annofabcli 1.102.1__py3-none-any.whl → 1.103.0__py3-none-any.whl

annofabcli/statistics/visualization/dataframe/user_performance.py

@@ -100,9 +100,7 @@ class UserPerformance:
         self.custom_production_volume_list = custom_production_volume_list if custom_production_volume_list is not None else []
         self.phase_list = phase_list
         if not self.required_columns_exist(df):
-            raise ValueError(
-                f"引数'df'の'columns'に次の列が存在していません。 {self.missing_columns(df)} :: 次の列が必須です。{self.columns}の列が必要です。"
-            )
+            raise ValueError(f"引数'df'の'columns'に次の列が存在していません。 {self.missing_columns(df)} :: 次の列が必須です。{self.columns}の列が必要です。")
 
         self.df = df
 
@@ -116,18 +114,14 @@ class UserPerformance:
         return len(self.df) == 0
 
     @staticmethod
-    def _add_ratio_column_for_productivity_per_user(
-        df: pandas.DataFrame, phase_list: Sequence[TaskPhaseString], production_volume_columns: list[str]
-    ) -> None:
+    def _add_ratio_column_for_productivity_per_user(df: pandas.DataFrame, phase_list: Sequence[TaskPhaseString], production_volume_columns: list[str]) -> None:
         """
         ユーザーの生産性に関する列を、DataFrameに追加します。
         """
 
         # 集計対象タスクから算出した計測作業時間（`monitored_worktime_hour`）に対応する実績作業時間を推定で算出する
         # 具体的には、実際の計測作業時間と十先作業時間の比（`real_monitored_worktime_hour/real_actual_worktime_hour`）になるように按分する
-        df[("actual_worktime_hour", "sum")] = (
-            df[("monitored_worktime_hour", "sum")] / df[("real_monitored_worktime_hour/real_actual_worktime_hour", "sum")]
-        )
+        df[("actual_worktime_hour", "sum")] = df[("monitored_worktime_hour", "sum")] / df[("real_monitored_worktime_hour/real_actual_worktime_hour", "sum")]
 
         for phase in phase_list:
 
@@ -156,23 +150,15 @@ class UserPerformance:
             # 生産性を算出
             ratio__actual_vs_monitored_worktime = df[("actual_worktime_hour", phase)] / df[("monitored_worktime_hour", phase)]
             for production_volume_column in production_volume_columns:
-                df[(f"monitored_worktime_hour/{production_volume_column}", phase)] = (
-                    df[("monitored_worktime_hour", phase)] / df[(production_volume_column, phase)]
-                )
-                df[(f"actual_worktime_hour/{production_volume_column}", phase)] = (
-                    df[("actual_worktime_hour", phase)] / df[(production_volume_column, phase)]
-                )
+                df[(f"monitored_worktime_hour/{production_volume_column}", phase)] = df[("monitored_worktime_hour", phase)] / df[(production_volume_column, phase)]
+                df[(f"actual_worktime_hour/{production_volume_column}", phase)] = df[("actual_worktime_hour", phase)] / df[(production_volume_column, phase)]
 
-                df[(f"stdev__actual_worktime_hour/{production_volume_column}", phase)] = (
-                    df[(f"stdev__monitored_worktime_hour/{production_volume_column}", phase)] * ratio__actual_vs_monitored_worktime
-                )
+                df[(f"stdev__actual_worktime_hour/{production_volume_column}", phase)] = df[(f"stdev__monitored_worktime_hour/{production_volume_column}", phase)] * ratio__actual_vs_monitored_worktime
 
         # 品質に関する情報
         phase = TaskPhase.ANNOTATION.value
         for production_volume_column in production_volume_columns:
-            df[(f"pointed_out_inspection_comment_count/{production_volume_column}", phase)] = (
-                df[("pointed_out_inspection_comment_count", phase)] / df[(production_volume_column, phase)]
-            )
+            df[(f"pointed_out_inspection_comment_count/{production_volume_column}", phase)] = df[("pointed_out_inspection_comment_count", phase)] / df[(production_volume_column, phase)]
 
         df[("rejected_count/task_count", phase)] = df[("rejected_count", phase)] / df[("task_count", phase)]
 
@@ -204,9 +190,7 @@ class UserPerformance:
         return cls(df, task_completion_criteria, custom_production_volume_list=custom_production_volume_list)
 
