PyFemtet 0.8.4__tar.gz → 0.8.5__tar.gz

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pyfemtet
-Version: 0.8.4
+Version: 0.8.5
 Summary: Design parameter optimization using Femtet.
 License: BSD-3-Clause
 Author: kazuma.naito

pyfemtet-0.8.5/pyfemtet/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.8.5"

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/_femopt.py

@@ -403,7 +403,7 @@ class FEMOpt:
 
         if names is not None:
             if isinstance(names, str):
-                names = [f'name_{i}' for i in range(n_return)]
+                names = [f'{names}_{i}' for i in range(n_return)]
             else:
                 # names = names
                 pass

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/_test_utils/record_history.py

@@ -59,32 +59,39 @@ def _get_obj_from_csv(csv_path, encoding=ENCODING):
         reader = csv.reader(f, delimiter=",")
         meta = reader.__next__()
     obj_indices = np.where(np.array(meta) == "obj")[0]
-    out = df.iloc[:, obj_indices]
+    out: pd.DataFrame = df.iloc[:, obj_indices]
+    out = out.dropna(axis=0)
     return out, columns
 
 
-def is_equal_result(ref_path, dif_path, log_path):
+def is_equal_result(ref_path, dif_path, log_path=None, threashold=0.05):
     """Check the equality of two result csv files."""
     ref_df, ref_columns = _get_obj_from_csv(ref_path)
     dif_df, dif_columns = _get_obj_from_csv(dif_path)
 
-    with open(log_path, "a", newline="\n", encoding=ENCODING) as f:
-        f.write("\n\n===== 結果の分析 =====\n\n")
-        f.write(f"        \tref\tdif\n")
-        f.write(f"---------------------\n")
-        f.write(f"len(col)\t{len(ref_columns)}\t{len(dif_columns)}\n")
-        f.write(f"len(df) \t{len(ref_df)}\t{len(dif_df)}\n")
-        try:
-            difference = (
+    if log_path is not None:
+        with open(log_path, "a", newline="\n", encoding=ENCODING) as f:
+            f.write("\n\n===== 結果の分析 =====\n\n")
+            f.write(f"        \tref\tdif\n")
+            f.write(f"---------------------\n")
+            f.write(f"len(col)\t{len(ref_columns)}\t{len(dif_columns)}\n")
+            f.write(f"len(df) \t{len(ref_df)}\t{len(dif_df)}\n")
+            try:
+                difference = (
+                    np.abs(ref_df.values - dif_df.values) / np.abs(dif_df.values)
+                ).mean()
+                f.write(f"diff    \t{int(difference*100)}%\n")
+            except Exception:
+                f.write(f"diff    \tcannot calc\n")
+
+    else:
+        difference = (
                 np.abs(ref_df.values - dif_df.values) / np.abs(dif_df.values)
-            ).mean()
-            f.write(f"diff    \t{int(difference*100)}%\n")
-        except Exception:
-            f.write(f"diff    \tcannot calc\n")
+        ).mean()
 
     assert len(ref_columns) == len(dif_columns), "結果 csv の column 数が異なります。"
     assert len(ref_df) == len(dif_df), "結果 csv の row 数が異なります。"
-    assert difference <= 0.05, "前回の結果との平均差異が 5% を超えています。"
+    assert difference <= threashold*100, f"前回の結果との平均差異が {int(difference)}% で {int(threashold*100)}% を超えています。"
 
 
 def _get_simplified_df_values(csv_path, exclude_columns=None):

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/interface/__init__.py

@@ -21,7 +21,9 @@ else:
     FemtetWithNXInterface = type('FemtetWithNXInterface', (FemtetInterface,), {})
     ExcelInterface = type('FemtetInterface', (NotAvailableForWindows,), {})
 
-from pyfemtet.opt.interface._surrogate import PoFBoTorchInterface
+from pyfemtet.opt.interface._surrogate._base import SurrogateModelInterfaceBase
+from pyfemtet.opt.interface._surrogate._singletaskgp import PoFBoTorchInterface
+
 
 __all__ =[
     'FEMInterface',
@@ -30,5 +32,6 @@ __all__ =[
     'FemtetWithSolidworksInterface',
     'FemtetWithNXInterface',
     'ExcelInterface',
+    'SurrogateModelInterfaceBase',
     'PoFBoTorchInterface',
 ]

