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

paddlex/inference/common/reader/image_reader.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.
@@ -13,20 +13,20 @@
 # limitations under the License.
 
 import numpy as np
-import cv2
 
-from ...utils.io import ImageReader, PDFReader
+from ....utils.deps import class_requires_deps, is_dep_available
+from ...utils.benchmark import benchmark
+from ...utils.io import ImageReader
 
+if is_dep_available("opencv-contrib-python"):
+    import cv2
 
+
+@benchmark.timeit_with_options(name=None, is_read_operation=True)
+@class_requires_deps("opencv-contrib-python")
 class ReadImage:
     """Load image from the file."""
 
-    _FLAGS_DICT = {
-        "BGR": cv2.IMREAD_COLOR,
-        "RGB": cv2.IMREAD_COLOR,
-        "GRAY": cv2.IMREAD_GRAYSCALE,
-    }
-
     def __init__(self, format="BGR"):
         """
         Initialize the instance.
@@ -37,7 +37,11 @@ class ReadImage:
         """
         super().__init__()
         self.format = format
-        flags = self._FLAGS_DICT[self.format]
+        flags = {
+            "BGR": cv2.IMREAD_COLOR,
+            "RGB": cv2.IMREAD_COLOR,
+            "GRAY": cv2.IMREAD_GRAYSCALE,
+        }[self.format]
         self._img_reader = ImageReader(backend="opencv", flags=flags)
 
     def __call__(self, imgs):
@@ -47,7 +51,7 @@ class ReadImage:
     def read(self, img):
         if isinstance(img, np.ndarray):
             if self.format == "RGB":
-                img = img[:, :, ::-1]
+                img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
             return img
         elif isinstance(img, str):
             blob = self._img_reader.read(img)
@@ -58,7 +62,7 @@ class ReadImage:
                 if blob.ndim != 3:
                     raise RuntimeError("Array is not 3-dimensional.")
                 # BGR to RGB
-                blob = blob[..., ::-1]
+                blob = cv2.cvtColor(blob, cv2.COLOR_BGR2RGB)
             return blob
         else:
             raise TypeError(

paddlex/inference/common/reader/ts_reader.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 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import numpy as np
 import pandas as pd
 
+from ...utils.benchmark import benchmark
 from ...utils.io import CSVReader
 
 
+@benchmark.timeit_with_options(name=None, is_read_operation=True)
 class ReadTS:
 
     def __init__(self):

paddlex/inference/common/reader/video_reader.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 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import numpy as np
-import cv2
 
+from ...utils.benchmark import benchmark
 from ...utils.io import VideoReader
 
 
+@benchmark.timeit_with_options(name=None, is_read_operation=True)
 class ReadVideo:
     """Load video from the file."""
 

paddlex/inference/common/result/__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.
@@ -12,18 +12,18 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from .base_result import BaseResult
 from .base_cv_result import BaseCVResult
+from .base_result import BaseResult
 from .base_ts_result import BaseTSResult
 from .base_video_result import BaseVideoResult
 from .mixin import (
-    StrMixin,
-    JsonMixin,
     Base64Mixin,
-    ImgMixin,
     CSVMixin,
     HtmlMixin,
-    XlsxMixin,
-    VideoMixin,
+    ImgMixin,
+    JsonMixin,
     MarkdownMixin,
+    StrMixin,
+    VideoMixin,
+    XlsxMixin,
 )

paddlex/inference/common/result/base_cv_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.
@@ -12,9 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from pathlib import Path
+
 from .base_result import BaseResult
 from .mixin import ImgMixin
-from ...utils.io import ImageWriter
 
 
 class BaseCVResult(BaseResult, ImgMixin):
@@ -29,3 +30,12 @@ class BaseCVResult(BaseResult, ImgMixin):
         """
         super().__init__(data)
         ImgMixin.__init__(self, "pillow")
+
+    def _get_input_fn(self):
+        fn = super()._get_input_fn()
+        if (page_idx := self.get("page_index", None)) is not None:
+            fp = Path(fn)
+            stem, suffix = fp.stem, fp.suffix
+            return f"{stem}_{page_idx}{suffix}"
+        else:
+            return fn

paddlex/inference/common/result/base_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.
@@ -13,13 +13,14 @@
 # limitations under the License.
 
 import inspect
-from pathlib import Path
-import time
 import random
+import time
+from pathlib import Path
+
 import numpy as np
 
 from ....utils import logging
-from .mixin import StrMixin, JsonMixin, ImgMixin
+from .mixin import JsonMixin, StrMixin
 
 
 class BaseResult(dict, JsonMixin, StrMixin):
@@ -55,7 +56,7 @@ class BaseResult(dict, JsonMixin, StrMixin):
                 func()
 
     def _get_input_fn(self):
-        if (fp := self["input_path"]) is None:
+        if self.get("input_path", None) is None:
             if self._rand_fn:
                 return self._rand_fn
 
