remove some deprecated files causing import errors

mmocr/datasets/__init__.py

@@ -1,23 +1,24 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 from . import utils
 from .base_dataset import BaseDataset
-from .builder import DATASETS, LOADERS, PARSERS, TRANSFORMS
+from .builder import LOADERS, PARSERS, TRANSFORMS
 from .icdar_dataset import IcdarDataset
+# TODO: check this
 from .kie_dataset import KIEDataset
 from .ner_dataset import NerDataset
 from .ocr_dataset import OCRDataset
 from .ocr_seg_dataset import OCRSegDataset
 from .openset_kie_dataset import OpensetKIEDataset
-from .pipelines import CustomFormatBundle, DBNetTargets, FCENetTargets
+from .pipelines import DBNetTargets, FCENetTargets
 from .text_det_dataset import TextDetDataset
 from .uniform_concat_dataset import UniformConcatDataset
 from .utils import *  # NOQA
 
 __all__ = [
-    'DATASETS', 'IcdarDataset', 'BaseDataset', 'OCRDataset', 'TextDetDataset',
+    'IcdarDataset', 'BaseDataset', 'OCRDataset', 'TextDetDataset',
-    'CustomFormatBundle', 'DBNetTargets', 'OCRSegDataset', 'KIEDataset',
+    'KIEDataset', 'DBNetTargets', 'OCRSegDataset', 'FCENetTargets',
-    'FCENetTargets', 'NerDataset', 'UniformConcatDataset', 'OpensetKIEDataset',
+    'NerDataset', 'UniformConcatDataset', 'OpensetKIEDataset', 'TRANSFORMS',
-    'TRANSFORMS', 'PARSERS', 'LOADERS'
+    'PARSERS', 'LOADERS'
 ]
 
 __all__ += utils.__all__

mmocr/datasets/pipelines/__init__.py

@@ -1,8 +1,7 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 from .box_utils import sort_vertex, sort_vertex8
-from .custom_format_bundle import CustomFormatBundle
 from .dbnet_transforms import EastRandomCrop, ImgAug
-from .kie_transforms import KIEFormatBundle, ResizeNoImg
+from .kie_transforms import ResizeNoImg
 from .loading import (LoadImageFromLMDB, LoadImageFromNdarray,
                       LoadTextAnnotations)
 from .ner_transforms import NerTransform, ToTensorNER
@@ -21,14 +20,14 @@ from .transforms import (ColorJitter, PyramidRescale, RandomCropFlip,
 
 __all__ = [
     'LoadTextAnnotations', 'NormalizeOCR', 'OnlineCropOCR', 'ResizeOCR',
-    'ToTensorOCR', 'CustomFormatBundle', 'DBNetTargets', 'PANetTargets',
+    'ToTensorOCR', 'DBNetTargets', 'PANetTargets', 'ColorJitter',
-    'ColorJitter', 'RandomCropInstances', 'RandomRotateTextDet',
+    'RandomCropInstances', 'RandomRotateTextDet', 'ScaleAspectJitter',
-    'ScaleAspectJitter', 'MultiRotateAugOCR', 'OCRSegTargets', 'FancyPCA',
+    'MultiRotateAugOCR', 'OCRSegTargets', 'FancyPCA',
     'RandomCropPolyInstances', 'RandomRotatePolyInstances', 'RandomPaddingOCR',
     'ImgAug', 'EastRandomCrop', 'RandomRotateImageBox', 'OpencvToPil',
-    'PilToOpencv', 'KIEFormatBundle', 'SquareResizePad', 'TextSnakeTargets',
+    'PilToOpencv', 'SquareResizePad', 'TextSnakeTargets', 'sort_vertex',
-    'sort_vertex', 'LoadImageFromNdarray', 'sort_vertex8', 'FCENetTargets',
+    'LoadImageFromNdarray', 'sort_vertex8', 'FCENetTargets', 'RandomScaling',
-    'RandomScaling', 'RandomCropFlip', 'NerTransform', 'ToTensorNER',
+    'RandomCropFlip', 'NerTransform', 'ToTensorNER', 'ResizeNoImg',
-    'ResizeNoImg', 'PyramidRescale', 'OneOfWrapper', 'RandomWrapper',
+    'PyramidRescale', 'OneOfWrapper', 'RandomWrapper', 'TorchVisionWrapper',
-    'TorchVisionWrapper', 'LoadImageFromLMDB'
+    'LoadImageFromLMDB'
 ]

mmocr/datasets/pipelines/custom_format_bundle.py

@@ -1,66 +0,0 @@
-# Copyright (c) OpenMMLab. All rights reserved.
-import numpy as np
-from mmcv.parallel import DataContainer as DC
-from mmdet.datasets.pipelines.formatting import DefaultFormatBundle
-
-from mmocr.core.visualize import overlay_mask_img, show_feature
-from mmocr.registry import TRANSFORMS
-
-
-@TRANSFORMS.register_module()
-class CustomFormatBundle(DefaultFormatBundle):
-    """Custom formatting bundle.
-
-    It formats common fields such as 'img' and 'proposals' as done in
-    DefaultFormatBundle, while other fields such as 'gt_kernels' and
-    'gt_effective_region_mask' will be formatted to DC as follows:
-
-    - gt_kernels: to DataContainer (cpu_only=True)
-    - gt_effective_mask: to DataContainer (cpu_only=True)
-
-    Args:
-        keys (list[str]): Fields to be formatted to DC only.
-        call_super (bool): If True, format common fields
-            by DefaultFormatBundle, else format fields in keys above only.
-        visualize (dict): If flag=True, visualize gt mask for debugging.
-    """
-
-    def __init__(self,
-                 keys=[],
-                 call_super=True,
-                 visualize=dict(flag=False, boundary_key=None)):
-
-        super().__init__()
-        self.visualize = visualize
-        self.keys = keys
-        self.call_super = call_super
-
-    def __call__(self, results):
-
-        if self.visualize['flag']:
-            img = results['img'].astype(np.uint8)
-            boundary_key = self.visualize['boundary_key']
-            if boundary_key is not None:
-                img = overlay_mask_img(img, results[boundary_key].masks[0])
-
-            features = [img]
-            names = ['img']
-            to_uint8 = [1]
-
-            for k in results['mask_fields']:
-                for iter in range(len(results[k].masks)):
-                    features.append(results[k].masks[iter])
-                    names.append(k + str(iter))
-                    to_uint8.append(0)
-            show_feature(features, names, to_uint8)
-
-        if self.call_super:
-            results = super().__call__(results)
-
-        for k in self.keys:
-            results[k] = DC(results[k], cpu_only=True)
-
-        return results
-
-    def __repr__(self):
-        return self.__class__.__name__