     @classmethod
-    def empty(
-        cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None
-    ) -> UserPerformance:
+    def empty(cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None) -> UserPerformance:
         """空のデータフレームを持つインスタンスを生成します。"""
         production_volume_columns = ["input_data_count", "annotation_count"]
         if custom_production_volume_list is not None:
@@ -321,10 +305,7 @@ class UserPerformance:
             # `to_csv()`で出力したときにKeyErrorが発生内容にするため、事前に列を追加しておく
             phase = TaskPhase.ANNOTATION.value
             columns = pandas.MultiIndex.from_tuples(
-                [
-                    (f"stdev__monitored_worktime_hour/{production_volume_column}", phase)
-                    for production_volume_column in task_worktime_by_phase_user.production_volume_columns
-                ]
+                [(f"stdev__monitored_worktime_hour/{production_volume_column}", phase) for production_volume_column in task_worktime_by_phase_user.production_volume_columns]
             )
             df_empty = pandas.DataFrame(columns=columns, index=pandas.Index([], name="account_id"), dtype="float64")
             return df_empty
@@ -340,9 +321,7 @@ class UserPerformance:
         df_stdev_per_volume_count_list = []
         for production_volume_column in task_worktime_by_phase_user.production_volume_columns:
             df_stdev_per_input_data_count = (
-                df2[df2[f"worktime_hour/{production_volume_column}"] != float("inf")]
-                .groupby(["account_id", "phase"])[[f"worktime_hour/{production_volume_column}"]]
-                .std(ddof=0)
+                df2[df2[f"worktime_hour/{production_volume_column}"] != float("inf")].groupby(["account_id", "phase"])[[f"worktime_hour/{production_volume_column}"]].std(ddof=0)
             )
             df_stdev_per_volume_count_list.append(df_stdev_per_input_data_count)
         df_stdev = pandas.concat(df_stdev_per_volume_count_list, axis=1)
@@ -351,9 +330,7 @@ class UserPerformance:
         # 前述の処理でinfを除外しているので、NaNが含まれることはないはず
         df_stdev2 = pandas.pivot_table(
             df_stdev,
-            values=[
-                f"worktime_hour/{production_volume_column}" for production_volume_column in task_worktime_by_phase_user.production_volume_columns
-            ],
+            values=[f"worktime_hour/{production_volume_column}" for production_volume_column in task_worktime_by_phase_user.production_volume_columns],
             index="account_id",
             columns="phase",
             dropna=False,
@@ -501,10 +478,7 @@ class UserPerformance:
 
         df = worktime_per_date.df
 
-        df4_list = [
-            _create_df_first_last_working_date(phase)
-            for phase in [None, TaskPhase.ANNOTATION.value, TaskPhase.INSPECTION.value, TaskPhase.ACCEPTANCE.value]
-        ]
+        df4_list = [_create_df_first_last_working_date(phase) for phase in [None, TaskPhase.ANNOTATION.value, TaskPhase.INSPECTION.value, TaskPhase.ACCEPTANCE.value]]
 
         # joinしない理由: レベル1の列名が空文字のDataFrameをjoinすると、Python3.12のpandas2.2.0で、列名が期待通りにならないため
         # https://github.com/pandas-dev/pandas/issues/57500
@@ -546,7 +520,7 @@ class UserPerformance:
             task_worktime_by_phase_user: タスク、フェーズ、ユーザーごとの作業時間や生産量が格納されたオブジェクト。生産量やタスクにかかった作業時間の取得に利用します。
 
 
-        """  # noqa: E501
+        """
 
         def drop_unnecessary_columns(df: pandas.DataFrame) -> pandas.DataFrame:
             """
@@ -593,9 +567,7 @@ class UserPerformance:
         df = df.join(cls._create_df_stdev_monitored_worktime(task_worktime_by_phase_user))
 
         # 比例関係の列を計算して追加する
-        cls._add_ratio_column_for_productivity_per_user(
-            df, phase_list=phase_list, production_volume_columns=task_worktime_by_phase_user.production_volume_columns
-        )
+        cls._add_ratio_column_for_productivity_per_user(df, phase_list=phase_list, production_volume_columns=task_worktime_by_phase_user.production_volume_columns)
 
         # 出力に不要な列を削除する
         df = drop_unnecessary_columns(df)
@@ -611,9 +583,7 @@ class UserPerformance:
 
         df = df.sort_values(["user_id"])
         # `df.reset_index()`を実行する理由：indexである`account_id`を列にするため
-        return cls(
-            df.reset_index(), task_completion_criteria, custom_production_volume_list=task_worktime_by_phase_user.custom_production_volume_list
-        )
+        return cls(df.reset_index(), task_completion_criteria, custom_production_volume_list=task_worktime_by_phase_user.custom_production_volume_list)
 
     @classmethod
     def _convert_column_dtypes(cls, df: pandas.DataFrame) -> pandas.DataFrame:
@@ -663,11 +633,7 @@ class UserPerformance:
             ("real_monitored_worktime_hour", "acceptance"),
         ]
 
