paddlex 3.0.0rc0__py3-none-any.whl → 3.0.1__py3-none-any.whl

paddlex/inference/models/ts_anomaly_detection/result.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,15 +12,20 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any
 import io
-import pandas as pd
-import matplotlib.pyplot as plt
+from typing import Any
+
 from PIL import Image
 
+from ....utils.deps import function_requires_deps, is_dep_available
 from ...common.result import BaseTSResult
 
+if is_dep_available("matplotlib"):
+    import matplotlib.pyplot as plt
+import pandas as pd
+
 
+@function_requires_deps("matplotlib")
 def visualize(forecast: pd.DataFrame) -> Image.Image:
     """
     Visualizes both the time series forecast and actual results, returning them as a Pillow image.
@@ -33,12 +38,18 @@ def visualize(forecast: pd.DataFrame) -> Image.Image:
     """
     plt.figure(figsize=(12, 6))
     forecast_columns = forecast.columns
-    index_name = forecast.index.name
+    forecast.index.name
     forecast.index = forecast.index.astype(str)
 
-    plt.step(forecast.index, forecast[forecast_columns[0]], where='post', label='Anomaly', color='red')
-    plt.title('Time Series Anomaly Detection')
-    plt.xlabel('Time')
+    plt.step(
+        forecast.index,
+        forecast[forecast_columns[0]],
+        where="post",
+        label="Anomaly",
+        color="red",
+    )
+    plt.title("Time Series Anomaly Detection")
+    plt.xlabel("Time")
     plt.ylabel(forecast_columns[0])
     plt.legend()
     plt.grid(True)
@@ -46,7 +57,7 @@ def visualize(forecast: pd.DataFrame) -> Image.Image:
     plt.xticks(rotation=45)
 
     buf = io.BytesIO()
-    plt.savefig(buf, bbox_inches='tight')
+    plt.savefig(buf, bbox_inches="tight")
     buf.seek(0)
     plt.close()
     image = Image.open(buf)
@@ -61,7 +72,7 @@ class TSAdResult(BaseTSResult):
         """apply"""
         anomaly = self["anomaly"]
         return {"res": visualize(anomaly)}
-    
+
     def _to_csv(self) -> Any:
         """
         Converts the anomaly detection results to a CSV format.

paddlex/inference/models/ts_classification/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

paddlex/inference/models/ts_classification/predictor.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,32 +12,23 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any, Union, Dict, List, Tuple
-import numpy as np
-import pandas as pd
-import os
 import copy
+import os
+from typing import Any, Dict, List, Tuple, Union
+
+import pandas as pd
 
 from ....modules.ts_classification.model_list import MODELS
 from ...common.batch_sampler import TSBatchSampler
 from ...common.reader import ReadTS
-from ..common import (
-    TSCutOff,
-    BuildTSDataset,
-    TSNormalize,
-    TimeFeature,
-    TStoArray,
-    TStoBatch,
-    StaticInfer,
-)
-
-from .processors import GetCls, BuildPadMask
-from ..base import BasicPredictor
+from ..base import BasePredictor
+from ..common import BuildTSDataset, TSCutOff, TSNormalize, TStoArray, TStoBatch
+from .processors import BuildPadMask, GetCls
 from .result import TSClsResult
 
 
-class TSClsPredictor(BasicPredictor):
-    """TSClsPredictor that inherits from BasicPredictor."""
+class TSClsPredictor(BasePredictor):
+    """TSClsPredictor that inherits from BasePredictor."""
 
