Static analysis improvements python (#3263)

* Simplify code and remove used vars.

* Simplify expressions and remove used parenthesis.

* Jit fixes.

* Making check more readable.

* fixing styles

torchvision/datasets/mnist.py

@@ -471,8 +471,8 @@ def read_sn3_pascalvincent_tensor(path: Union[str, IO], strict: bool = True) ->
     magic = get_int(data[0:4])
     nd = magic % 256
     ty = magic // 256
-    assert nd >= 1 and nd <= 3
-    assert ty >= 8 and ty <= 14
+    assert 1 <= nd <= 3
+    assert 8 <= ty <= 14
     m = SN3_PASCALVINCENT_TYPEMAP[ty]
     s = [get_int(data[4 * (i + 1): 4 * (i + 2)]) for i in range(nd)]
     parsed = np.frombuffer(data, dtype=m[1], offset=(4 * (nd + 1)))

torchvision/datasets/phototour.py

@@ -192,7 +192,6 @@ def read_info_file(data_dir: str, info_file: str) -> torch.Tensor:
     """Return a Tensor containing the list of labels
        Read the file and keep only the ID of the 3D point.
     """
-    labels = []
     with open(os.path.join(data_dir, info_file), 'r') as f:
         labels = [int(line.split()[0]) for line in f]
     return torch.LongTensor(labels)

torchvision/io/_video_opt.py

@@ -303,7 +303,7 @@ def _read_video_timestamps_from_file(filename):
         1,  # audio_timebase_den
     )
     _vframes, vframe_pts, vtimebase, vfps, vduration, \
-        _aframes, aframe_pts, atimebase, asample_rate, aduration = (result)
+        _aframes, aframe_pts, atimebase, asample_rate, aduration = result
     info = _fill_info(vtimebase, vfps, vduration, atimebase, asample_rate, aduration)
 
     vframe_pts = vframe_pts.numpy().tolist()

torchvision/models/detection/backbone_utils.py

@@ -87,7 +87,7 @@ def resnet_fpn_backbone(
         norm_layer=norm_layer)
 
     # select layers that wont be frozen
-    assert trainable_layers <= 5 and trainable_layers >= 0
+    assert 0 <= trainable_layers <= 5
     layers_to_train = ['layer4', 'layer3', 'layer2', 'layer1', 'conv1'][:trainable_layers]
     # freeze layers only if pretrained backbone is used
     for name, parameter in backbone.named_parameters():

torchvision/models/detection/generalized_rcnn.py

@@ -106,6 +106,6 @@ class GeneralizedRCNN(nn.Module):
             if not self._has_warned:
                 warnings.warn("RCNN always returns a (Losses, Detections) tuple in scripting")
                 self._has_warned = True
-            return (losses, detections)
+            return losses, detections
         else:
             return self.eager_outputs(losses, detections)

torchvision/models/detection/roi_heads.py

@@ -190,7 +190,7 @@ def _onnx_heatmaps_to_keypoints(maps, maps_i, roi_map_width, roi_map_height,
 
     xy_preds_i_0 = x + offset_x_i.to(dtype=torch.float32)
     xy_preds_i_1 = y + offset_y_i.to(dtype=torch.float32)
-    xy_preds_i_2 = torch.ones((xy_preds_i_1.shape), dtype=torch.float32)
+    xy_preds_i_2 = torch.ones(xy_preds_i_1.shape, dtype=torch.float32)
     xy_preds_i = torch.stack([xy_preds_i_0.to(dtype=torch.float32),
                               xy_preds_i_1.to(dtype=torch.float32),
                               xy_preds_i_2.to(dtype=torch.float32)], 0)
@@ -795,7 +795,6 @@ class RoIHeads(nn.Module):
                 mask_features = self.mask_head(mask_features)
                 mask_logits = self.mask_predictor(mask_features)
             else:
-                mask_logits = torch.tensor(0)
                 raise Exception("Expected mask_roi_pool to be not None")
 
             loss_mask = {}

torchvision/models/video/resnet.py

@@ -30,7 +30,7 @@ class Conv3DSimple(nn.Conv3d):
 
     @staticmethod
     def get_downsample_stride(stride):
-        return (stride, stride, stride)
+        return stride, stride, stride
 
 
 class Conv2Plus1D(nn.Sequential):
@@ -53,7 +53,7 @@ class Conv2Plus1D(nn.Sequential):
 
     @staticmethod
     def get_downsample_stride(stride):
-        return (stride, stride, stride)
+        return stride, stride, stride
 
 
 class Conv3DNoTemporal(nn.Conv3d):
@@ -75,7 +75,7 @@ class Conv3DNoTemporal(nn.Conv3d):
 
     @staticmethod
     def get_downsample_stride(stride):
-        return (1, stride, stride)
+        return 1, stride, stride
 
 
 class BasicBlock(nn.Module):

torchvision/ops/feature_pyramid_network.py

@@ -99,9 +99,7 @@ class FeaturePyramidNetwork(nn.Module):
         This is equivalent to self.inner_blocks[idx](x),
         but torchscript doesn't support this yet
         """
-        num_blocks = 0
-        for m in self.inner_blocks:
-            num_blocks += 1
+        num_blocks = len(self.inner_blocks)
         if idx < 0:
             idx += num_blocks
         i = 0
@@ -117,9 +115,7 @@ class FeaturePyramidNetwork(nn.Module):
         This is equivalent to self.layer_blocks[idx](x),
         but torchscript doesn't support this yet
         """
-        num_blocks = 0
-        for m in self.layer_blocks:
-            num_blocks += 1
+        num_blocks = len(self.layer_blocks)
         if idx < 0:
             idx += num_blocks
         i = 0

torchvision/transforms/functional_tensor.py

@@ -45,7 +45,7 @@ def _max_value(dtype: torch.dtype) -> float:
             max_value = next_value
             bits *= 2
         else:
-            return max_value.item()
+            break
     return max_value.item()