-        monitored_worktime_columns = (
-            [("monitored_worktime_hour", "sum")]
-            + [("monitored_worktime_hour", phase) for phase in phase_list]
-            + [("monitored_worktime_ratio", phase) for phase in phase_list]
-        )
+        monitored_worktime_columns = [("monitored_worktime_hour", "sum")] + [("monitored_worktime_hour", phase) for phase in phase_list] + [("monitored_worktime_ratio", phase) for phase in phase_list]
         production_columns = [("task_count", phase) for phase in phase_list]
         for production_volume_column in production_volume_columns:
             production_columns.extend([(production_volume_column, phase) for phase in phase_list])
@@ -685,10 +651,7 @@ class UserPerformance:
 
         inspection_comment_columns = [
             ("pointed_out_inspection_comment_count", TaskPhase.ANNOTATION.value),
-            *[
-                (f"pointed_out_inspection_comment_count/{production_volume_column}", TaskPhase.ANNOTATION.value)
-                for production_volume_column in production_volume_columns
-            ],
+            *[(f"pointed_out_inspection_comment_count/{production_volume_column}", TaskPhase.ANNOTATION.value) for production_volume_column in production_volume_columns],
         ]
 
         rejected_count_columns = [
@@ -815,48 +778,32 @@ class UserPerformance:
         """
         # ゼロ割の警告を無視する
         with numpy.errstate(divide="ignore", invalid="ignore"):
-            series[("real_monitored_worktime_hour/real_actual_worktime_hour", "sum")] = (
-                series[("real_monitored_worktime_hour", "sum")] / series[("real_actual_worktime_hour", "sum")]
-            )
+            series[("real_monitored_worktime_hour/real_actual_worktime_hour", "sum")] = series[("real_monitored_worktime_hour", "sum")] / series[("real_actual_worktime_hour", "sum")]
 
             for phase in phase_list:
                 # Annofab時間の比率を算出
-                # 計測作業時間の合計値が0により、monitored_worktime_ratioはnanになる場合は、教師付の実績作業時間を実績作業時間の合計値になるようなmonitored_worktime_ratioに変更する  # noqa: E501
+                # 計測作業時間の合計値が0により、monitored_worktime_ratioはnanになる場合は、教師付の実績作業時間を実績作業時間の合計値になるようなmonitored_worktime_ratioに変更する
                 if series[("monitored_worktime_hour", "sum")] == 0:
                     if phase == TaskPhase.ANNOTATION.value:
                         series[("monitored_worktime_ratio", phase)] = 1
                     else:
                         series[("monitored_worktime_ratio", phase)] = 0
                 else:
-                    series[("monitored_worktime_ratio", phase)] = (
-                        series[("monitored_worktime_hour", phase)] / series[("monitored_worktime_hour", "sum")]
-                    )
+                    series[("monitored_worktime_ratio", phase)] = series[("monitored_worktime_hour", phase)] / series[("monitored_worktime_hour", "sum")]
 
                 # Annofab時間の比率から、Annowork時間を予測する
                 series[("actual_worktime_hour", phase)] = series[("actual_worktime_hour", "sum")] * series[("monitored_worktime_ratio", phase)]
 
                 # 生産性を算出
-                series[("monitored_worktime_hour/input_data_count", phase)] = (
-                    series[("monitored_worktime_hour", phase)] / series[("input_data_count", phase)]
-                )
-                series[("actual_worktime_hour/input_data_count", phase)] = (
-                    series[("actual_worktime_hour", phase)] / series[("input_data_count", phase)]
-                )
+                series[("monitored_worktime_hour/input_data_count", phase)] = series[("monitored_worktime_hour", phase)] / series[("input_data_count", phase)]
+                series[("actual_worktime_hour/input_data_count", phase)] = series[("actual_worktime_hour", phase)] / series[("input_data_count", phase)]
 
-                series[("monitored_worktime_hour/annotation_count", phase)] = (
-                    series[("monitored_worktime_hour", phase)] / series[("annotation_count", phase)]
-                )
-                series[("actual_worktime_hour/annotation_count", phase)] = (
-                    series[("actual_worktime_hour", phase)] / series[("annotation_count", phase)]
-                )
+                series[("monitored_worktime_hour/annotation_count", phase)] = series[("monitored_worktime_hour", phase)] / series[("annotation_count", phase)]
+                series[("actual_worktime_hour/annotation_count", phase)] = series[("actual_worktime_hour", phase)] / series[("annotation_count", phase)]
 
             phase = TaskPhase.ANNOTATION.value
-            series[("pointed_out_inspection_comment_count/annotation_count", phase)] = (
-                series[("pointed_out_inspection_comment_count", phase)] / series[("annotation_count", phase)]
-            )
-            series[("pointed_out_inspection_comment_count/input_data_count", phase)] = (
-                series[("pointed_out_inspection_comment_count", phase)] / series[("input_data_count", phase)]
-            )
+            series[("pointed_out_inspection_comment_count/annotation_count", phase)] = series[("pointed_out_inspection_comment_count", phase)] / series[("annotation_count", phase)]
+            series[("pointed_out_inspection_comment_count/input_data_count", phase)] = series[("pointed_out_inspection_comment_count", phase)] / series[("input_data_count", phase)]
             series[("rejected_count/task_count", phase)] = series[("rejected_count", phase)] / series[("task_count", phase)]
 