     entities = MODELS
 
@@ -90,11 +81,7 @@ class TSClsPredictor(BasicPredictor):
         preprocessors["BuildPadMask"] = BuildPadMask(self.config["input_data"])
         preprocessors["TStoArray"] = TStoArray(self.config["input_data"])
         preprocessors["TStoBatch"] = TStoBatch()
-        infer = StaticInfer(
-            model_dir=self.model_dir,
-            model_prefix=self.MODEL_FILE_PREFIX,
-            option=self.pp_option,
-        )
+        infer = self.create_static_infer()
         postprocessors = {}
         postprocessors["GetCls"] = GetCls()
         return preprocessors, infer, postprocessors
@@ -109,7 +96,7 @@ class TSClsPredictor(BasicPredictor):
         Returns:
             Dict[str, Any]: A dictionary containing the paths to the input data, the raw input time series, and the classification results.
         """
-        batch_raw_ts = self.preprocessors["ReadTS"](ts_list=batch_data)
+        batch_raw_ts = self.preprocessors["ReadTS"](ts_list=batch_data.instances)
         batch_raw_ts_ori = copy.deepcopy(batch_raw_ts)
 
         if "TSNormalize" in self.preprocessors:
@@ -127,9 +114,9 @@ class TSClsPredictor(BasicPredictor):
         batch_ts_preds = self.postprocessors["GetCls"](pred_list=batch_preds)
 
         return {
-            "input_path": batch_data,
+            "input_path": batch_data.input_paths,
             "input_ts": batch_raw_ts,
             "input_ts_data": batch_raw_ts_ori,
             "classification": batch_ts_preds,
-            "target_cols": [self.config["info_params"]["target_cols"]]
+            "target_cols": [self.config["info_params"]["target_cols"]],
         }

paddlex/inference/models/ts_classification/processors.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,11 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from typing import Any, Dict, List
+
 import numpy as np
 import pandas as pd
-from typing import List, Any, Dict
+
+from ...utils.benchmark import benchmark
 
 
+@benchmark.timeit
 class GetCls:
     """A class to process prediction outputs and return class IDs and scores."""
 
@@ -56,6 +60,7 @@ class GetCls:
         return result
 
 
+@benchmark.timeit
 class BuildPadMask:
     """A class to build padding masks for time series data."""
 

paddlex/inference/models/ts_classification/result.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,15 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any
 import io
-import pandas as pd
-import matplotlib.pyplot as plt
+from typing import Any
+
 from PIL import Image
 
+from ....utils.deps import function_requires_deps, is_dep_available
 from ...common.result import BaseTSResult
 
+if is_dep_available("matplotlib"):
+    import matplotlib.pyplot as plt
 
+
+@function_requires_deps("matplotlib")
 def visualize(predicted_label, input_ts, target_cols):
     """
     Visualize time series data and its prediction results.
@@ -34,16 +38,21 @@ def visualize(predicted_label, input_ts, target_cols):
     """
     # 设置图形大小
     plt.figure(figsize=(12, 6))
-    input_ts_columns = input_ts.columns
+    input_ts.columns
     input_ts.index = input_ts.index.astype(str)
     length = len(input_ts)
-    value = predicted_label.loc[0, 'classid']
-    plt.plot(input_ts.index, input_ts[target_cols[0]], label=f'Predicted classid: {value}', color='blue')
+    value = predicted_label.loc[0, "classid"]
+    plt.plot(
+        input_ts.index,
+        input_ts[target_cols[0]],
+        label=f"Predicted classid: {value}",
+        color="blue",
+    )
 
     # 设置图形标题和标签
-    plt.title('Time Series input_ts with Predicted Labels')
-    plt.xlabel('Time')
-    plt.ylabel('Value')
+    plt.title("Time Series input_ts with Predicted Labels")
+    plt.xlabel("Time")
+    plt.ylabel("Value")
     plt.legend()
     plt.grid(True)
     plt.xticks(ticks=range(0, length, 10))
@@ -51,7 +60,7 @@ def visualize(predicted_label, input_ts, target_cols):
 
     # 保存图像到内存
     buf = io.BytesIO()
-    plt.savefig(buf, bbox_inches='tight')
+    plt.savefig(buf, bbox_inches="tight")
     buf.seek(0)
     plt.close()
     image = Image.open(buf)
@@ -67,7 +76,7 @@ class TSClsResult(BaseTSResult):
         classification = self["classification"]
         ts_input = self["input_ts_data"]
         return {"res": visualize(classification, ts_input, self["target_cols"])}
-    
+
     def _to_csv(self) -> Any:
         """
         Converts the classification results to a CSV format.

