pyfemtet 0.8.8__py3-none-any.whl → 0.8.10__py3-none-any.whl

pyfemtet/__init__.py

@@ -1 +1 @@
-__version__ = "0.8.8"
+__version__ = "0.8.10"

pyfemtet/_message/locales/ja/LC_MESSAGES/messages.po

@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2025-01-29 15:47+0900\n"
+"POT-Creation-Date: 2025-02-19 13:47+0900\n"
 "PO-Revision-Date: 2024-07-22 14:05+0900\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: ja\n"
@@ -16,7 +16,7 @@ msgstr ""
 "MIME-Version: 1.0\n"
 "Content-Type: text/plain; charset=utf-8\n"
 "Content-Transfer-Encoding: 8bit\n"
-"Generated-By: Babel 2.16.0\n"
+"Generated-By: Babel 2.17.0\n"
 
 #: pyfemtet/_message/messages.py:30
 msgid "hello!"
@@ -542,6 +542,14 @@ msgstr "ひとつの設計変数に対する目的関数の変化"
 msgid "The vertical axis is objective, and the horizontal axis is parameter."
 msgstr "縦軸は目的関数、横軸は設計変数です。"
 
+#: pyfemtet/_message/messages.py:201
+msgid "The importance of parameters evaluated by fANOVA"
+msgstr "fANOVA に基づくパラメータ重要度"
+
+#: pyfemtet/_message/messages.py:202
+msgid "The normalized relative importance of parameters. Please note that the importance is calculated from the overall relationship of the input-output response, rather than from a specific solution."
+msgstr "相対的なパラメータの重要度です。この重要度は特定の解におけるものではなく、入力-出力応答全体の関係から計算されていることに注意してください。"
+
 #~ msgid "The magnitude relationship is incorrect. "
 #~ msgstr ""
 

pyfemtet/_message/locales/messages.pot

@@ -8,14 +8,14 @@ msgid ""
 msgstr ""
 "Project-Id-Version: PROJECT VERSION\n"
 "Report-Msgid-Bugs-To: EMAIL@ADDRESS\n"
-"POT-Creation-Date: 2025-01-29 15:47+0900\n"
+"POT-Creation-Date: 2025-02-19 13:47+0900\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language-Team: LANGUAGE <LL@li.org>\n"
 "MIME-Version: 1.0\n"
 "Content-Type: text/plain; charset=utf-8\n"
 "Content-Transfer-Encoding: 8bit\n"
-"Generated-By: Babel 2.16.0\n"
+"Generated-By: Babel 2.17.0\n"
 
 #: pyfemtet/_message/messages.py:30
 msgid "hello!"
@@ -541,3 +541,11 @@ msgstr ""
 msgid "The vertical axis is objective, and the horizontal axis is parameter."
 msgstr ""
 
+#: pyfemtet/_message/messages.py:201
+msgid "The importance of parameters evaluated by fANOVA"
+msgstr ""
+
+#: pyfemtet/_message/messages.py:202
+msgid "The normalized relative importance of parameters. Please note that the importance is calculated from the overall relationship of the input-output response, rather than from a specific solution."
+msgstr ""
+

pyfemtet/_message/messages.py

@@ -198,4 +198,5 @@ class Message:
     DETAIL_PAGE_CONTOUR_DESCRIPTION = _('The axes are parameters, and the color shows objective value.')
     DETAIL_PAGE_SLICE_HEADER = _('The response of an objective versus one parameter')
     DETAIL_PAGE_SLICE_DESCRIPTION = _('The vertical axis is objective, and the horizontal axis is parameter.')
-
+    DETAIL_PAGE_IMPORTANCE_HEADER = _('The importance of parameters evaluated by fANOVA')
+    DETAIL_PAGE_IMPORTANCE_DESCRIPTION = _('The normalized relative importance of parameters. Please note that the importance is calculated from the overall relationship of the input-output response, rather than from a specific solution.')