     def get_production_volume_name(self, production_volume_column: str) -> str:
@@ -929,9 +876,7 @@ class UserPerformance:
         y_column = f"{worktime_type.value}_worktime_minute/{production_volume_column}"
         # 分単位の生産性を算出する
         for phase in self.phase_list:
-            df[(f"{worktime_type.value}_worktime_minute/{production_volume_column}", phase)] = (
-                df[(f"{worktime_type.value}_worktime_hour/{production_volume_column}", phase)] * 60
-            )
+            df[(f"{worktime_type.value}_worktime_minute/{production_volume_column}", phase)] = df[(f"{worktime_type.value}_worktime_hour/{production_volume_column}", phase)] * 60
 
         for biography_index, biography in enumerate(sorted(set(df["biography"]))):
             for scatter_obj, phase in zip(scatter_obj_list, self.phase_list):
@@ -1074,9 +1019,7 @@ class UserPerformance:
 
         write_bokeh_graph(bokeh.layouts.column(element_list), output_file)
 
-    def plot_quality_and_productivity(
-        self, output_file: Path, worktime_type: WorktimeType, production_volume_column: str, *, metadata: Optional[dict[str, Any]] = None
-    ) -> None:
+    def plot_quality_and_productivity(self, output_file: Path, worktime_type: WorktimeType, production_volume_column: str, *, metadata: Optional[dict[str, Any]] = None) -> None:
         """
         作業時間を元に算出した生産性と品質の関係を、メンバごとにプロットする
         """
@@ -1157,9 +1100,7 @@ class UserPerformance:
         df = self.convert_df_suitable_for_bokeh(self.df)
         PHASE = TaskPhase.ANNOTATION.value  # noqa: N806
 
-        df[(f"{worktime_type.value}_worktime_minute/{production_volume_column}", PHASE)] = (
-            df[(f"{worktime_type.value}_worktime_hour/{production_volume_column}", PHASE)] * 60
-        )
+        df[(f"{worktime_type.value}_worktime_minute/{production_volume_column}", PHASE)] = df[(f"{worktime_type.value}_worktime_hour/{production_volume_column}", PHASE)] * 60
         logger.debug(f"{output_file} を出力します。")
 
         production_volume_name = self.get_production_volume_name(production_volume_column)

annofabcli/statistics/visualization/dataframe/whole_performance.py

@@ -84,9 +84,7 @@ class WholePerformance:
         df_task = df_task_worktime_by_phase_user[["project_id", "task_id", "status"]].drop_duplicates()
 
         unique_keys_for_worktime = ["project_id", "task_id", "phase", "phase_stage"]
-        addable_columns_for_task = list(
-            set(df_task_worktime_by_phase_user.columns) - set(user_info_columns) - set(unique_keys_for_worktime) - {"status"}
-        )
+        addable_columns_for_task = list(set(df_task_worktime_by_phase_user.columns) - set(user_info_columns) - set(unique_keys_for_worktime) - {"status"})
         df_task_worktime_by_phase_user = df_task_worktime_by_phase_user.groupby(unique_keys_for_worktime)[addable_columns_for_task].sum()
         df_task_worktime_by_phase_user[user_info_columns] = PSEUDO_VALUE
         df_task_worktime_by_phase_user = df_task_worktime_by_phase_user.reset_index()
@@ -95,9 +93,7 @@ class WholePerformance:
 
         return UserPerformance.from_df_wrapper(
             worktime_per_date=WorktimePerDate(df_worktime_per_date),
-            task_worktime_by_phase_user=TaskWorktimeByPhaseUser(
-                df_task_worktime_by_phase_user, custom_production_volume_list=task_worktime_by_phase_user.custom_production_volume_list
-            ),
+            task_worktime_by_phase_user=TaskWorktimeByPhaseUser(df_task_worktime_by_phase_user, custom_production_volume_list=task_worktime_by_phase_user.custom_production_volume_list),
             task_completion_criteria=task_completion_criteria,
         )
 
@@ -115,7 +111,7 @@ class WholePerformance:
             worktime_per_date: 日ごとの作業時間が記載されたDataFrameを格納したオブジェクト。ユーザー情報の取得や、実際の作業時間（集計タスクに影響しない）の算出に利用します。
             task_worktime_by_phase_user: タスク、フェーズ、ユーザーごとの作業時間や生産量が格納されたオブジェクト。生産量やタスクにかかった作業時間の取得に利用します。
 
-        """  # noqa: E501
+        """
         # 1人が作業した場合のパフォーマンス情報を生成する
         all_user_performance = cls._create_all_user_performance(worktime_per_date, task_worktime_by_phase_user, task_completion_criteria)
 
@@ -142,9 +138,7 @@ class WholePerformance:
         return cls(df_all.iloc[0], task_completion_criteria, custom_production_volume_list=task_worktime_by_phase_user.custom_production_volume_list)
 
