Update for 2.1 fixes

xopt/generators/bayesian/bax_generator.py

@@ -14,7 +14,7 @@
 
 
 class BaxGenerator(BayesianGenerator):
-    alias = "BAX"
+    name = "BAX"
     algorithm: Algorithm = Field(description="algorithm evaluated in the BAX process")
     algorithm_results: Dict = Field(
         None, description="dictionary results from algorithm", exclude=True

xopt/generators/bayesian/bayesian_generator.py

@@ -66,12 +66,13 @@ class BayesianGenerator(Generator, ABC):
         None,
         description="limits for travel distance between points in normalized space",
     )
-    fixed_features: Dict[str, float] = Field(
+    fixed_features: Optional[Dict[str, float]] = Field(
         None, description="fixed features used in Bayesian optimization"
     )
 
     @field_validator("model_constructor", mode='before')
     def validate_model_constructor(cls, value):
+        print(f'Verifying model {value}')
         constructor_dict = {"standard": StandardModelConstructor}
         if value is None:
             value = StandardModelConstructor()
@@ -83,9 +84,7 @@ def validate_model_constructor(cls, value):
             else:
                 raise ValueError(f"{value} not found")
         elif isinstance(value, dict):
-            #name = value.pop("name")
-            # name is ClassVar, not instance field
-            name = cls.name
+            name = value.pop("name")
             if name in constructor_dict:
                 value = constructor_dict[name](**value)
             else:
@@ -106,7 +105,7 @@ def validate_numerical_optimizer(cls, value):
             else:
                 raise ValueError(f"{value} not found")
         elif isinstance(value, dict):
-            name = cls.name
+            name = value.pop("name")
             if name in optimizer_dict:
                 value = optimizer_dict[name](**value)
             else:

xopt/generators/bayesian/models/time_dependent.py

@@ -1,5 +1,5 @@
 from copy import deepcopy
-from typing import Union
+from typing import Dict, List, Union
 
 import pandas as pd
 

xopt/numerical_optimizer.py

@@ -23,7 +23,6 @@ def optimize(self, function, bounds, n_candidates=1):
 
 class LBFGSOptimizer(NumericalOptimizer):
     name: str = Field("LBFGS", frozen=True)
-    name2: Literal['test'] = 'test2'
     n_raw_samples: PositiveInt = Field(
         20,
         description="number of raw samples used to seed optimization",
@@ -71,7 +70,6 @@ class GridOptimizer(NumericalOptimizer):
     """
 
     name: str = Field("grid", frozen=True)
-    name3: Literal['test'] = "test"
     n_grid_points: PositiveInt = Field(
         10, description="number of grid points per axis used for optimization"
     )

xopt/pydantic.py

@@ -50,7 +50,7 @@
 def recursive_serialize(v, base_key=""):
     # This will iterate model fields
     for key, value in dict(v).items():
-        # print(f'{v=} {key=} {value=}')
+        print(f'{v=} {key=} {value=}')
         if isinstance(value, dict):
             v[key] = recursive_serialize(value, base_key=key)
         elif isinstance(value, torch.nn.Module):
@@ -99,10 +99,20 @@ def orjson_dumps(v: BaseModel, *, base_key=""):
     json_encoder = partial(custom_pydantic_encoder, JSON_ENCODERS)
     return orjson_dumps_custom(v, default=json_encoder, base_key=base_key)
 
+def orjson_dumps_except_root(v: BaseModel, *, base_key=""):
+    # to ensure pydantic custom serializer works, needs to return dict and not string
+    # so, serialize subfields instead of base model
+    json_encoder = partial(custom_pydantic_encoder, JSON_ENCODERS)
+    v2 = recursive_serialize(v.model_dump(), base_key=base_key)
+    data = {}
+    for field, fv in dict(v2).items():
+        data[field] = orjson.dumps(fv, default=json_encoder).decode()
+    return data
 
 def orjson_dumps_custom(v: BaseModel, *, default, base_key=""):
     v = recursive_serialize(v.model_dump(), base_key=base_key)
-    return orjson.dumps(v, default=default).decode()
+    dump = orjson.dumps(v, default=default).decode()
+    return dump
 
 
 def orjson_loads(v, default=None):
@@ -139,9 +149,9 @@ def validate_files(cls, value, info: FieldValidationInfo):
 
         return value
 
-    @model_serializer(mode='plain', when_used='json', return_type='str')
+    @model_serializer(mode='plain', when_used='json')
     def serialize_json(self, base_key=""):
-        return orjson_dumps(self, base_key=base_key)
+        return orjson_dumps_except_root(self, base_key=base_key)
 
     # TODO: implement json load parsing on main object (json_loads is gone)