v2: More validation (#381)

* Check priors (related to #374)
* Check observables
* Fix missing prior parameters after v1->v2 conversion of uniform priors
* Fix some pydantic validation / serialization 
* Fix style

---------

Co-authored-by: Dilan Pathirana <[email protected]>

Author: dweindl

petab/v2/core.py

@@ -73,8 +73,11 @@ def _not_nan(v: float, info: ValidationInfo) -> float:
 
 
 def _convert_nan_to_none(v):
+    """Convert NaN or "" to None."""
     if isinstance(v, float) and np.isnan(v):
         return None
+    if isinstance(v, str) and v == "":
+        return None
     return v
 
 
@@ -503,9 +506,17 @@ class ExperimentPeriod(BaseModel):
     @field_validator("condition_ids", mode="before")
     @classmethod
     def _validate_ids(cls, condition_ids):
+        if condition_ids in [None, "", [], [""]]:
+            # unspecified, or "use-model-as-is"
+            return []
+
         for condition_id in condition_ids:
+            # The empty condition ID for "use-model-as-is" has been handled
+            #  above. Having a combination of empty and non-empty IDs is an
+            #  error, since the targets of conditions to be combined must be
+            #  disjoint.
             if not is_valid_identifier(condition_id):
-                raise ValueError(f"Invalid ID: {condition_id}")
+                raise ValueError(f"Invalid {C.CONDITION_ID}: `{condition_id}'")
         return condition_ids
 
 
@@ -854,17 +865,23 @@ class Parameter(BaseModel):
     #: Parameter ID.
     id: str = Field(alias=C.PARAMETER_ID)
     #: Lower bound.
-    lb: float | None = Field(alias=C.LOWER_BOUND, default=None)
+    lb: Annotated[float | None, BeforeValidator(_convert_nan_to_none)] = Field(
+        alias=C.LOWER_BOUND, default=None
+    )
     #: Upper bound.
-    ub: float | None = Field(alias=C.UPPER_BOUND, default=None)
+    ub: Annotated[float | None, BeforeValidator(_convert_nan_to_none)] = Field(
+        alias=C.UPPER_BOUND, default=None
+    )
     #: Nominal value.
-    nominal_value: float | None = Field(alias=C.NOMINAL_VALUE, default=None)
+    nominal_value: Annotated[
+        float | None, BeforeValidator(_convert_nan_to_none)
+    ] = Field(alias=C.NOMINAL_VALUE, default=None)
     #: Is the parameter to be estimated?
     estimate: bool = Field(alias=C.ESTIMATE, default=True)
     #: Type of parameter prior distribution.
-    prior_distribution: PriorDistribution | None = Field(
-        alias=C.PRIOR_DISTRIBUTION, default=None
-    )
+    prior_distribution: Annotated[
+        PriorDistribution | None, BeforeValidator(_convert_nan_to_none)
+    ] = Field(alias=C.PRIOR_DISTRIBUTION, default=None)
     #: Prior distribution parameters.
     prior_parameters: list[float] = Field(
         alias=C.PRIOR_PARAMETERS, default_factory=list
@@ -889,8 +906,18 @@ def _validate_id(cls, v):
 
     @field_validator("prior_parameters", mode="before")
     @classmethod
-    def _validate_prior_parameters(cls, v):
+    def _validate_prior_parameters(
+        cls, v: str | list[str] | float | None | np.ndarray
+    ):
+        if v is None:
+            return []
+
+        if isinstance(v, float) and np.isnan(v):
+            return []
+
         if isinstance(v, str):
+            if v == "":
+                return []
             v = v.split(C.PARAMETER_SEPARATOR)
         elif not isinstance(v, Sequence):
             v = [v]
@@ -899,7 +926,7 @@ def _validate_prior_parameters(cls, v):
 
     @field_validator("estimate", mode="before")
     @classmethod
-    def _validate_estimate_before(cls, v):
+    def _validate_estimate_before(cls, v: bool | str):
         if isinstance(v, bool):
             return v
 
@@ -918,12 +945,17 @@ def _validate_estimate_before(cls, v):
     def _serialize_estimate(self, estimate: bool, _info):
         return str(estimate).lower()
 
-    @field_validator("lb", "ub", "nominal_value")
-    @classmethod
-    def _convert_nan_to_none(cls, v):
-        if isinstance(v, float) and np.isnan(v):
-            return None
-        return v
+    @field_serializer("prior_distribution")
+    def _serialize_prior_distribution(
+        self, prior_distribution: PriorDistribution | None, _info
+    ):
+        if prior_distribution is None:
+            return ""
+        return str(prior_distribution)
+
+    @field_serializer("prior_parameters")
+    def _serialize_prior_parameters(self, prior_parameters: list[str], _info):
+        return C.PARAMETER_SEPARATOR.join(prior_parameters)
 
     @model_validator(mode="after")
     def _validate(self) -> Self:
@@ -952,7 +984,7 @@ def _validate(self) -> Self:
 
     @property
     def prior_dist(self) -> Distribution:
-        """Get the pior distribution of the parameter."""
+        """Get the prior distribution of the parameter."""
         if self.estimate is False:
             raise ValueError(f"Parameter `{self.id}' is not estimated.")
 
@@ -980,6 +1012,13 @@ def prior_dist(self) -> Distribution:
                     "transformation."
                 )
             return cls(*self.prior_parameters, trunc=[self.lb, self.ub])
+
+        if cls == Uniform:
+            # `Uniform.__init__` does not accept the `trunc` parameter
+            low = max(self.prior_parameters[0], self.lb)
+            high = min(self.prior_parameters[1], self.ub)
+            return cls(low, high, log=log)
+
         return cls(*self.prior_parameters, log=log, trunc=[self.lb, self.ub])