annofabcli 1.111.1__tar.gz → 1.111.2__tar.gz

{annofabcli-1.111.1 → annofabcli-1.111.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: annofabcli
-Version: 1.111.1
+Version: 1.111.2
 Summary: Utility Command Line Interface for AnnoFab
 Author: Kurusugawa Computer Inc.
 License: MIT

{annofabcli-1.111.1 → annofabcli-1.111.2}/annofabcli/statistics/visualization/dataframe/annotation_duration.py

@@ -23,7 +23,7 @@ class AnnotationDuration:
         return [
             "project_id",
             "task_id",
-            "annotation_duration_second",
+            "annotation_duration_minute",
         ]
 
     @classmethod
@@ -62,7 +62,7 @@ class AnnotationDuration:
         """
         logger.debug(f"アノテーションZIPファイルから区間アノテーションの長さを計算します。 :: project_id='{project_id}', file='{annotation_zip!s}'")
 
-        result: dict[tuple[str, str], float] = defaultdict(float)  # key:(project_id, task_id), value:合計アノテーション時間（秒）
+        result: dict[tuple[str, str], float] = defaultdict(float)  # key:(project_id, task_id), value:合計アノテーション時間（分）
 
         for index, parser in enumerate(lazy_parse_simple_annotation_zip(annotation_zip)):
             simple_annotation = parser.load_json()
@@ -83,7 +83,7 @@ class AnnotationDuration:
                     # 区間アノテーションの場合
                     begin = data["begin"]
                     end = data["end"]
-                    total_duration += (end - begin) / 1000.0  # ミリ秒から秒に変換
+                    total_duration += (end - begin) / 1000.0 / 60.0  # ミリ秒から分に変換
 
             result[(project_id, parser.task_id)] += total_duration
 
@@ -114,7 +114,7 @@ class AnnotationDuration:
         df_dtype: dict[str, str] = {
             "project_id": "string",
             "task_id": "string",
-            "annotation_duration_second": "float64",
+            "annotation_duration_minute": "float64",
         }
 
         df = pandas.DataFrame(columns=cls.columns()).astype(df_dtype)

{annofabcli-1.111.1 → annofabcli-1.111.2}/annofabcli/statistics/visualization/visualization_source_files.py

@@ -39,6 +39,7 @@ class VisualizationSourceFiles:
         self.task_history_json_path = target_dir / f"{self.project_id}__task-history.json"
         self.task_history_event_json_path = target_dir / f"{self.project_id}__task-history-event.json"
         self.annotation_zip_path = target_dir / f"{self.project_id}__annotation.zip"
+        self.input_data_json_path = target_dir / f"{self.project_id}__input_data.json"
 
         self.logging_prefix = f"project_id='{project_id}'"
 
@@ -99,6 +100,52 @@ class VisualizationSourceFiles:
         logger.debug(f"{self.logging_prefix}: '{self.comment_json_path}'を読み込みました。{len(comment_list)}件のコメントが含まれています。")
         return comment_list
 
+    def read_input_data_json(self) -> list[dict[str, Any]]:
+        """
+        入力データ全件ファイルを読み込みます。
+
+        Returns:
+            全入力データの一覧
+        """
+        with self.input_data_json_path.open(encoding="utf-8") as f:
+            input_data_list = json.load(f)
+
+        logger.debug(f"{self.logging_prefix}: '{self.input_data_json_path}'を読み込みました。{len(input_data_list)}件の入力データが含まれています。")
+        return input_data_list
+
+    def get_video_duration_minutes_by_task_id(self) -> dict[str, float]:
+        """
+        動画プロジェクトの場合、タスクIDごとの動画の長さ（分単位）を取得します。
+
+        Returns:
+            key: task_id, value: 動画の長さ（分）
+        """
+        tasks = self.read_tasks_json()
+        input_data_list = self.read_input_data_json()
+
+        # 入力データIDをキーとした辞書を作成
+        dict_input_data_by_id = {input_data["input_data_id"]: input_data for input_data in input_data_list}
+
+        result = {}
+        for task in tasks:
+            task_id = task["task_id"]
+            input_data_id_list = task["input_data_id_list"]
+            assert len(input_data_id_list) == 1, f"task_id='{task_id}'には複数の入力データが含まれています。"
+            input_data_id = input_data_id_list[0]
+            input_data = dict_input_data_by_id.get(input_data_id)
+
+            if input_data is None:
+                logger.warning(f"task_id='{task_id}' :: タスクに含まれている入力データ（input_data_id='{input_data_id}'）は、見つかりません。")
+                result[task_id] = 0.0
+                continue
+
+            video_duration_second = input_data["system_metadata"]["input_duration"]
+
+            # 秒から分に変換
+            result[task_id] = video_duration_second / 60.0
+
+        return result
+
     def write_files(self, *, is_latest: bool = False, should_get_task_histories_one_of_each: bool = False, should_download_annotation_zip: bool = True) -> None:
         """
         可視化に必要なファイルを作成します。
@@ -117,6 +164,8 @@ class VisualizationSourceFiles:
 
         downloading_obj.download_task_json(self.project_id, dest_path=self.task_json_path, is_latest=is_latest, wait_options=wait_options)
 