paddlex/inference/models/ts_forecasting/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

paddlex/inference/models/ts_forecasting/predictor.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,31 +12,30 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any, Union, Dict, List, Tuple
-import numpy as np
-import pandas as pd
-import os
 import copy
+import os
+from typing import Any, Dict, List, Tuple, Union
+
+import pandas as pd
 
 from ....modules.ts_forecast.model_list import MODELS
 from ...common.batch_sampler import TSBatchSampler
 from ...common.reader import ReadTS
+from ..base import BasePredictor
 from ..common import (
-    TSCutOff,
     BuildTSDataset,
-    TSNormalize,
     TimeFeature,
+    TSCutOff,
+    TSNormalize,
     TStoArray,
     TStoBatch,
-    StaticInfer,
 )
 from .processors import ArraytoTS, TSDeNormalize
-from ..base import BasicPredictor
 from .result import TSFcResult
 
 
-class TSFcPredictor(BasicPredictor):
-    """TSFcPredictor that inherits from BasicPredictor."""
+class TSFcPredictor(BasePredictor):
+    """TSFcPredictor that inherits from BasePredictor."""
 
     entities = MODELS
 
@@ -95,11 +94,7 @@ class TSFcPredictor(BasicPredictor):
             )
         preprocessors["TStoArray"] = TStoArray(self.config["input_data"])
         preprocessors["TStoBatch"] = TStoBatch()
-        infer = StaticInfer(
-            model_dir=self.model_dir,
-            model_prefix=self.MODEL_FILE_PREFIX,
-            option=self.pp_option,
-        )
+        infer = self.create_static_infer()
         postprocessors = {}
         postprocessors["ArraytoTS"] = ArraytoTS(self.config["info_params"])
         if self.config.get("scale", None):
@@ -122,7 +117,7 @@ class TSFcPredictor(BasicPredictor):
             dict: A dictionary containing the input path, raw image, class IDs, scores, and label names for every instance of the batch. Keys include 'input_path', 'input_img', 'class_ids', 'scores', and 'label_names'.
         """
 
-        batch_raw_ts = self.preprocessors["ReadTS"](ts_list=batch_data)
+        batch_raw_ts = self.preprocessors["ReadTS"](ts_list=batch_data.instances)
         batch_raw_ts_ori = copy.deepcopy(batch_raw_ts)
         batch_cutoff_ts = self.preprocessors["TSCutOff"](ts_list=batch_raw_ts)
 
@@ -152,7 +147,7 @@ class TSFcPredictor(BasicPredictor):
             )
 
         return {
-            "input_path": batch_data,
+            "input_path": batch_data.input_paths,
             "input_ts": batch_raw_ts,
             "cutoff_ts": batch_raw_ts_ori,
             "forecast": batch_ts_preds,

paddlex/inference/models/ts_forecasting/processors.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,12 +12,20 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import List, Dict, Any, Union
-import joblib
+from typing import Any, Dict, List
+
 import numpy as np
 import pandas as pd
 
+from ....utils.deps import class_requires_deps, is_dep_available
+from ...utils.benchmark import benchmark
+
+if is_dep_available("joblib"):
+    import joblib
+
 
+@benchmark.timeit
+@class_requires_deps("joblib")
 class TSDeNormalize:
     """A class to de-normalize time series prediction data using a pre-fitted scaler."""
 