pyfemtet/_util/excel_parse_util.py

@@ -13,15 +13,15 @@ __all__ = [
 ]
 
 
-def parse_excel(book_path, sheet_name, keyword, required, optional) -> pd.DataFrame:
+def parse_excel(book_path, sheet_name, keyword, required, optional, raise_if_no_keyword=True) -> pd.DataFrame:
     """Excel シートからパラメータを取得します。
 
     シートのパースプロセスは以下の通りです。
 
     1. シート全体 (A1 セルから、値が入力されている最終セルまで) をデータに取り込みます。
-    2. すべてのセルが空白である列をデータから除きます。
-    3. すべてのセルが空白である行をデータから除きます。
-    4. 最も左上（上が優先）にある keyword に一致するセルより上および左の行・列をデータから除きます。
+    2. 最も左上（上が優先）にある keyword に一致するセルより上および左の行・列をデータから除きます。
+    3. その時点のデータの上から順に見てすべてのセルが空白である行以下の行をデータから除きます。
+    4. すべてのセルが空白である列をデータから除きます。
 
     Args:
         book_path: Excel book のパス。
@@ -29,6 +29,7 @@ def parse_excel(book_path, sheet_name, keyword, required, optional) -> pd.DataFr
         keyword (str): 必ず含まれるべき、表データの最初の列名として使う文字列。
         required (list[str]): 必ず含まれるべき、表データの列名として使う文字列のリスト。
         optional (list[str]): 表データの列名として使ってよい文字列のリスト。
+        raise_if_no_keyword (bool, optional): 指定されたシートに keyword が含まれない場合エラーにするかどうか。
 
     Returns:
 
@@ -41,21 +42,31 @@ def parse_excel(book_path, sheet_name, keyword, required, optional) -> pd.DataFr
     # 読み込み
     df = pd.read_excel(book_path, sheet_name, header=None)
 
-    # NaN のみからなる列を削除する
-    valid_columns = [col for col in df.columns if df[col].notna().sum()]
-    df = df[valid_columns]
-
-    # NaN のみからなる行を削除する
-    valid_rows = [row for row in df.index if df.loc[row].notna().sum()]
-    df = df.loc[valid_rows]
-
     # 「変数名」を左上とする表にする
-    df: pd.DataFrame
     idx = np.where(df.values == keyword)
+    if len(idx[0]) == 0:
+        if raise_if_no_keyword:
+            raise RuntimeError(f'keyword "{keyword}" is lacked in {sheet_name}. ')
+        else:
+            return pd.DataFrame()
+
     r = idx[0][0]
     c = idx[1][0]
     df = pd.DataFrame(df.iloc[1+r:, c:].values, columns=df.iloc[r, c:].values)
 
+    # NaN のみからなる行を最初に見つけるまでデータに追加する
+    valid_rows = []
+    for row in df.index:
+        if df.loc[row].notna().sum():
+            valid_rows.append(row)
+        else:
+            break
+    df = df.loc[valid_rows]
+
+    # NaN のみからなる列を削除する
+    valid_columns = [col for i, col in enumerate(df.columns) if df.iloc[:, i].notna().sum()]
+    df = df[valid_columns]
+
     # パースが成功しているかチェックする
     lack = True
     for col in df.columns:
@@ -76,8 +87,15 @@ class ParseBase:
     OPTIONAL_COLUMNS = []
 