@@ -67,4 +68,5 @@ class BaseResult(dict, JsonMixin, StrMixin):
             )
             self._rand_fn = Path(fp).name
             return self._rand_fn
+        fp = self["input_path"]
         return Path(fp).name

paddlex/inference/common/result/base_ts_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.
@@ -14,7 +14,6 @@
 
 from .base_result import BaseResult
 from .mixin import CSVMixin, ImgMixin
-from ...utils.io import CSVWriter
 
 
 class BaseTSResult(BaseResult, CSVMixin, ImgMixin):

paddlex/inference/common/result/base_video_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.
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 from .base_result import BaseResult
-from .mixin import ImgMixin, VideoMixin
+from .mixin import VideoMixin
 
 
 class BaseVideoResult(BaseResult, VideoMixin):

paddlex/inference/common/result/mixin.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.
@@ -12,29 +12,28 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Union, Tuple, List, Dict, Any, Iterator, Callable, Optional
+import copy
+import json
+import mimetypes
+import os
 from abc import abstractmethod
 from pathlib import Path
-import os
-import mimetypes
-import json
-import copy
-import re
+from typing import Any, Callable, Dict, List, Optional, Union
+
 import numpy as np
-from PIL import Image
 import pandas as pd
+from PIL import Image
 
 from ....utils import logging
 from ...utils.io import (
-    JsonWriter,
-    ImageReader,
-    ImageWriter,
     CSVWriter,
     HtmlWriter,
-    XlsxWriter,
+    ImageWriter,
+    JsonWriter,
+    MarkdownWriter,
     TextWriter,
     VideoWriter,
-    MarkdownWriter,
+    XlsxWriter,
 )
 
 
@@ -162,7 +161,7 @@ class JsonMixin:
         else:
             if len(json_data) > 1:
                 logging.warning(
-                    f"The result has multiple json files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple json files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._json_writer.write(
                 save_path,
@@ -265,7 +264,7 @@ class Base64Mixin:
         else:
             if len(base64) > 1:
                 logging.warning(
-                    f"The result has multiple base64 files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple base64 files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._base64_writer.write(
                 save_path, base64[list(base64.keys())[0]], *args, **kwargs
@@ -329,7 +328,7 @@ class ImgMixin:
         else:
             if len(img) > 1:
                 logging.warning(
-                    f"The result has multiple img files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple img files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._img_writer.write(save_path, img[list(img.keys())[0]], *args, **kwargs)
 
@@ -393,7 +392,7 @@ class CSVMixin:
         else:
             if len(csv) > 1:
                 logging.warning(
-                    f"The result has multiple csv files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple csv files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._csv_writer.write(save_path, csv[list(csv.keys())[0]], *args, **kwargs)
 
@@ -456,7 +455,7 @@ class HtmlMixin:
         else:
             if len(html) > 1:
                 logging.warning(
-                    f"The result has multiple html files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple html files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._html_writer.write(
                 save_path, html[list(html.keys())[0]], *args, **kwargs
@@ -525,7 +524,7 @@ class XlsxMixin:
         else:
             if len(xlsx) > 1:
                 logging.warning(
-                    f"The result has multiple xlsx files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple xlsx files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             self._xlsx_writer.write(
                 save_path, xlsx[list(xlsx.keys())[0]], *args, **kwargs
@@ -590,7 +589,7 @@ class VideoMixin:
         else:
             if len(video) > 1:
                 logging.warning(
-                    f"The result has multiple video files need to be saved. But the `save_path` has been specfied as `{save_path}`!"
+                    f"The result has multiple video files need to be saved. But the `save_path` has been specified as `{save_path}`!"
                 )
             video_writer.write(save_path, video[list(video.keys())[0]], *args, **kwargs)
 
@@ -610,10 +609,13 @@ class MarkdownMixin:
         self._save_funcs.append(self.save_to_markdown)
 
     @abstractmethod
-    def _to_markdown(self) -> Dict[str, Union[str, Dict[str, Any]]]:
+    def _to_markdown(self, pretty=True) -> Dict[str, Union[str, Dict[str, Any]]]:
         """
         Convert the result to markdown format.
 