@@ -60,6 +68,7 @@ class TSDeNormalize:
         return pred
 
 
+@benchmark.timeit
 class ArraytoTS:
     """A class to convert arrays of predictions into time series format."""
 

paddlex/inference/models/ts_forecasting/result.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,15 +12,20 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any
 import io
+from typing import Any
+
 import pandas as pd
-import matplotlib.pyplot as plt
 from PIL import Image
 
+from ....utils.deps import function_requires_deps, is_dep_available
 from ...common.result import BaseTSResult
 
+if is_dep_available("matplotlib"):
+    import matplotlib.pyplot as plt
 
+
+@function_requires_deps("matplotlib")
 def visualize(forecast: pd.DataFrame, actual_data: pd.DataFrame) -> Image.Image:
     """
     Visualizes both the time series forecast and actual results, returning them as a Pillow image.
@@ -43,20 +48,28 @@ def visualize(forecast: pd.DataFrame, actual_data: pd.DataFrame) -> Image.Image:
 
     length = min(len(forecast), len(actual_data))
     actual_data = actual_data.tail(length)
-    
-    plt.plot(actual_data.index, actual_data[forecast_columns[0]], label='Actual Data', color='blue', linestyle='--')
-    plt.plot(forecast.index, forecast[forecast_columns[0]], label='Forecast', color='red')
 
-    plt.title('Time Series Forecast')
-    plt.xlabel('Time')
+    plt.plot(
+        actual_data.index,
+        actual_data[forecast_columns[0]],
+        label="Actual Data",
+        color="blue",
+        linestyle="--",
+    )
+    plt.plot(
+        forecast.index, forecast[forecast_columns[0]], label="Forecast", color="red"
+    )
+
+    plt.title("Time Series Forecast")
+    plt.xlabel("Time")
     plt.ylabel(forecast_columns[0])
     plt.legend()
     plt.grid(True)
-    plt.xticks(ticks=range(0, 2*length, 10))
+    plt.xticks(ticks=range(0, 2 * length, 10))
     plt.xticks(rotation=45)
 
     buf = io.BytesIO()
-    plt.savefig(buf, bbox_inches='tight')
+    plt.savefig(buf, bbox_inches="tight")
     buf.seek(0)
     plt.close()
     image = Image.open(buf)
@@ -72,7 +85,7 @@ class TSFcResult(BaseTSResult):
         forecast = self["forecast"]
         ts_input = self["cutoff_ts"]
         return {"res": visualize(forecast, ts_input)}
-    
+
     def _to_csv(self) -> Any:
         """
         Converts the forecasting results to a CSV format.

paddlex/inference/models/video_classification/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

paddlex/inference/models/video_classification/predictor.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,27 +12,25 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any, Union, Dict, List, Tuple
-from ....utils.func_register import FuncRegister
+from typing import Union
+
 from ....modules.video_classification.model_list import MODELS
+from ....utils.func_register import FuncRegister
 from ...common.batch_sampler import VideoBatchSampler
 from ...common.reader import ReadVideo
-from ..common import (
-    StaticInfer,
-)
-from ..base import BasicPredictor
+from ..base import BasePredictor
 from .processors import (
-    Scale,
     CenterCrop,
     Image2Array,
     NormalizeVideo,
-    VideoClasTopk,
+    Scale,
     ToBatch,
+    VideoClasTopk,
 )
 from .result import TopkVideoResult
 
 
-class VideoClasPredictor(BasicPredictor):
+class VideoClasPredictor(BasePredictor):
 
     entities = MODELS
 
@@ -62,11 +60,7 @@ class VideoClasPredictor(BasicPredictor):
                 pre_tfs[name] = op
         pre_tfs["ToBatch"] = ToBatch()
 
-        infer = StaticInfer(
-            model_dir=self.model_dir,
-            model_prefix=self.MODEL_FILE_PREFIX,
-            option=self.pp_option,
-        )
+        infer = self.create_static_infer()
 
         post_op = {}
         for key in self.config["PostProcess"]:

paddlex/inference/models/video_classification/processors.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -13,19 +13,19 @@
 # limitations under the License.
 