     @classmethod
-    def empty(
-        cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None
-    ) -> WholePerformance:
+    def empty(cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None) -> WholePerformance:
         """空のデータフレームを持つインスタンスを生成します。"""
 
         production_volume_columns = ["input_data_count", "annotation_count"]

annofabcli/statistics/visualization/dataframe/whole_productivity_per_date.py

@@ -141,9 +141,7 @@ class WholeProductivityPerCompletedDate:
         return pandas.DataFrame(index=[e.strftime("%Y-%m-%d") for e in pandas.date_range(start_date, end_date)])
 
     @classmethod
-    def from_df_wrapper(
-        cls, task: Task, worktime_per_date: WorktimePerDate, task_completion_criteria: TaskCompletionCriteria
-    ) -> WholeProductivityPerCompletedDate:
+    def from_df_wrapper(cls, task: Task, worktime_per_date: WorktimePerDate, task_completion_criteria: TaskCompletionCriteria) -> WholeProductivityPerCompletedDate:
         """
         完了日毎の全体の生産量、生産性を算出する。
 
@@ -250,9 +248,7 @@ class WholeProductivityPerCompletedDate:
             """速度情報の列を追加"""
             df[f"{numerator_column}/{denominator_column}"] = df[numerator_column] / df[denominator_column]
             # 1週間移動平均も出力
-            df[f"{numerator_column}/{denominator_column}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(
-                df[numerator_column]
-            ) / get_weekly_sum(df[denominator_column])
+            df[f"{numerator_column}/{denominator_column}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df[numerator_column]) / get_weekly_sum(df[denominator_column])
 
         # 累計情報を追加
         add_cumsum_column(df, column="task_count")
@@ -307,16 +303,10 @@ class WholeProductivityPerCompletedDate:
                     "unmonitored",
                 ]:
                     df[f"{category}_worktime_minute/{denominator}"] = df[f"{category}_worktime_hour"] * 60 / df[denominator]
-                    df[f"{category}_worktime_minute/{denominator}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = (
-                        get_weekly_sum(df[f"{category}_worktime_hour"]) * 60 / get_weekly_sum(df[denominator])
-                    )
+                    df[f"{category}_worktime_minute/{denominator}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df[f"{category}_worktime_hour"]) * 60 / get_weekly_sum(df[denominator])
 
-            df[f"actual_worktime_hour/task_count{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df["actual_worktime_hour"]) / get_weekly_sum(
-                df["task_count"]
-            )
-            df[f"monitored_worktime_hour/task_count{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(
-                df["monitored_worktime_hour"]
-            ) / get_weekly_sum(df["task_count"])
+            df[f"actual_worktime_hour/task_count{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df["actual_worktime_hour"]) / get_weekly_sum(df["task_count"])
+            df[f"monitored_worktime_hour/task_count{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df["monitored_worktime_hour"]) / get_weekly_sum(df["task_count"])
 
             for column in [
                 "task_count",
@@ -349,9 +339,7 @@ class WholeProductivityPerCompletedDate:
             )
             line_graph.add_secondary_y_axis(
                 "作業時間[時間]",
-                secondary_y_axis_range=DataRange1d(
-                    end=max(df["actual_worktime_hour"].max(), df["monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO
-                ),
+                secondary_y_axis_range=DataRange1d(end=max(df["actual_worktime_hour"].max(), df["monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO),
                 primary_y_axis_range=DataRange1d(end=df["task_count"].max() * SECONDARY_Y_RANGE_RATIO),
             )
 
@@ -396,9 +384,7 @@ class WholeProductivityPerCompletedDate:
             )
             line_graph.add_secondary_y_axis(
                 "作業時間[時間]",
-                secondary_y_axis_range=DataRange1d(
-                    end=max(df["actual_worktime_hour"].max(), df["monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO
-                ),
+                secondary_y_axis_range=DataRange1d(end=max(df["actual_worktime_hour"].max(), df["monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO),
                 primary_y_axis_range=DataRange1d(end=df["input_data_count"].max() * SECONDARY_Y_RANGE_RATIO),
             )
 
@@ -460,7 +446,7 @@ class WholeProductivityPerCompletedDate:
             ("monitored_acceptance_worktime", "計測作業時間(受入)"),
         ]
         if df["actual_worktime_hour"].sum() > 0:
-            # 条件分岐の理由：実績作業時間がないときは、非計測作業時間がマイナス値になり、分かりづらいグラフになるため。必要なときのみ非計測作業時間をプロットする  # noqa: E501
+            # 条件分岐の理由：実績作業時間がないときは、非計測作業時間がマイナス値になり、分かりづらいグラフになるため。必要なときのみ非計測作業時間をプロットする
             phase_prefix.append(("unmonitored_worktime", "非計測作業時間"))
 
         fig_info_list = [
@@ -572,9 +558,7 @@ class WholeProductivityPerCompletedDate:
             )
             line_graph.add_secondary_y_axis(
                 "作業時間[時間]",
-                secondary_y_axis_range=DataRange1d(
-                    end=max(df["cumsum_actual_worktime_hour"].max(), df["cumsum_monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO
-                ),
+                secondary_y_axis_range=DataRange1d(end=max(df["cumsum_actual_worktime_hour"].max(), df["cumsum_monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO),
                 primary_y_axis_range=DataRange1d(end=df["cumsum_task_count"].max() * SECONDARY_Y_RANGE_RATIO),
             )
 