mmocr/datasets/pipelines/kie_transforms.py

@@ -1,8 +1,6 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import numpy as np
 from mmcv import rescale_size
-from mmcv.parallel import DataContainer as DC
-from mmdet.datasets.pipelines.formatting import DefaultFormatBundle, to_tensor
 
 from mmocr.registry import TRANSFORMS
 
@@ -40,52 +38,53 @@ class ResizeNoImg:
         return results
 
 
-@TRANSFORMS.register_module()
+# @TRANSFORMS.register_module()
-class KIEFormatBundle(DefaultFormatBundle):
+# class KIEFormatBundle(DefaultFormatBundle):
-    """Key information extraction formatting bundle.
+#     """Key information extraction formatting bundle.
 
-    Based on the DefaultFormatBundle, itt simplifies the pipeline of formatting
+#     Based on the DefaultFormatBundle, itt simplifies the pipeline of
-    common fields, including "img", "proposals", "gt_bboxes", "gt_labels",
+#     formatting
-    "gt_masks", "gt_semantic_seg", "relations" and "texts".
+#     common fields, including "img", "proposals", "gt_bboxes", "gt_labels",
-    These fields are formatted as follows.
+#     "gt_masks", "gt_semantic_seg", "relations" and "texts".
+#     These fields are formatted as follows.
 
-    - img: (1) transpose, (2) to tensor, (3) to DataContainer (stack=True)
+#     - img: (1) transpose, (2) to tensor, (3) to DataContainer (stack=True)
-    - proposals: (1) to tensor, (2) to DataContainer
+#     - proposals: (1) to tensor, (2) to DataContainer
-    - gt_bboxes: (1) to tensor, (2) to DataContainer
+#     - gt_bboxes: (1) to tensor, (2) to DataContainer
-    - gt_bboxes_ignore: (1) to tensor, (2) to DataContainer
+#     - gt_bboxes_ignore: (1) to tensor, (2) to DataContainer
-    - gt_labels: (1) to tensor, (2) to DataContainer
+#     - gt_labels: (1) to tensor, (2) to DataContainer
-    - gt_masks: (1) to tensor, (2) to DataContainer (cpu_only=True)
+#     - gt_masks: (1) to tensor, (2) to DataContainer (cpu_only=True)
-    - gt_semantic_seg: (1) unsqueeze dim-0 (2) to tensor,
+#     - gt_semantic_seg: (1) unsqueeze dim-0 (2) to tensor,
-                       (3) to DataContainer (stack=True)
+#                        (3) to DataContainer (stack=True)
-    - relations: (1) scale, (2) to tensor, (3) to DataContainer
+#     - relations: (1) scale, (2) to tensor, (3) to DataContainer
-    - texts: (1) to tensor, (2) to DataContainer
+#     - texts: (1) to tensor, (2) to DataContainer
-    """
+#     """
 
-    def __call__(self, results):
+#     def __call__(self, results):
-        """Call function to transform and format common fields in results.
+#         """Call function to transform and format common fields in results.
 
-        Args:
+#         Args:
-            results (dict): Result dict contains the data to convert.
+#             results (dict): Result dict contains the data to convert.
 
-        Returns:
+#         Returns:
-            dict: The result dict contains the data that is formatted with
+#             dict: The result dict contains the data that is formatted with
-                default bundle.
+#                 default bundle.
-        """
+#         """
-        super().__call__(results)
+#         super().__call__(results)
-        if 'ann_info' in results:
+#         if 'ann_info' in results:
-            for key in ['relations', 'texts']:
+#             for key in ['relations', 'texts']:
-                value = results['ann_info'][key]
+#                 value = results['ann_info'][key]
-                if key == 'relations' and 'scale_factor' in results:
+#                 if key == 'relations' and 'scale_factor' in results:
-                    scale_factor = results['scale_factor']
+#                     scale_factor = results['scale_factor']
-                    if isinstance(scale_factor, float):
+#                     if isinstance(scale_factor, float):
-                        sx = sy = scale_factor
+#                         sx = sy = scale_factor
-                    else:
+#                     else:
-                        sx, sy = results['scale_factor'][:2]
+#                         sx, sy = results['scale_factor'][:2]
-                    r = sx / sy
+#                     r = sx / sy
-                    factor = np.array([sx, sy, r, 1, r]).astype(np.float32)
+#                     factor = np.array([sx, sy, r, 1, r]).astype(np.float32)
-                    value = value * factor[None, None]
+#                     value = value * factor[None, None]
-                results[key] = DC(to_tensor(value))
+#                 results[key] = DC(to_tensor(value))
-        return results
+#         return results
 
-    def __repr__(self):
+#     def __repr__(self):
-        return self.__class__.__name__
+#         return self.__class__.__name__