+        Args:
+            pretty (Optional[bool]): whether to pretty markdown by HTML, default by True.
+
         Returns:
             Dict[str, Union[str, Dict[str, Any]]]: A dictionary containing markdown text and image data.
         """
@@ -628,7 +630,7 @@ class MarkdownMixin:
         """
         return self._to_markdown()
 
-    def save_to_markdown(self, save_path, *args, **kwargs) -> None:
+    def save_to_markdown(self, save_path, pretty=True, *args, **kwargs) -> None:
         """Save the markdown data to a file.
 
         Args:
@@ -666,7 +668,7 @@ class MarkdownMixin:
             self._markdown_writer.write,
             self._img_writer.write,
             self.save_path,
-            self._to_markdown(),
+            self._to_markdown(pretty=pretty),
             *args,
             **kwargs,
         )
@@ -699,5 +701,9 @@ class MarkdownMixin:
             if isinstance(value, dict):
                 base_save_path = save_path.parent
                 for img_path, img_data in value.items():
-                    save_path = base_save_path / img_path
-                    save_img_func(save_path.as_posix(), img_data, *args, **kwargs)
+                    save_img_func(
+                        (base_save_path / img_path).as_posix(),
+                        img_data,
+                        *args,
+                        **kwargs,
+                    )

paddlex/inference/models/__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.
@@ -13,70 +13,44 @@
 # limitations under the License.
 
 
-from pathlib import Path
-from typing import Any, Dict, Optional
 from importlib import import_module
+from pathlib import Path
+from typing import Any, Dict, Optional, Union
+
 from ...utils import errors
+from ..utils.hpi import HPIConfig
 from ..utils.official_models import official_models
-from .base import BasePredictor, BasicPredictor
-
-from .image_classification import ClasPredictor
-from .object_detection import DetPredictor
-from .keypoint_detection import KptPredictor
-from .text_detection import TextDetPredictor
-from .text_recognition import TextRecPredictor
-from .table_structure_recognition import TablePredictor
-from .formula_recognition import FormulaRecPredictor
-from .instance_segmentation import InstanceSegPredictor
-from .semantic_segmentation import SegPredictor
-from .image_feature import ImageFeaturePredictor
-from .ts_forecasting import TSFcPredictor
-from .ts_anomaly_detection import TSAdPredictor
-from .ts_classification import TSClsPredictor
-from .image_unwarping import WarpPredictor
-from .image_multilabel_classification import MLClasPredictor
-from .face_feature import FaceFeaturePredictor
-from .open_vocabulary_detection import OVDetPredictor
-from .open_vocabulary_segmentation import OVSegPredictor
-
 
 # from .table_recognition import TablePredictor
 # from .general_recognition import ShiTuRecPredictor
 from .anomaly_detection import UadPredictor
+from .base import BasePredictor
+from .doc_vlm import DocVLMPredictor
+from .face_feature import FaceFeaturePredictor
+from .formula_recognition import FormulaRecPredictor
+from .image_classification import ClasPredictor
+from .image_feature import ImageFeaturePredictor
+from .image_multilabel_classification import MLClasPredictor
+from .image_unwarping import WarpPredictor
+from .instance_segmentation import InstanceSegPredictor
+from .keypoint_detection import KptPredictor
+from .m_3d_bev_detection import BEVDet3DPredictor
 
 # from .face_recognition import FaceRecPredictor
 from .multilingual_speech_recognition import WhisperPredictor
+from .object_detection import DetPredictor
+from .open_vocabulary_detection import OVDetPredictor
+from .open_vocabulary_segmentation import OVSegPredictor
+from .semantic_segmentation import SegPredictor
+from .table_structure_recognition import TablePredictor
+from .text_detection import TextDetPredictor
+from .text_recognition import TextRecPredictor
+from .ts_anomaly_detection import TSAdPredictor
+from .ts_classification import TSClsPredictor
+from .ts_forecasting import TSFcPredictor
 from .video_classification import VideoClasPredictor
 from .video_detection import VideoDetPredictor
 
-module_3d_bev_detection = import_module(".3d_bev_detection", "paddlex.inference.models")
-BEVDet3DPredictor = getattr(module_3d_bev_detection, "BEVDet3DPredictor")
-
-
-def _create_hp_predictor(
-    model_name, model_dir, device, config, hpi_params, *args, **kwargs
-):
-    try:
-        from paddlex_hpi.models import HPPredictor
-    except ModuleNotFoundError:
-        raise RuntimeError(
-            "The PaddleX HPI plugin is not properly installed, and the high-performance model inference features are not available."
-        ) from None
-    try:
-        predictor = HPPredictor.get(model_name)(
-            model_dir=model_dir,
-            config=config,
-            device=device,
-            *args,
-            hpi_params=hpi_params,
-            **kwargs,
-        )
-    except errors.others.ClassNotFoundException:
-        raise ValueError(
-            f"{model_name} is not supported by the PaddleX HPI plugin."
-        ) from None
-    return predictor
-
 