+        downloading_obj.download_input_data_json(self.project_id, dest_path=self.input_data_json_path, is_latest=is_latest, wait_options=wait_options)
+
         if should_download_annotation_zip:
             downloading_obj.download_annotation_zip(
                 self.project_id,

{annofabcli-1.111.1 → annofabcli-1.111.2}/annofabcli/statistics/visualize_statistics.py

@@ -12,7 +12,7 @@ from typing import Any, Callable, Optional
 
 import annofabapi
 import pandas
-from annofabapi.models import InputDataType, ProjectMemberRole, TaskPhase
+from annofabapi.models import ProjectMemberRole, TaskPhase
 
 import annofabcli
 from annofabcli.common.cli import (
@@ -82,6 +82,7 @@ class WriteCsvGraph:
         production_volume_include_labels: Optional[list[str]] = None,
         production_volume_exclude_labels: Optional[list[str]] = None,
         include_annotation_duration_seconds: bool = False,
+        include_video_duration_minutes: bool = False,
     ) -> None:
         self.service = service
         self.project_id = project_id
@@ -98,6 +99,7 @@ class WriteCsvGraph:
         self.production_volume_include_labels = production_volume_include_labels
         self.production_volume_exclude_labels = production_volume_exclude_labels
         self.include_annotation_duration_seconds = include_annotation_duration_seconds
+        self.include_video_duration_minutes = include_video_duration_minutes
 
         self.task: Optional[Task] = None
         self.worktime_per_date: Optional[WorktimePerDate] = None
@@ -119,14 +121,18 @@ class WriteCsvGraph:
 
     def _get_task(self) -> Task:
         if self.task is None:
+            custom_production_volume = self._prepare_custom_production_volume()
+
             if self.annotation_count is None:
-                # アノテーションZIPからアノテーション数を取得
-                annotation_count = AnnotationCount.from_annotation_zip(
-                    self.visualize_source_files.annotation_zip_path,
-                    project_id=self.project_id,
-                    include_labels=self.production_volume_include_labels,
-                    exclude_labels=self.production_volume_exclude_labels,
-                )
+                if self.visualize_source_files.annotation_zip_path.exists():
+                    annotation_count = AnnotationCount.from_annotation_zip(
+                        self.visualize_source_files.annotation_zip_path,
+                        project_id=self.project_id,
+                        include_labels=self.production_volume_include_labels,
+                        exclude_labels=self.production_volume_exclude_labels,
+                    )
+                else:
+                    annotation_count = AnnotationCount.empty()
             else:
                 annotation_count = self.annotation_count
 
@@ -137,36 +143,6 @@ class WriteCsvGraph:
             new_tasks = filter_tasks(tasks, self.task_completion_criteria, self.filtering_query, task_histories=task_histories)
             logger.debug(f"project_id='{self.project_id}' :: 集計対象タスクは {len(new_tasks)} / {len(tasks)} 件です。")
 
-            # annotation_duration_secondsを生産量に含める場合、アノテーション時間を計算
-            custom_production_volume = self.custom_production_volume
-            if self.include_annotation_duration_seconds:
-                logger.debug(f"project_id='{self.project_id}' :: 区間アノテーションの長さ（'annotation_duration_second'）を計算します。")
-                annotation_duration_obj = AnnotationDuration.from_annotation_zip(
-                    self.visualize_source_files.annotation_zip_path,
-                    project_id=self.project_id,
-                    include_labels=self.production_volume_include_labels,
-                    exclude_labels=self.production_volume_exclude_labels,
-                )
-
-                if custom_production_volume is not None:
-                    # 既存のCustomProductionVolumeのデータと結合
-                    if not custom_production_volume.is_empty():
-                        annotation_duration_df = pandas.merge(custom_production_volume.df, annotation_duration_obj.df, on=["project_id", "task_id"], how="outer")
-                    else:
-                        annotation_duration_df = annotation_duration_obj.df
-
-                    # annotation_duration_secondを含む新しいProductionVolumeColumnリストを作成
-                    annotation_duration_column = ProductionVolumeColumn(value="annotation_duration_second", name="区間アノテーションの長さ（秒）")
-                    new_production_volume_list = list(custom_production_volume.custom_production_volume_list)
-                    if annotation_duration_column not in new_production_volume_list:
-                        new_production_volume_list.append(annotation_duration_column)
-
-                    custom_production_volume = CustomProductionVolume(annotation_duration_df, custom_production_volume_list=new_production_volume_list)
-                else:
-                    # CustomProductionVolumeが存在しない場合、新規作成
-                    annotation_duration_column = ProductionVolumeColumn(value="annotation_duration_second", name="区間アノテーションの長さ（秒）")
-                    custom_production_volume = CustomProductionVolume(annotation_duration_obj.df, custom_production_volume_list=[annotation_duration_column])
-
             self.task = Task.from_api_content(
                 tasks=new_tasks,
                 task_histories=task_histories,
@@ -180,6 +156,80 @@ class WriteCsvGraph:
 