@@ -628,9 +612,7 @@ class WholeProductivityPerCompletedDate:
             )
             line_graph.add_secondary_y_axis(
                 "作業時間[時間]",
-                secondary_y_axis_range=DataRange1d(
-                    end=max(df["cumsum_actual_worktime_hour"].max(), df["cumsum_monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO
-                ),
+                secondary_y_axis_range=DataRange1d(end=max(df["cumsum_actual_worktime_hour"].max(), df["cumsum_monitored_worktime_hour"].max()) * SECONDARY_Y_RANGE_RATIO),
                 primary_y_axis_range=DataRange1d(end=df["cumsum_input_data_count"].max() * SECONDARY_Y_RANGE_RATIO),
             )
 
@@ -762,9 +744,7 @@ class WholeProductivityPerCompletedDate:
         write_bokeh_graph(bokeh.layouts.column(element_list), output_file)
 
     @classmethod
-    def empty(
-        cls, *, task_completion_criteria: TaskCompletionCriteria, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None
-    ) -> WholeProductivityPerCompletedDate:
+    def empty(cls, *, task_completion_criteria: TaskCompletionCriteria, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None) -> WholeProductivityPerCompletedDate:
         df = pandas.DataFrame(columns=cls.get_columns(custom_production_volume_list=custom_production_volume_list))
         return cls(df, task_completion_criteria, custom_production_volume_list=custom_production_volume_list)
 
@@ -868,9 +848,7 @@ class WholeProductivityPerFirstAnnotationStartedDate:
         def add_velocity_column(df: pandas.DataFrame, numerator_column: str, denominator_column: str) -> None:
             df[f"{numerator_column}/{denominator_column}"] = df[numerator_column] / df[denominator_column]
 
-            df[f"{numerator_column}/{denominator_column}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_moving_average(
-                df[numerator_column]
-            ) / get_weekly_moving_average(df[denominator_column])
+            df[f"{numerator_column}/{denominator_column}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_moving_average(df[numerator_column]) / get_weekly_moving_average(df[denominator_column])
 
         # annofab 計測時間から算出したvelocityを追加
         for column in production_volume_columns:
@@ -907,9 +885,7 @@ class WholeProductivityPerFirstAnnotationStartedDate:
                 "acceptance_worktime_hour",
             ]
         ].copy()
-        df_sub_task["first_annotation_started_date"] = df_sub_task["first_annotation_started_datetime"].map(
-            lambda e: datetime_to_date(e) if not pandas.isna(e) else None
-        )
+        df_sub_task["first_annotation_started_date"] = df_sub_task["first_annotation_started_datetime"].map(lambda e: datetime_to_date(e) if not pandas.isna(e) else None)
 
         value_columns = [
             *production_volume_columns,
@@ -924,18 +900,14 @@ class WholeProductivityPerFirstAnnotationStartedDate:
             aggfunc="sum",
         ).fillna(0)
         if len(df_agg_sub_task) > 0:
-            df_agg_sub_task["task_count"] = df_sub_task.pivot_table(
-                values=["task_id"], index="first_annotation_started_date", aggfunc="count"
-            ).fillna(0)
+            df_agg_sub_task["task_count"] = df_sub_task.pivot_table(values=["task_id"], index="first_annotation_started_date", aggfunc="count").fillna(0)
         else:
             # 列だけ作る
             df_agg_sub_task = df_agg_sub_task.assign(**dict.fromkeys(value_columns, 0), task_count=0)
 
         # 日付の一覧を生成
         if len(df_agg_sub_task) > 0:
-            df_date_base = pandas.DataFrame(
-                index=[e.strftime("%Y-%m-%d") for e in pandas.date_range(start=df_agg_sub_task.index.min(), end=df_agg_sub_task.index.max())]
-            )
+            df_date_base = pandas.DataFrame(index=[e.strftime("%Y-%m-%d") for e in pandas.date_range(start=df_agg_sub_task.index.min(), end=df_agg_sub_task.index.max())])
         else:
             df_date_base = pandas.DataFrame()
 
@@ -953,9 +925,7 @@ class WholeProductivityPerFirstAnnotationStartedDate:
         return True
 
     @classmethod
-    def empty(
-        cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None
-    ) -> WholeProductivityPerFirstAnnotationStartedDate:
+    def empty(cls, task_completion_criteria: TaskCompletionCriteria, *, custom_production_volume_list: Optional[list[ProductionVolumeColumn]] = None) -> WholeProductivityPerFirstAnnotationStartedDate:
         df = pandas.DataFrame(columns=cls.get_columns(custom_production_volume_list=custom_production_volume_list))
         return cls(df, task_completion_criteria, custom_production_volume_list=custom_production_volume_list)
 
