[fix] formatting

Author: chyomin06

compressai_vision/codecs/encdec_utils/png_yuv.py

@@ -89,9 +89,9 @@ def __call__(self, input: Dict, file_prefix: str):
             images_in_folder = len(list(parent.glob(ext)))
             nb_frames = input["last_frame"] - input["frame_skip"]
 
-            assert (
-                images_in_folder == nb_frames
-            ), f"input folder contains {images_in_folder} images, {nb_frames} were expected"
+            assert images_in_folder == nb_frames, (
+                f"input folder contains {images_in_folder} images, {nb_frames} were expected"
+            )
 
             input_info = [
                 "-pattern_type",
@@ -122,9 +122,9 @@ def __call__(self, input: Dict, file_prefix: str):
 
         # Use existing YUV (if found and indicated for use):
         if self.use_yuv:
-            assert (
-                yuv_file is not None
-            ), "Parameter 'use_yuv' set True but YUV file not found."
+            assert yuv_file is not None, (
+                "Parameter 'use_yuv' set True but YUV file not found."
+            )
             size = yuv_file.stat().st_size
             bytes_per_luma_sample = {"yuv420p": 1.5}[chroma_format]
             bytes_per_sample = (input_bitdepth + 7) >> 3
@@ -135,9 +135,9 @@ def __call__(self, input: Dict, file_prefix: str):
                 * bytes_per_sample
                 * nb_frames
             )
-            assert (
-                size == expected_size
-            ), f"YUV found for input but expected size of {expected_size} bytes differs from actual size of {size} bytes"
+            assert size == expected_size, (
+                f"YUV found for input but expected size of {expected_size} bytes differs from actual size of {size} bytes"
+            )
             shutil.copy(yuv_file, yuv_in_path)
             print(f"Using pre-existing YUV file: {yuv_file}")
             return (yuv_in_path, nb_frames, frame_width, frame_height, file_prefix)
@@ -204,9 +204,9 @@ def __call__(
         frame_width = video_info["width"]
         frame_height = video_info["height"]
 
-        assert (
-            "420" in video_info["format"].value
-        ), f"Only support yuv420, but got {video_info['format']}"
+        assert "420" in video_info["format"].value, (
+            f"Only support yuv420, but got {video_info['format']}"
+        )
         pix_fmt_suffix = "10le" if video_info["bitdepth"] == 10 else ""
         chroma_format = "yuv420p"
 

compressai_vision/codecs/sic_sfu2022.py

@@ -151,13 +151,13 @@ def __init__(self, device: str, **kwargs):
         # root_url = "https://dspub.blob.core.windows.net/compressai/sic_sfu2022"
 
         self.target_tlayer = int(kwargs["target_task_layer"])
-        assert (
-            self.num_tasks == 2 or self.num_tasks == 3
-        ), f"SIC_SFU2023 supports only 2 or 3 task layers, but got {self.num_tasks}"
-        assert (
-            self.target_tlayer < self.num_tasks
-        ), f"target task layer must be lower than the number of tasks, \
+        assert self.num_tasks == 2 or self.num_tasks == 3, (
+            f"SIC_SFU2023 supports only 2 or 3 task layers, but got {self.num_tasks}"
+        )
+        assert self.target_tlayer < self.num_tasks, (
+            f"target task layer must be lower than the number of tasks, \
             but got {self.target_tlayer} < {self.num_tasks}"
+        )
 
         self.trg_vmodel = self.vmodels[self.target_tlayer]
 
@@ -572,9 +572,9 @@ def compress(self, x, target_layer=0, feature_only=False):
                 "models (the entropy coder is run sequentially on CPU)."
             )
 
-        assert (
-            target_layer < self.NUM_LAYERS
-        ), f"Got the target layer {target_layer}, but should be less than {self.NUM_LAYERS}"
+        assert target_layer < self.NUM_LAYERS, (
+            f"Got the target layer {target_layer}, but should be less than {self.NUM_LAYERS}"
+        )
 
         y = self.g_a(x)
         z = self.h_a(y)

compressai_vision/codecs/std_codecs.py

@@ -436,7 +436,7 @@ def encode(
             nb_frames, frame_height, frame_width = frames.size()
             input_bitdepth = self.enc_cfgs["input_bitdepth"]
             chroma_format = self.enc_cfgs["chroma_format"]
-            file_prefix = f"{file_prefix}_{frame_width}x{frame_height}_{self.frame_rate }fps_{input_bitdepth}bit_p{chroma_format}"
+            file_prefix = f"{file_prefix}_{frame_width}x{frame_height}_{self.frame_rate}fps_{input_bitdepth}bit_p{chroma_format}"
             yuv_in_path = f"{file_prefix}_input.yuv"
 
             self.yuvio.setWriter(
@@ -480,9 +480,9 @@ def encode(
                 for partial in list_of_bitstreams:
                     Path(partial).unlink()
 
