task_key, parent_task_output_identifier) are set, but not both."""
 
-    has_asset_key = bool(values.get("asset_key"))
-    has_parent = bool(values.get("parent_task_key")) and bool(values.get("parent_task_output_identifier"))
+    has_asset_key = bool(dict(values).get("asset_key"))
+    has_parent = bool(dict(values).get("parent_task_key")) and bool(dict(values).get("parent_task_output_identifier"))
 
     if has_asset_key != has_parent:  # xor
         return values
@@ -185,7 +187,7 @@ class InputRef(_PydanticConfig):
     parent_task_output_identifier: Optional[str] = None
 
     # either (asset_key) or (parent_task_key, parent_task_output_identifier) must be specified
-    _check_asset_key_or_parent_ref = pydantic.root_validator(allow_reuse=True)(check_asset_key_or_parent_ref)
+    _check_asset_key_or_parent_ref = pydantic.model_validator(mode="before")(check_asset_key_or_parent_ref)
 
 
 class ComputeTaskOutputSpec(_PydanticConfig):
@@ -206,7 +208,7 @@ class ComputePlanTaskSpec(_Spec):
     function_key: str
     worker: str
     tag: Optional[str] = None
-    metadata: Optional[Dict[str, str]] = None
+    metadata: Optional[Union[Dict[str, str], Dict[str, int]]] = None
     inputs: Optional[List[InputRef]] = None
     outputs: Optional[Dict[str, ComputeTaskOutputSpec]] = None
 
@@ -289,7 +291,7 @@ class FunctionInputSpec(_Spec):
     optional: bool
     kind: AssetKind
 
-    @pydantic.root_validator
+    @pydantic.model_validator(mode="before")
     def _check_identifiers(cls, values):
         """Checks that the multiplicity and the optionality of a data manager is always set to False"""
         if values["kind"] == AssetKind.data_manager:
@@ -323,7 +325,7 @@ class FunctionOutputSpec(_Spec):
     kind: AssetKind
     multiple: bool
 
-    @pydantic.root_validator
+    @pydantic.model_validator(mode="before")
     def _check_performance(cls, values):
         """Checks that the performance is always set to False"""
         if values.get("kind") == AssetKind.performance and values.get("multiple"):
@@ -373,10 +375,12 @@ def build_request_kwargs(self):
         # Computed fields using `@property` are not dumped when `exclude_unset` flag is enabled,
         # this is why we need to reimplement this custom function.
         data["inputs"] = (
-            {input.identifier: input.dict(exclude={"identifier"}) for input in self.inputs} if self.inputs else {}
+            {input.identifier: input.dict(exclude={"identifier"}) for input in self.inputs} if self.inputs else dict()
         )
         data["outputs"] = (
-            {output.identifier: output.dict(exclude={"identifier"}) for output in self.outputs} if self.outputs else {}
+            {output.identifier: output.dict(exclude={"identifier"}) for output in self.outputs}
+            if self.outputs
+            else dict()
         )
 
         if self.Meta.file_attributes:
@@ -404,7 +408,7 @@ class TaskSpec(_Spec):
     key: str = pydantic.Field(default_factory=lambda: str(uuid.uuid4()))
     tag: Optional[str] = None
     compute_plan_key: Optional[str] = None
-    metadata: Optional[Dict[str, str]] = None
+    metadata: Optional[Union[Dict[str, str], Dict[str, int]]] = None
     function_key: str
     worker: str
     rank: Optional[int] = None
@@ -419,7 +423,7 @@ def build_request_kwargs(self):
         # this is why we need to reimplement this custom function.
         data = json.loads(self.json(exclude_unset=True))
         data["key"] = self.key
-        data["inputs"] = [input.dict() for input in self.inputs] if self.inputs else []
+        data["inputs"] = [input.model_dump() for input in self.inputs] if self.inputs else []
         data["outputs"] = {k: v.dict(by_alias=True) for k, v in self.outputs.items()} if self.outputs else {}
         yield data, None