feat: add get_model classmethod to BaseModel (deepmodeling#4002)

Fix deepmodeling#3968. External and new models can implement this method (if
different from default) without changing the old `get_model` methods
(which cannot be done by a plugin).

Note: I don't modify old `get_model` methods in this PR.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Introduced a new method for model instantiation that enhances
flexibility in parameter configuration.
- Improved the model retrieval process to support dynamic model
selection based on specified types.
  
- **Bug Fixes**
- Enhanced control flow to ensure correct model type selection,
addressing potential issues with model retrieval.

- **Refactor**
- Updated existing model retrieval functions to streamline logic and
improve maintainability.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>

deepmd/dpmodel/model/base_model.py

@@ -1,5 +1,6 @@
 # SPDX-License-Identifier: LGPL-3.0-or-later
 import inspect
+import json
 from abc import (
     ABC,
     abstractmethod,
@@ -193,6 +194,30 @@ def update_sel(
             cls = cls.get_class_by_type(model_type)
             return cls.update_sel(train_data, type_map, local_jdata)
 
+        @classmethod
+        def get_model(cls, model_params: dict) -> "BaseBaseModel":
+            """Get the model by the parameters.
+
+            By default, all the parameters are directly passed to the constructor.
+            If not, override this method.
+
+            Parameters
+            ----------
+            model_params : dict
+                The model parameters
+
+            Returns
+            -------
+            BaseBaseModel
+                The model
+            """
+            model_params_old = model_params.copy()
+            model_params = model_params.copy()
+            model_params.pop("type", None)
+            model = cls(**model_params)
+            model.model_def_script = json.dumps(model_params_old)
+            return model
+
     return BaseBaseModel
 
 

deepmd/dpmodel/model/model.py

@@ -5,6 +5,9 @@
 from deepmd.dpmodel.fitting.ener_fitting import (
     EnergyFittingNet,
 )
+from deepmd.dpmodel.model.base_model import (
+    BaseModel,
+)
 from deepmd.dpmodel.model.ener_model import (
     EnergyModel,
 )
@@ -93,7 +96,11 @@ def get_model(data: dict):
     data : dict
         The data to construct the model.
     """
-    if "spin" in data:
-        return get_spin_model(data)
+    model_type = data.get("type", "standard")
+    if model_type == "standard":
+        if "spin" in data:
+            return get_spin_model(data)
+        else:
+            return get_standard_model(data)
     else:
-        return get_standard_model(data)
+        return BaseModel.get_class_by_type(model_type).get_model(data)

deepmd/pt/model/model/__init__.py

@@ -197,12 +197,16 @@ def get_standard_model(model_params):
 
 
 def get_model(model_params):
-    if "spin" in model_params:
-        return get_spin_model(model_params)
-    elif "use_srtab" in model_params:
-        return get_zbl_model(model_params)
+    model_type = model_params.get("type", "standard")
+    if model_type == "standard":
+        if "spin" in model_params:
+            return get_spin_model(model_params)
+        elif "use_srtab" in model_params:
+            return get_zbl_model(model_params)
+        else:
+            return get_standard_model(model_params)
     else:
-        return get_standard_model(model_params)
+        return BaseModel.get_class_by_type(model_type).get_model(model_params)
 
 
 __all__ = [