 
-import os
-import os.path as osp
-from typing import List, Sequence, Union, Optional, Tuple
+from typing import List, Optional, Sequence, Tuple, Union
 
-import re
 import numpy as np
-import cv2
-import math
-import json
-import tempfile
-import lazy_paddle
 
+from ....utils.deps import class_requires_deps, is_dep_available
+from ...utils.benchmark import benchmark
 
+if is_dep_available("opencv-contrib-python"):
+    import cv2
+
+
+@benchmark.timeit
+@class_requires_deps("opencv-contrib-python")
 class Scale:
     """Scale images."""
 
@@ -135,6 +135,7 @@ class Scale:
         return [self.scale(video) for video in videos]
 
 
+@benchmark.timeit
 class CenterCrop:
     """Center crop images."""
 
@@ -163,22 +164,16 @@ class CenterCrop:
 
         crop_imgs = []
         th, tw = self.target_size, self.target_size
-        if isinstance(imgs, lazy_paddle.Tensor):
-            h, w = imgs.shape[-2:]
+        for img in imgs:
+            h, w, _ = img.shape
+            assert (w >= self.target_size) and (
+                h >= self.target_size
+            ), "image width({}) and height({}) should be larger than crop size".format(
+                w, h, self.target_size
+            )
             x1 = int(round((w - tw) / 2.0)) if self.do_round else (w - tw) // 2
             y1 = int(round((h - th) / 2.0)) if self.do_round else (h - th) // 2
-            crop_imgs = imgs[:, :, y1 : y1 + th, x1 : x1 + tw]
-        else:
-            for img in imgs:
-                h, w, _ = img.shape
-                assert (w >= self.target_size) and (
-                    h >= self.target_size
-                ), "image width({}) and height({}) should be larger than crop size".format(
-                    w, h, self.target_size
-                )
-                x1 = int(round((w - tw) / 2.0)) if self.do_round else (w - tw) // 2
-                y1 = int(round((h - th) / 2.0)) if self.do_round else (h - th) // 2
-                crop_imgs.append(img[y1 : y1 + th, x1 : x1 + tw])
+            crop_imgs.append(img[y1 : y1 + th, x1 : x1 + tw])
         return crop_imgs
 
     def __call__(self, videos: List[np.ndarray]) -> List[np.ndarray]:
@@ -194,6 +189,7 @@ class CenterCrop:
         return [self.center_crop(video) for video in videos]
 
 
+@benchmark.timeit
 class Image2Array:
     """Convert a sequence of images to a numpy array with optional transposition."""
 
@@ -247,6 +243,8 @@ class Image2Array:
         return [self.img2array(video) for video in videos]
 
 
+@benchmark.timeit
+@class_requires_deps("opencv-contrib-python")
 class NormalizeVideo:
     """
     Normalize video frames by subtracting the mean and dividing by the standard deviation.
@@ -324,6 +322,7 @@ class NormalizeVideo:
         return [self.normalize_video(video) for video in videos]
 
 
+@benchmark.timeit
 class VideoClasTopk:
     """Applies a top-k transformation on video classification predictions."""
 
@@ -394,6 +393,7 @@ class VideoClasTopk:
         return indexes, scores, label_names
 
 
+@benchmark.timeit
 class ToBatch:
     """A class for batching videos."""
 

paddlex/inference/models/video_classification/result.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -12,17 +12,21 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import cv2
 import numpy as np
 import PIL
 from PIL import Image, ImageDraw, ImageFont
 
+from ....utils.deps import class_requires_deps, is_dep_available
 from ....utils.fonts import PINGFANG_FONT_FILE_PATH
+from ...common.result import BaseVideoResult
 from ...utils.color_map import get_colormap
 from ...utils.io import VideoReader
-from ...common.result import BaseVideoResult
+
+if is_dep_available("opencv-contrib-python"):
+    import cv2
 
 
+@class_requires_deps("opencv-contrib-python")
 class TopkVideoResult(BaseVideoResult):
 
     def _to_video(self):

paddlex/inference/models/video_detection/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.