         return self.task
 
+    def _prepare_custom_production_volume(self) -> Optional[CustomProductionVolume]:
+        """カスタム生産量の準備を行う"""
+        custom_production_volume = self.custom_production_volume
+
+        # annotation_duration_secondsを生産量に含める場合、アノテーション時間を計算
+        if self.include_annotation_duration_seconds:
+            custom_production_volume = self._add_annotation_duration(custom_production_volume)
+
+        # 動画プロジェクトの場合、動画の長さ（分）を生産量に含める
+        if self.include_video_duration_minutes:
+            custom_production_volume = self._add_video_duration(custom_production_volume)
+
+        return custom_production_volume
+
+    def _add_annotation_duration(self, custom_production_volume: Optional[CustomProductionVolume]) -> CustomProductionVolume:
+        """区間アノテーションの長さを生産量に追加する"""
+        logger.debug(f"project_id='{self.project_id}' :: 区間アノテーションの長さ（'annotation_duration_minute'）を計算します。")
+        annotation_duration_obj = AnnotationDuration.from_annotation_zip(
+            self.visualize_source_files.annotation_zip_path,
+            project_id=self.project_id,
+            include_labels=self.production_volume_include_labels,
+            exclude_labels=self.production_volume_exclude_labels,
+        )
+        annotation_duration_column = ProductionVolumeColumn(value="annotation_duration_minute", name="区間アノテーションの長さ（分）")
+
+        if custom_production_volume is not None:
+            # 既存のCustomProductionVolumeのデータと結合
+            if not custom_production_volume.is_empty():
+                annotation_duration_df = pandas.merge(custom_production_volume.df, annotation_duration_obj.df, on=["project_id", "task_id"], how="outer")
+            else:
+                annotation_duration_df = annotation_duration_obj.df
+
+            # annotation_duration_minuteを含む新しいProductionVolumeColumnリストを作成
+            new_production_volume_list = list(custom_production_volume.custom_production_volume_list)
+            if annotation_duration_column not in new_production_volume_list:
+                new_production_volume_list.append(annotation_duration_column)
+
+            return CustomProductionVolume(annotation_duration_df, custom_production_volume_list=new_production_volume_list)
+        else:
+            # CustomProductionVolumeが存在しない場合、新規作成
+            return CustomProductionVolume(annotation_duration_obj.df, custom_production_volume_list=[annotation_duration_column])
+
+    def _add_video_duration(self, custom_production_volume: Optional[CustomProductionVolume]) -> CustomProductionVolume:
+        """動画の長さ（分）を生産量に追加する"""
+        logger.debug(f"project_id='{self.project_id}' :: 動画の長さ（'video_duration_minute'）を計算します。")
+        video_duration_by_task_id = self.visualize_source_files.get_video_duration_minutes_by_task_id()
+
+        # DataFrameの作成
+        video_duration_data = [{"project_id": self.project_id, "task_id": task_id, "video_duration_minute": duration} for task_id, duration in video_duration_by_task_id.items()]
+        if len(video_duration_data) == 0:
+            video_duration_df = pandas.DataFrame(columns=["project_id", "task_id", "video_duration_minute"])
+        else:
+            video_duration_df = pandas.DataFrame(video_duration_data)
+
+        video_duration_df = video_duration_df.astype({"project_id": "string", "task_id": "string", "video_duration_minute": "float64"})
+        video_duration_column = ProductionVolumeColumn(value="video_duration_minute", name="動画の長さ（分）")
+
+        if custom_production_volume is not None:
+            # 既存のCustomProductionVolumeのデータと結合
+            if not custom_production_volume.is_empty():
+                merged_df = pandas.merge(custom_production_volume.df, video_duration_df, on=["project_id", "task_id"], how="outer")
+            else:
+                merged_df = video_duration_df
+
+            # video_duration_minuteを含む新しいProductionVolumeColumnリストを作成
+            new_production_volume_list = list(custom_production_volume.custom_production_volume_list)
+            if video_duration_column not in new_production_volume_list:
+                new_production_volume_list.append(video_duration_column)
+
+            return CustomProductionVolume(merged_df, custom_production_volume_list=new_production_volume_list)
+        else:
+            # CustomProductionVolumeが存在しない場合、新規作成
+            return CustomProductionVolume(video_duration_df, custom_production_volume_list=[video_duration_column])
+
     def _get_task_worktime_obj(self) -> TaskWorktimeByPhaseUser:
         if self.task_worktime_obj is None:
             task_history = TaskHistory.from_api_content(self.visualize_source_files.read_task_histories_json())
@@ -370,6 +420,9 @@ class VisualizingStatisticsMain:
         # 動画プロジェクトの場合、annotation_duration_secondを生産量に含める
         custom_production_volume = self.custom_production_volume
 