     @classmethod
-    def parse(cls, book_path, sheet_name) -> pd.DataFrame:
-        return parse_excel(book_path, sheet_name, cls.KEYWORD, cls.REQUIRED_COLUMNS, cls.OPTIONAL_COLUMNS)
+    def parse(cls, book_path, sheet_name, raise_if_no_keyword=True) -> pd.DataFrame:
+        return parse_excel(
+            book_path,
+            sheet_name,
+            cls.KEYWORD,
+            cls.REQUIRED_COLUMNS,
+            cls.OPTIONAL_COLUMNS,
+            raise_if_no_keyword,
+        )
 
 
 class ParseAsParameter(ParseBase):

pyfemtet/opt/_femopt.py

@@ -131,7 +131,7 @@ class FEMOpt:
         # メンバーの宣言
         self.client = None
         self.status = None  # actor
-        self.history = None  # actor
+        self.history: History = None
         self.worker_status_list = None  # [actor]
         self.monitor_process_future = None
         self.monitor_server_kwargs = dict()
@@ -139,6 +139,7 @@ class FEMOpt:
         self.monitor_host_record = None
         self._hv_reference = None
         self._extra_space_dir = None
+        self._opt_exceptions = []
 
     # multiprocess 時に pickle できないオブジェクト参照の削除
     def __getstate__(self):
@@ -925,6 +926,7 @@ class FEMOpt:
             sleep(1)  # monitor が terminated 状態で少なくとも一度更新されなければ running のまま固まる
 
             # 全ての Exception を再表示
+            self._opt_exceptions = opt_exceptions
             for i, opt_exception in enumerate(opt_exceptions):
                 if opt_exception is not None:
                     print()

pyfemtet/opt/_femopt_core.py

@@ -325,7 +325,7 @@ class Function:
         args = self.args
         # Femtet 特有の処理
         if isinstance(fem, FemtetInterface):
-            args = (fem.Femtet, *args)
+            args = (fem.object_passed_to_functions, *args)
         return float(self.fun(*args, **self.kwargs))
 
     def _restore_constants(self):
@@ -553,7 +553,7 @@ class History:
             hv_reference=None,
     ):
         # hypervolume 計算メソッド
-        self._hv_reference = hv_reference or 'dynamic-pareto'
+        self._hv_reference = 'dynamic-pareto' if hv_reference is None else hv_reference
 
         # 引数の処理
         self.path = history_path  # .csv
@@ -728,6 +728,14 @@ class History:
 
         return columns, meta_columns
 
+    def generate_hidden_infeasible_result(self):
+        y = np.full_like(np.zeros(len(self.obj_names)), np.nan)
+        c = np.full_like(np.zeros(len(self.cns_names)), np.nan)
+        return y, c
+
+    def is_hidden_infeasible_result(self, y):
+        return np.all(np.isnan(y))
+
     def record(
             self,
             parameters,
@@ -780,7 +788,7 @@ class History:
             feasible_list.append(is_feasible(cns_value, cns.lb, cns.ub))
 
         # feasibility
-        row.append(all(feasible_list))
+        row.append(all(feasible_list) and not self.is_hidden_infeasible_result(obj_values))
 
         # the others
         row.append(-1.)  # dummy hypervolume
@@ -892,7 +900,29 @@ class History:
                     ret.append(row)
             return np.array(ret).max(axis=0)
 
-        if self._hv_reference == 'dynamic-pareto':
+        if (
+                isinstance(self._hv_reference, np.ndarray)
+                or isinstance(self._hv_reference, list)
+        ):
+            _buff = np.array(self._hv_reference)
+            assert _buff.shape == (len(self.obj_names),)
+
+            ref_point = np.array(
+                [obj.convert(raw_value) for obj, raw_value in zip(objectives.values(), _buff)]
+            )
+
+            _buff = get_pareto(objective_values)
+
+            pareto_set = np.empty((0, len(objectives)))
+            for pareto_sol in _buff:
+                if all(pareto_sol < ref_point):
+                    pareto_set = np.concatenate([pareto_set, [pareto_sol]], axis=0)
+
+            hv = compute_hypervolume(pareto_set, ref_point)
+            df.loc[len(df) - 1, 'hypervolume'] = hv
+            return
+
+        elif self._hv_reference == 'dynamic-pareto':
             pareto_set, pareto_set_list = get_pareto(objective_values, with_partial=True)
             for i, partial_pareto_set in enumerate(pareto_set_list):
                 # 並列計算時など Valid な解がまだ一つもない場合は pareto_set が長さ 0 になる
@@ -943,28 +973,6 @@ class History:
                 df.loc[len(df) - 1, 'hypervolume'] = hv
             return
 