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/interface/_surrogate/_base.py

@@ -17,6 +17,7 @@ class SurrogateModelInterfaceBase(FEMInterface, ABC):
             self,
             history_path: str = None,
             history: History = None,
+            override_objective: bool = True,
     ):
 
         self.history: History
@@ -25,6 +26,7 @@ class SurrogateModelInterfaceBase(FEMInterface, ABC):
         self.obj: dict[str, float] = dict()
         self.df_prm: pd.DataFrame
         self.df_obj: pd.DataFrame
+        self.override_objective: bool = override_objective
 
         # history_path が与えられた場合、history をコンストラクトする
         if history_path is not None:
@@ -57,27 +59,25 @@ class SurrogateModelInterfaceBase(FEMInterface, ABC):
         FEMInterface.__init__(
             self,
             history=history,  # コンストラクト済み history を渡せば並列計算時も何もしなくてよい
+            override_objective=self.override_objective
         )
 
     def filter_feasible(self, x: np.ndarray, y: np.ndarray, return_feasibility=False):
         feasible_idx = np.where(~np.isnan(y.sum(axis=1)))
         if return_feasibility:
             # calculated or not
-            y = np.zeros_like(y)
-            y[feasible_idx] = 1.
-            # satisfy weak feasibility or not
-            infeasible_idx = np.where(~self.history.get_df()['feasible'].values)
-            y[infeasible_idx] = .0
-            return x, y.reshape((-1, 1))
+            feas = np.zeros((len(y), 1), dtype=float)
+            feas[feasible_idx] = 1.
+            return x, feas
         else:
             return x[feasible_idx], y[feasible_idx]
 
     def _setup_after_parallel(self, *args, **kwargs):
-
-        opt: AbstractOptimizer = kwargs['opt']
-        obj: Objective
-        for obj_name, obj in opt.objectives.items():
-            obj.fun = lambda: self.obj[obj_name]
+        if self.override_objective:
+            opt: AbstractOptimizer = kwargs['opt']
+            obj: Objective
+            for obj_name, obj in opt.objectives.items():
+                obj.fun = lambda obj_name_=obj_name: self.obj[obj_name_]
 
     def update_parameter(self, parameters: pd.DataFrame, with_warning=False) -> Optional[List[str]]:
         for i, row in parameters.iterrows():

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/interface/_surrogate/_singletaskgp.py

@@ -27,8 +27,8 @@ class PoFBoTorchInterface(SurrogateModelInterfaceBase):
     def train_f(self):
         # df そのまま用いて training する
         x, y = self.filter_feasible(self.df_prm.values, self.df_obj.values, return_feasibility=True)
-        if y.min() == 1:
-            self.model_f.predict = lambda *args, **kwargs: (1., 0.001)
+        if y.min() == 1:  # feasible values only
+            self.model_f.predict = lambda *args, **kwargs: (1., 0.001)  # mean, std
         self.model_f.fit(x, y)
 
     def _setup_after_parallel(self, *args, **kwargs):
@@ -64,7 +64,8 @@ class PoFBoTorchInterface(SurrogateModelInterfaceBase):
             raise SolveError(Msg.INFO_POF_IS_LESS_THAN_THRESHOLD)
 
         # 実際の計算(mean は history.obj_names 順)
-        mean, _ = self.model.predict(np.array([x]))
+        _mean, _std = self.model.predict(np.array([x]))
+        mean = _mean[0]
 
         # 目的関数の更新
         self.obj = {obj_name: value for obj_name, value in zip(self.history.obj_names, mean)}

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyfemtet/opt/prediction/single_task_gp.py

@@ -85,6 +85,8 @@ class SingleTaskGPModel(PredictionModelBase):
         fit_gpytorch_mll(mll)
 
     def predict(self, x: np.ndarray) -> tuple[np.ndarray, np.ndarray]:
+        assert len(x.shape) >= 2
+
         X = tensor(x)
 
         post = self.gp.posterior(X)

{pyfemtet-0.8.4 → pyfemtet-0.8.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pyfemtet"
-version = "0.8.4"  # ignored by versioning plugin
+version = "0.8.5"  # ignored by versioning plugin
 description = "Design parameter optimization using Femtet."
 authors = ["kazuma.naito <kazuma.naito@murata.com>"]
 readme = "README.md"

pyfemtet-0.8.4/pyfemtet/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.8.4"