+        # 動画プロジェクトかどうかを判定
+        is_video_project = project_info.input_data_type == "movie"
+
         project_dir = ProjectDir(
             output_project_dir,
             self.task_completion_criteria,
@@ -429,7 +482,8 @@ class VisualizingStatisticsMain:
             output_only_text=self.output_only_text,
             production_volume_include_labels=self.production_volume_include_labels,
             production_volume_exclude_labels=self.production_volume_exclude_labels,
-            include_annotation_duration_seconds=(project_info.input_data_type == InputDataType.MOVIE.value),
+            include_annotation_duration_seconds=is_video_project,
+            include_video_duration_minutes=is_video_project,
         )
 
         write_obj._catch_exception(write_obj.write_user_performance)()  # noqa: SLF001

{annofabcli-1.111.1 → annofabcli-1.111.2}/docs/command_reference/statistics/visualize.rst

@@ -241,8 +241,10 @@ CSVには以下の列が存在している必要があります。
 
 .. note::
     
-    動画プロジェクトでは、区間アノテーションの長さ ``annotation_duration_second`` が、生産量として追加されます。
+    動画プロジェクトでは、以下の生産量が追加されます。
     
+    * ``annotation_duration_minute`` ： 区間アノテーションの長さ（分）
+    * ``video_duration_minute`` ： 動画の長さ（分）
 
 
 .. _production_volume_include_or_exclude_label:
@@ -253,7 +255,7 @@ CSVには以下の列が存在している必要があります。
 ``--production_volume_include_label`` または ``--production_volume_exclude_label`` にラベル名を指定することで、以下の生産量の集計対象に含めるラベルや除外するラベルを指定できます。
 
 * ``annotation_count``
-* ``annotation_duration_second`` （動画プロジェクトの場合のみ）
+* ``annotation_duration_minute`` （動画プロジェクトの場合のみ）
 
 
 

annofabcli-1.111.2/tests/statistics/test_visualization_source_files.py

@@ -0,0 +1,55 @@
+import json
+from unittest.mock import Mock
+
+import pytest
+
+from annofabcli.statistics.visualization.visualization_source_files import VisualizationSourceFiles
+
+
+class TestVisualizationSourceFiles:
+    @pytest.fixture
+    def mock_service(self):
+        return Mock()
+
+    @pytest.fixture
+    def visualization_source_files(self, mock_service, tmp_path):
+        return VisualizationSourceFiles(annofab_service=mock_service, project_id="test_project", target_dir=tmp_path)
+
+    def test_init(self, visualization_source_files, tmp_path):
+        """初期化のテスト"""
+        assert visualization_source_files.project_id == "test_project"
+        assert visualization_source_files.target_dir == tmp_path
+        assert visualization_source_files.task_json_path == tmp_path / "test_project__task.json"
+        assert visualization_source_files.input_data_json_path == tmp_path / "test_project__input_data.json"
+
+    def test_get_video_duration_minutes_by_task_id_with_data(self, visualization_source_files, tmp_path):
+        """データありでの動画長さ計算テスト"""
+        # テストデータを作成
+        task_data = [{"task_id": "task1", "input_data_id_list": ["input1"]}, {"task_id": "task2", "input_data_id_list": ["input2"]}]
+
+        input_data = [
+            {
+                "input_data_id": "input1",
+                "system_metadata": {
+                    "input_duration": 120.0  # 2分 = 120秒
+                },
+            },
+            {
+                "input_data_id": "input2",
+                "system_metadata": {
+                    "input_duration": 180.0  # 3分 = 180秒
+                },
+            },
+        ]
+
+        task_json_path = tmp_path / "test_project__task.json"
+        input_data_json_path = tmp_path / "test_project__input_data.json"
+
+        task_json_path.write_text(json.dumps(task_data), encoding="utf-8")
+        input_data_json_path.write_text(json.dumps(input_data), encoding="utf-8")
+
+        result = visualization_source_files.get_video_duration_minutes_by_task_id()
+
+        # 期待値: 120秒 = 2分、180秒 = 3分
+        expected = {"task1": 2.0, "task2": 3.0}
+        assert result == expected