-        elif (
-                isinstance(self._hv_reference, np.ndarray)
-                or isinstance(self._hv_reference, list)
-        ):
-            _buff = np.array(self._hv_reference)
-            assert _buff.shape == (len(self.obj_names),)
-
-            ref_point = np.array(
-                [obj.convert(raw_value) for obj, raw_value in zip(objectives.values(), _buff)]
-            )
-
-            _buff = get_pareto(objective_values)
-
-            pareto_set = np.empty((0, len(objectives)))
-            for pareto_sol in _buff:
-                if all(pareto_sol < ref_point):
-                    pareto_set = np.concatenate([pareto_set, [pareto_sol]], axis=0)
-
-            hv = compute_hypervolume(pareto_set, ref_point)
-            df.loc[len(df) - 1, 'hypervolume'] = hv
-            return
-
         else:
             raise NotImplementedError(f'Invalid Hypervolume reference point calculation method: {self._hv_reference}')
 

pyfemtet/opt/_test_utils/record_history.py

@@ -92,6 +92,7 @@ def is_equal_result(ref_path, dif_path, log_path=None, threashold=0.05):
     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 <= 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。"
+    print(f"OK! 前回の結果との平均差異が {int(difference)}% で {int(threashold*100)}% 以下です。")
 
 
 def _get_simplified_df_values(csv_path, exclude_columns=None):