-        assert Path(
-            bitstream_path
-        ).is_file(), f"bitstream {bitstream_path} was not created"
+        assert Path(bitstream_path).is_file(), (
+            f"bitstream {bitstream_path} was not created"
+        )
 
         if not remote_inference:
             inner_codec_bitstream = load_bitstream(bitstream_path)
@@ -601,12 +601,12 @@ def decode(
                 for file_path in sorted(Path(dec_path).glob(f"*{file_prefix}*.png")):
                     rec_frames.append(str(file_path))
 
-                assert (
-                    file_prefix in rec_frames[0]
-                ), f"Can't find a correct filename with {file_prefix} in {dec_path}"
-                assert (
-                    len(rec_frames) == 1
-                ), f"Number of retrieved file must be 1, but got {len(rec_frames)}"
+                assert file_prefix in rec_frames[0], (
+                    f"Can't find a correct filename with {file_prefix} in {dec_path}"
+                )
+                assert len(rec_frames) == 1, (
+                    f"Number of retrieved file must be 1, but got {len(rec_frames)}"
+                )
 
             conversion_time = 0
             output = {"file_names": rec_frames}
@@ -1090,9 +1090,7 @@ def add_descriptor_modes(descriptors, descriptor_mode):
             "vcm_ctc": ["load", "UsingDescriptor", "load", "load"],
             "load": ["load", "UsingDescriptor", "load", "load"],
             "generate": ["save", "GeneratingDescriptor", "save", "save"],
-        }[
-            descriptor_mode
-        ]
+        }[descriptor_mode]
         items = str(bitstream_path).split("/")
         dataset = {
             "SFUHW": "SFU",

compressai_vision/codecs/utils.py

@@ -257,9 +257,9 @@ def reshape_frame_to_feature_pyramid(
         """reshape a frame of channels into the feature pyramid"""
 
         assert isinstance(x, (Tensor, Dict))
-        assert (
-            packing_all_in_one is True
-        ), "packing_all_in_one = False is not supported yet"
+        assert packing_all_in_one is True, (
+            "packing_all_in_one = False is not supported yet"
+        )
 
         top_y = 0
         tiled_frames = {}

compressai_vision/datasets/__init__.py

@@ -31,9 +31,9 @@
     DataCatalog,
     DefaultDataset,
     Detectron2Dataset,
+    SamDataset,
     TrackingDataset,
     deccode_compressed_rle,
-    SamDataset,
 )
 from .utils import get_seq_info
 

compressai_vision/datasets/image.py

@@ -54,8 +54,8 @@
     JDECustomMapper,
     LinearMapper,
     MMPOSECustomMapper,
-    YOLOXCustomMapper,
     SAMCustomMapper,
+    YOLOXCustomMapper,
 )
 
 
@@ -246,9 +246,9 @@ def __init__(self, root, dataset_name, imgs_folder, **kwargs):
         if kwargs["linear_mapper"] is True:
             mapper = LinearMapper()
         else:
-            assert (
-                kwargs["cfg"] is not None
-            ), "A proper mapper information via cfg must be provided"
+            assert kwargs["cfg"] is not None, (
+                "A proper mapper information via cfg must be provided"
+            )
             mapper = DatasetMapper(kwargs["cfg"], False)
 
         self.mapDataset = MapDataset(_dataset, mapper)

compressai_vision/datasets/utils.py

@@ -37,8 +37,8 @@
 
 from jde.utils.datasets import letterbox
 from mmpose.structures.bbox import get_warp_matrix
-from torchvision import transforms
 from torch.nn import functional as F
+from torchvision import transforms
 
 __all__ = [
     "MMPOSECustomMapper",
@@ -139,9 +139,9 @@ def __call__(self, dataset_dict):
         # Read image
         org_img = cv2.imread(dataset_dict["file_name"])  # return img in BGR by default
 
-        assert (
-            len(org_img.shape) == 3
-        ), f"detect an input image with 2 chs, {dataset_dict['file_name']}"
+        assert len(org_img.shape) == 3, (
+            f"detect an input image with 2 chs, {dataset_dict['file_name']}"
+        )
 
         img_h, img_w, _ = org_img.shape
 
@@ -220,9 +220,9 @@ def __call__(self, dataset_dict):
         # Read image
         org_img = cv2.imread(dataset_dict["file_name"])  # return img in BGR by default
 