@@ -1018,9 +988,7 @@ class WholeProductivityPerFirstAnnotationStartedDate:
             for denominator in [e.value for e in production_volume_list]:
                 for numerator in ["worktime", "annotation_worktime", "inspection_worktime", "acceptance_worktime"]:
                     df[f"{numerator}_minute/{denominator}"] = df[f"{numerator}_hour"] * 60 / df[denominator]
-                    df[f"{numerator}_minute/{denominator}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = (
-                        get_weekly_sum(df[f"{numerator}_hour"]) * 60 / get_weekly_sum(df[denominator])
-                    )
+                    df[f"{numerator}_minute/{denominator}{WEEKLY_MOVING_AVERAGE_COLUMN_SUFFIX}"] = get_weekly_sum(df[f"{numerator}_hour"]) * 60 / get_weekly_sum(df[denominator])
 
         def create_div_element() -> Div:
             """

annofabcli/statistics/visualization/dataframe/worktime_per_date.py

@@ -110,9 +110,7 @@ class WorktimePerDate:
         """
         df = self.df.copy()
         df["monitored_acceptance_worktime_hour"] = 0
-        df["monitored_worktime_hour"] = (
-            df["monitored_annotation_worktime_hour"] + df["monitored_inspection_worktime_hour"] + df["monitored_acceptance_worktime_hour"]
-        )
+        df["monitored_worktime_hour"] = df["monitored_annotation_worktime_hour"] + df["monitored_inspection_worktime_hour"] + df["monitored_acceptance_worktime_hour"]
         return WorktimePerDate(df)
 
     def is_empty(self) -> bool:
@@ -180,9 +178,7 @@ class WorktimePerDate:
             if f"monitored_{phase}_worktime_hour" not in df.columns:
                 df[f"monitored_{phase}_worktime_hour"] = 0
 
-        df["monitored_worktime_hour"] = (
-            df["monitored_annotation_worktime_hour"] + df["monitored_inspection_worktime_hour"] + df["monitored_acceptance_worktime_hour"]
-        )
+        df["monitored_worktime_hour"] = df["monitored_annotation_worktime_hour"] + df["monitored_inspection_worktime_hour"] + df["monitored_acceptance_worktime_hour"]
 
         if not actual_worktime.is_empty():
             df = df.merge(
@@ -450,9 +446,7 @@ class WorktimePerDate:
         # またpandas.NAを持つDataFrameも描画できないので、numpy.nanに変換する
         # TODO この問題が解決されたら、削除する
         # https://qiita.com/yuji38kwmt/items/b5da6ed521e827620186
-        df_cumulative = df_cumulative.astype(
-            {"date": "object", "account_id": "object", "user_id": "object", "username": "object", "biography": "object"}
-        )
+        df_cumulative = df_cumulative.astype({"date": "object", "account_id": "object", "user_id": "object", "username": "object", "biography": "object"})
         df_cumulative.replace(pandas.NA, numpy.nan, inplace=True)
 
         line_count = 0

annofabcli/statistics/visualization/filtering_query.py

@@ -31,9 +31,7 @@ def _get_first_annotation_started_datetime(sub_task_history_list: list[dict[str,
     1個のタスクのタスク履歴一覧から、最初に教師付フェーズを作業した日時を取得します。
     """
     task_history_list_with_annotation_phase = [
-        e
-        for e in sub_task_history_list
-        if e["phase"] == "annotation" and e["account_id"] is not None and isoduration_to_hour(e["accumulated_labor_time_milliseconds"]) > 0
+        e for e in sub_task_history_list if e["phase"] == "annotation" and e["account_id"] is not None and isoduration_to_hour(e["accumulated_labor_time_milliseconds"]) > 0
     ]
     if len(task_history_list_with_annotation_phase) == 0:
         return None
@@ -41,9 +39,7 @@ def _get_first_annotation_started_datetime(sub_task_history_list: list[dict[str,
     return task_history_list_with_annotation_phase[0]["started_datetime"]
 
 
-def filter_task_histories(
-    task_histories: dict[str, list[dict[str, Any]]], *, start_date: Optional[str] = None, end_date: Optional[str] = None
-) -> dict[str, list[dict[str, Any]]]:
+def filter_task_histories(task_histories: dict[str, list[dict[str, Any]]], *, start_date: Optional[str] = None, end_date: Optional[str] = None) -> dict[str, list[dict[str, Any]]]:
     """
     タスク履歴を絞り込みます。
 

annofabcli/statistics/visualization/project_dir.py

@@ -173,9 +173,7 @@ class ProjectDir(DataClassJsonMixin):
         phase_name = self.get_phase_name_for_filename(phase)
         obj.to_csv(self.project_dir / Path(f"{phase_name}者_{phase_name}開始日list.csv"))
 