pyfemtet/opt/advanced_samples/meta_script/meta_script.py

@@ -0,0 +1,163 @@
+# for meta script
+import os
+import pyfemtet
+
+if __name__ == '__main__':
+    print(f'pyfemtet {pyfemtet.__version__} starting.')
+
+from pyfemtet._message.messages import encoding
+
+from fire import Fire
+import yaml
+
+# for concrete script
+from pyfemtet.opt import FEMOpt
+from pyfemtet.opt.interface import *
+from pyfemtet.opt.optimizer import *
+from optuna.samplers import *
+from pyfemtet.opt.interface._femtet_excel import FemtetWithExcelSettingsInterface
+
+
+# fore debug
+class ContentContext:
+
+    def __init__(self, content_):
+        self.content = content_
+
+    def __enter__(self):
+        return self.content
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        pass
+
+
+def main(
+        yaml_path: str = None,
+
+        interface_class: str = None,  # including parameter definition excel
+        interface_kwargs: str = None,  # including Parametric Analysis Output
+        optimizer_class: str = None,
+        optimizer_kwargs: str = None,
+        femopt_kwargs: str = None,
+        seed: str = 'null',
+        optimize_kwargs: str = None,
+):
+    """
+
+    Args:
+        yaml_path:
+            If this argument is passed, the other arguments will be ignored
+            and load by .yaml file.
+            The yaml file must contain the other arguments.
+
+        interface_class: FemtetWithExcelSettingsInterface or SurrogateModelInterface.
+        interface_kwargs: See documentation of each interface class.
+        optimizer_class: OptunaOptimizer or ScipyOptimizer.
+        optimizer_kwargs: See documentation of each optimizer class.
+        femopt_kwargs: See documentation of FEMOpt.
+        seed: int or None.
+        optimize_kwargs: See documentation of FEMOpt.optimize().
+
+    """
+
+    if yaml_path is not None:
+        if os.path.isfile(yaml_path):
+            context = open(yaml_path, 'r', encoding='utf-8')
+        else:
+            print('debug mode')
+            context = ContentContext(yaml_path)
+        with context as f:
+            d = yaml.safe_load(f)
+            interface_class = yaml.safe_dump(d['interface_class'], allow_unicode=True)
+            interface_kwargs = yaml.safe_dump(d['interface_kwargs'], allow_unicode=True)
+            optimizer_class = yaml.safe_dump(d['optimizer_class'], allow_unicode=True)
+            optimizer_kwargs = yaml.safe_dump(d['optimizer_kwargs'], allow_unicode=True)
+            femopt_kwargs = yaml.safe_dump(d['femopt_kwargs'], allow_unicode=True)
+            seed = yaml.safe_dump(d['seed'], allow_unicode=True)
+            optimize_kwargs = yaml.safe_dump(d['optimize_kwargs'], allow_unicode=True)
+
+    Interface = eval(yaml.safe_load(interface_class))
+    interface_kwargs_ = yaml.safe_load(interface_kwargs)
+
+    Optimizer = eval(yaml.safe_load(optimizer_class))
+    optimizer_kwargs_ = yaml.safe_load(optimizer_kwargs)
+
+    femopt_kwargs_ = yaml.safe_load(femopt_kwargs)
+
+    seed_ = yaml.safe_load(seed)
+
+    optimize_kwargs_ = yaml.safe_load(optimize_kwargs)
+
+    fem = Interface(**interface_kwargs_)
+    opt = Optimizer(**optimizer_kwargs_)
+    femopt = FEMOpt(fem=fem, opt=opt, **femopt_kwargs_)
+    femopt.set_random_seed(seed_)
+    femopt.optimize(**optimize_kwargs_)
+
+
+# if __name__ == '__main__':
+#     Fire(main)
+
+
+# for debugging file input
+if __name__ == '__main__':
+    os.chdir(os.path.dirname(__file__))
+    path = 'sample.yaml'
+    main(yaml_path=path)
+
+
+# for debugging yaml input
+# if __name__ == '__main__':
+#     content = r"""
+# interface_class: FemtetInterface
+# interface_kwargs:
+#   femprj_path: C:\日本語ファイル.femprj
+# optimizer_class: OptunaOptimizer
+# optimizer_kwargs:
+#   sampler_class: TPESampler
+#   sampler_kwargs:
+#     n_startup_trials: 10
+# femopt_kwargs:
+#   history_path: sample.csv
+# optimize_kwargs:
+#   n_trials: 15
+#   confirm_before_exit: False
+# seed: null
+# """
+#
+#     main(yaml_path=content)
+
+
+# for debugging CLI input
+# if __name__ == '__main__':
+#
+#     interface_kwargs = dict(
+#         femprj_path=r'sample.femprj',
+#         parametric_output_indexes_use_as_objective={0: 0, 1: 'minimize'}
+#     )
+#     optimizer_kwargs = dict(
+#         sampler_class='TPESampler',
+#         sampler_kwargs=dict(
+#             n_startup_trials=10,
+#         )
+#     )
+#     femopt_kwargs = dict(
+#         history_path='sample.csv'
+#     )
+#
+#     seed = None
+#
+#     optimize_kwargs = dict(
+#         n_trials=15,
+#         confirm_before_exit=False,
+#     )
+#
+#     main(
+#         interface_class='FemtetWithExcelSettingsInterface',
+#         interface_kwargs=yaml.safe_dump(interface_kwargs, allow_unicode=True),
+#         optimizer_class='OptunaOptimizer',
+#         optimizer_kwargs=yaml.safe_dump(optimizer_kwargs, allow_unicode=True),
+#         femopt_kwargs=yaml.safe_dump(femopt_kwargs, allow_unicode=True),
+#         seed=yaml.safe_dump(femopt_kwargs, allow_unicode=True),
+#         optimize_kwargs=yaml.safe_dump(optimize_kwargs, allow_unicode=True),
+#     )