-        assert (
-            len(org_img.shape) == 3
-        ), f"detect an input image with 2 chs, {dataset_dict['file_name']}"
+        assert len(org_img.shape) == 3, (
+            f"detect an input image with 2 chs, {dataset_dict['file_name']}"
+        )
 
         dataset_dict["height"], dataset_dict["width"], _ = org_img.shape
 
@@ -412,5 +412,5 @@ def get_seq_info(seq_info_path):
     config.read(seq_info_path)
     fps = config["Sequence"]["frameRate"]
     total_frame = config["Sequence"]["seqLength"]
-    name = f'{config["Sequence"]["name"]}_{config["Sequence"]["imWidth"]}x{config["Sequence"]["imHeight"]}_{fps}'
+    name = f"{config['Sequence']['name']}_{config['Sequence']['imWidth']}x{config['Sequence']['imHeight']}_{fps}"
     return name, int(fps), int(total_frame)

compressai_vision/evaluators/evaluators.py

@@ -725,9 +725,9 @@ def digest_summary(summary):
         return ret
 
     def mot_eval(self):
-        assert len(self.dataset) == len(
-            self._predictions
-        ), "Total number of frames are mismatch"
+        assert len(self.dataset) == len(self._predictions), (
+            "Total number of frames are mismatch"
+        )
 
         # skip the very first frame
         for gt_frame in self.dataset[1:]:
@@ -818,9 +818,9 @@ def __init__(
         assert self.seqinfo_path is not None, "Sequence Information must be provided"
 
     def mot_eval(self):
-        assert len(self.dataset) == len(
-            self._predictions
-        ), "Total number of frames are mismatch"
+        assert len(self.dataset) == len(self._predictions), (
+            "Total number of frames are mismatch"
+        )
 
         self._save_all_eval_info(self._predictions)
         _pd_pd = self._format_pd_in_motchallenge(self._predictions)
@@ -874,9 +874,9 @@ def __init__(
         )
 
     def mot_eval(self):
-        assert len(self.dataset) == len(
-            self._predictions
-        ), "Total number of frames are mismatch"
+        assert len(self.dataset) == len(self._predictions), (
+            "Total number of frames are mismatch"
+        )
 
         self._save_all_eval_info(self._predictions)
         _pd_pd = self._format_pd_in_motchallenge(self._predictions)

compressai_vision/evaluators/tf_evaluation_utils/np_box_list_ops.py

@@ -310,7 +310,7 @@ def multi_class_non_max_suppression(boxlist, score_thresh, iou_thresh, max_outpu
     elif len(scores.shape) == 2:
         if scores.shape[1] is None:
             raise ValueError(
-                "scores field must have statically defined second " "dimension"
+                "scores field must have statically defined second dimension"
             )
     else:
         raise ValueError("scores field must be of rank 1 or 2")
@@ -551,7 +551,7 @@ def filter_scores_greater_than(boxlist, thresh):
         raise ValueError("Scores should have rank 1 or 2")
     if len(scores.shape) == 2 and scores.shape[1] != 1:
         raise ValueError(
-            "Scores should have rank 1 or have shape " "consistent with [None, 1]"
+            "Scores should have rank 1 or have shape consistent with [None, 1]"
         )
     high_score_indices = np.reshape(np.where(np.greater(scores, thresh)), [-1]).astype(
         np.int32

compressai_vision/evaluators/tf_evaluation_utils/np_box_mask_list_ops.py

@@ -312,7 +312,7 @@ def multi_class_non_max_suppression(
     elif len(scores.shape) == 2:
         if scores.shape[1] is None:
             raise ValueError(
-                "scores field must have statically defined second " "dimension"
+                "scores field must have statically defined second dimension"
             )
     else:
         raise ValueError("scores field must be of rank 1 or 2")
@@ -430,7 +430,7 @@ def filter_scores_greater_than(box_mask_list, thresh):
         raise ValueError("Scores should have rank 1 or 2")
     if len(scores.shape) == 2 and scores.shape[1] != 1:
         raise ValueError(
-            "Scores should have rank 1 or have shape " "consistent with [None, 1]"
+            "Scores should have rank 1 or have shape consistent with [None, 1]"
         )
     high_score_indices = np.reshape(np.where(np.greater(scores, thresh)), [-1]).astype(
         np.int32