 def create_predictor(
     model_name: str,
@@ -84,12 +58,15 @@ def create_predictor(
     device=None,
     pp_option=None,
     use_hpip: bool = False,
-    hpi_params: Optional[Dict[str, Any]] = None,
+    hpi_config: Optional[Union[Dict[str, Any], HPIConfig]] = None,
     *args,
     **kwargs,
 ) -> BasePredictor:
     if model_dir is None:
-        model_dir = check_model(model_name)
+        assert (
+            model_name in official_models
+        ), f"The model ({model_name}) is not supported! Please using directory of local model files or model name supported by PaddleX!"
+        model_dir = official_models[model_name]
     else:
         assert Path(model_dir).exists(), f"{model_dir} is not exists!"
         model_dir = Path(model_dir)
@@ -97,34 +74,13 @@ def create_predictor(
     assert (
         model_name == config["Global"]["model_name"]
     ), f"Model name mismatch，please input the correct model dir."
-
-    if use_hpip:
-        return _create_hp_predictor(
-            model_name=model_name,
-            model_dir=model_dir,
-            config=config,
-            hpi_params=hpi_params,
-            device=device,
-            *args,
-            **kwargs,
-        )
-    else:
-        return BasicPredictor.get(model_name)(
-            model_dir=model_dir,
-            config=config,
-            device=device,
-            pp_option=pp_option,
-            *args,
-            **kwargs,
-        )
-
-
-def check_model(model):
-    if Path(model).exists():
-        return Path(model)
-    elif model in official_models:
-        return official_models[model]
-    else:
-        raise Exception(
-            f"The model ({model}) is no exists! Please using directory of local model files or model name supported by PaddleX!"
-        )
+    return BasePredictor.get(model_name)(
+        model_dir=model_dir,
+        config=config,
+        device=device,
+        pp_option=pp_option,
+        use_hpip=use_hpip,
+        hpi_config=hpi_config,
+        *args,
+        **kwargs,
+    )

paddlex/inference/models/anomaly_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.

paddlex/inference/models/anomaly_detection/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.
@@ -12,28 +12,22 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any, Union, Dict, List, Tuple
+from typing import Any, Dict, List, Tuple, Union
+
 import numpy as np
 
-from ....utils.func_register import FuncRegister
 from ....modules.anomaly_detection.model_list import MODELS
+from ....utils.func_register import FuncRegister
 from ...common.batch_sampler import ImageBatchSampler
 from ...common.reader import ReadImage
-from ..common import (
-    Resize,
-    ResizeByShort,
-    Normalize,
-    ToCHWImage,
-    ToBatch,
-    StaticInfer,
-)
+from ..base import BasePredictor
+from ..common import Normalize, Resize, ToBatch, ToCHWImage
 from .processors import MapToMask
-from ..base import BasicPredictor
 from .result import UadResult
 
 
-class UadPredictor(BasicPredictor):
-    """UadPredictor that inherits from BasicPredictor."""
+class UadPredictor(BasePredictor):
+    """UadPredictor that inherits from BasePredictor."""
 
     entities = MODELS
 
@@ -82,11 +76,7 @@ class UadPredictor(BasicPredictor):
             preprocessors[name] = op
         preprocessors["ToBatch"] = ToBatch()
 
-        infer = StaticInfer(
-            model_dir=self.model_dir,
-            model_prefix=self.MODEL_FILE_PREFIX,
-            option=self.pp_option,
-        )
+        infer = self.create_static_infer()
         postprocessors = {"Map_to_mask": MapToMask()}
         return preprocessors, infer, postprocessors
 

paddlex/inference/models/anomaly_detection/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.
@@ -13,9 +13,16 @@
 # limitations under the License.
 
 import numpy as np
-from skimage import measure, morphology
 
+from ....utils.deps import class_requires_deps, is_dep_available
+from ...utils.benchmark import benchmark
 
+if is_dep_available("scikit-image"):
+    from skimage import morphology
+
+
+@benchmark.timeit
+@class_requires_deps("scikit-image")
 class MapToMask:
     """Map_to_mask"""
 

paddlex/inference/models/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.
@@ -13,10 +13,11 @@
 # limitations under the License.
 
 import copy
+
 import numpy as np
 from PIL import Image
 
-from ...common.result import BaseCVResult, StrMixin, JsonMixin
+from ...common.result import BaseCVResult, JsonMixin
 
 
 class UadResult(BaseCVResult):

paddlex/inference/models/base/__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.
@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from .predictor import BasePredictor, BasicPredictor
+from .predictor import BasePredictor