pyfemtet/opt/advanced_samples/meta_script/sample.yaml

@@ -0,0 +1,14 @@
+ interface_class: FemtetInterface
+ interface_kwargs:
+   femprj_path: C:\日本語ファイル.femprj
+ optimizer_class: OptunaOptimizer
+ optimizer_kwargs:
+   sampler_class: TPESampler
+   sampler_kwargs:
+     n_startup_trials: 10
+ femopt_kwargs:
+   history_path: sample.csv
+ optimize_kwargs:
+   n_trials: 15
+   confirm_before_exit: False
+ seed: null

pyfemtet/opt/advanced_samples/restart/gal_ex13_parametric_restart.py

@@ -0,0 +1,99 @@
+import os
+from time import sleep
+
+from optuna.samplers import RandomSampler, NSGAIISampler, GPSampler, BaseSampler
+
+from pyfemtet.opt import FEMOpt, FemtetInterface, OptunaOptimizer
+
+os.chdir(os.path.dirname(__file__))
+
+
+def get_res_freq(Femtet):
+    Galileo = Femtet.Gogh.Galileo
+    Galileo.Mode = 0
+    sleep(0.01)
+    return Galileo.GetFreq().Real
+
+
+def main(n_trials, sampler_class: type[BaseSampler], sampler_kwargs: dict):
+    """Main function
+
+    length
+    using different algorithms for each restarting.
+
+    So this main function requires n_trials and sampler_class.
+
+    Args:
+
+        n_trials (int):
+            How many additional succeeded trials
+            to terminate optimization.
+
+        sampler_class (type[optuna.samplers.BaseSampler]):
+            The algorithm we use.
+
+        sampler_kwargs (dict):
+            The arguments of sampler.
+
+    """
+
+
+    # Connect to Femtet.
+    fem = FemtetInterface(
+        femprj_path='gal_ex13_parametric.femprj',
+    )
+
+    # Initialize the optimization object.
+    opt = OptunaOptimizer(
+        sampler_class=sampler_class,
+        sampler_kwargs=sampler_kwargs,
+    )
+
+    # To restart, it is necessary to inform the new optimization program
+    # about the history of the previous optimization.
+    # By specifying a csv for the `history_path` argument of FEMOpt,
+    # if it does not exist, a new csv file will be created,
+    # and if it exists, optimization will continue from that csv file.
+    #
+    # Note:
+    #   When restarting, the number and names of variables,
+    #   as well as the number and names of objective functions
+    #   and constraints must be consistent.
+    #   However, you can change the bounds of variables,
+    #   direction of objective functions, and content of constraints.
+    #
+    # Note:
+    #   When using OptunaOptimizer, the .db file with the same name
+    #   (in this case restarting-sample.db) that is saved along with
+    #   csv is required to be  in the same folder as the csv file.
+    femopt = FEMOpt(
+        fem=fem,
+        opt=opt,
+        history_path='restarting-sample.csv'
+    )
+
+    # Set the design variables.
+    femopt.add_parameter('length', 0.1, 0.02, 0.2)
+    femopt.add_parameter('width', 0.01, 0.001, 0.02)
+    femopt.add_parameter('base_radius', 0.008, 0.006, 0.01)
+
+    # Set the objective function.
+    femopt.add_objective(fun=get_res_freq, name='First Resonant Frequency (Hz)', direction=800)
+
+    # Run optimization.
+    femopt.set_random_seed(42)
+    femopt.optimize(n_trials=n_trials, confirm_before_exit=False)
+
+
+if __name__ == '__main__':
+    # First, we will perform 3 optimizations using the RandomSampler.
+    main(3, RandomSampler, {})
+
+    # Next, we will perform 3 optimizations using the NSGAIISampler.
+    main(3, NSGAIISampler, {})
+
+    # Finally, we will perform 3 optimizations using the GPSampler.
+    main(3, GPSampler, {'n_startup_trials': 0, 'deterministic_objective': True})
+
+    # After this program ends, you can continue further with
+    # restarting-sample.csv and the .db file.