-    def write_performance_line_graph_per_date(
-        self, obj: AbstractPhaseProductivityPerDate, phase: TaskPhase, *, user_id_list: Optional[list[str]] = None
-    ) -> None:
+    def write_performance_line_graph_per_date(self, obj: AbstractPhaseProductivityPerDate, phase: TaskPhase, *, user_id_list: Optional[list[str]] = None) -> None:
         """
         指定したフェーズの開始日ごとの作業時間や生産性情報を、折れ線グラフとして出力します。
         """
@@ -199,8 +197,7 @@ class ProjectDir(DataClassJsonMixin):
             obj.plot_production_volume_metrics(
                 production_volume_column=custom_production_volume.value,
                 production_volume_name=custom_production_volume.name,
-                output_file=output_dir
-                / Path(f"{phase_name}者用/折れ線-横軸_{phase_name}開始日-縦軸_{custom_production_volume.name}単位の指標-{phase_name}者用.html"),
+                output_file=output_dir / Path(f"{phase_name}者用/折れ線-横軸_{phase_name}開始日-縦軸_{custom_production_volume.name}単位の指標-{phase_name}者用.html"),
                 target_user_id_list=user_id_list,
                 metadata=self.metadata,
             )
@@ -209,14 +206,10 @@ class ProjectDir(DataClassJsonMixin):
         """`全体の生産性と品質.csv`を読み込む。"""
         file = self.project_dir / self.FILENAME_WHOLE_PERFORMANCE
         if file.exists():
-            return WholePerformance.from_csv(
-                file, custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria
-            )
+            return WholePerformance.from_csv(file, custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria)
         else:
             logger.warning(f"'{file!s}'を読み込もうとしましたが、ファイルは存在しません。")
-            return WholePerformance.empty(
-                custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria
-            )
+            return WholePerformance.empty(custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria)
 
     def write_whole_performance(self, whole_performance: WholePerformance) -> None:
         """`全体の生産性と品質.csv`を出力します。"""
@@ -234,9 +227,7 @@ class ProjectDir(DataClassJsonMixin):
                 custom_production_volume_list=self.custom_production_volume_list,
             )
         else:
-            return WholeProductivityPerCompletedDate.empty(
-                task_completion_criteria=self.task_completion_criteria, custom_production_volume_list=self.custom_production_volume_list
-            )
+            return WholeProductivityPerCompletedDate.empty(task_completion_criteria=self.task_completion_criteria, custom_production_volume_list=self.custom_production_volume_list)
 
     def write_whole_productivity_per_date(self, obj: WholeProductivityPerCompletedDate) -> None:
         """
@@ -263,9 +254,7 @@ class ProjectDir(DataClassJsonMixin):
                 custom_production_volume_list=self.custom_production_volume_list,
             )
         else:
-            return WholeProductivityPerFirstAnnotationStartedDate.empty(
-                self.task_completion_criteria, custom_production_volume_list=self.custom_production_volume_list
-            )
+            return WholeProductivityPerFirstAnnotationStartedDate.empty(self.task_completion_criteria, custom_production_volume_list=self.custom_production_volume_list)
 
     def write_whole_productivity_per_first_annotation_started_date(self, obj: WholeProductivityPerFirstAnnotationStartedDate) -> None:
         """
@@ -285,14 +274,10 @@ class ProjectDir(DataClassJsonMixin):
         """
         file = self.project_dir / self.FILENAME_USER_PERFORMANCE
         if file.exists():
-            return UserPerformance.from_csv(
-                file, custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria
-            )
+            return UserPerformance.from_csv(file, custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria)
         else:
             logger.warning(f"'{file!s}'を読み込もうとしましたが、ファイルは存在しません。")
-            return UserPerformance.empty(
-                custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria
-            )
+            return UserPerformance.empty(custom_production_volume_list=self.custom_production_volume_list, task_completion_criteria=self.task_completion_criteria)
 
     def write_user_performance(self, user_performance: UserPerformance) -> None:
         """
@@ -407,9 +392,7 @@ class ProjectDir(DataClassJsonMixin):
 
     def write_worktime_line_graph(self, obj: WorktimePerDate, user_id_list: Optional[list[str]] = None) -> None:
         """横軸が日付、縦軸がユーザごとの作業時間である折れ線グラフを出力します。"""
-        obj.plot_cumulatively(
-            self.project_dir / "line-graph/累積折れ線-横軸_日-縦軸_作業時間.html", target_user_id_list=user_id_list, metadata=self.metadata
-        )
+        obj.plot_cumulatively(self.project_dir / "line-graph/累積折れ線-横軸_日-縦軸_作業時間.html", target_user_id_list=user_id_list, metadata=self.metadata)
 
     def read_project_info(self) -> ProjectInfo:
         """