lint fixes

Summary:
Ran the linter.
TODO: need to update the linter as per D21353065.

Reviewed By: bottler

Differential Revision: D21362270

fbshipit-source-id: ad0e781de0a29f565ad25c43bc94a19b1828c020

Author: nikhilaravi

pytorch3d/io/mtl_io.py

@@ -80,7 +80,7 @@ def make_mesh_texture_atlas(
         faces_material_ind = torch.from_numpy(face_material_names == material_name).to(
             faces_verts_uvs.device
         )
-        if (faces_material_ind).sum() > 0:
+        if faces_material_ind.sum() > 0:
             # For these faces, update the base color to the
             # diffuse material color.
             if "diffuse_color" not in props:

pytorch3d/loss/chamfer.py

@@ -117,8 +117,8 @@ def chamfer_distance(
     P2 = y.shape[1]
 
     # Check if inputs are heterogeneous and create a lengths mask.
-    is_x_heterogeneous = ~(x_lengths == P1).all()
-    is_y_heterogeneous = ~(y_lengths == P2).all()
+    is_x_heterogeneous = (x_lengths != P1).any()
+    is_y_heterogeneous = (y_lengths != P2).any()
     x_mask = (
         torch.arange(P1, device=x.device)[None] >= x_lengths[:, None]
     )  # shape [N, P1]

pytorch3d/renderer/mesh/rasterize_meshes.py

@@ -259,7 +259,7 @@ def rasterize_meshes_python(
     N = len(meshes)
     # Assume only square images.
     # TODO(T52813608) extend support for non-square images.
-    H, W, = image_size, image_size
+    H, W = image_size, image_size
     K = faces_per_pixel
     device = meshes.device
 
@@ -479,7 +479,7 @@ def point_line_distance(p, v0, v1):
     if l2 <= kEpsilon:
         return (p - v1).dot(p - v1)  # v0 == v1
 
-    t = (v1v0).dot(p - v0) / l2
+    t = v1v0.dot(p - v0) / l2
     t = torch.clamp(t, min=0.0, max=1.0)
     p_proj = v0 + t * v1v0
     delta_p = p_proj - p

pytorch3d/renderer/utils.py

@@ -308,7 +308,7 @@ def convert_to_tensors_and_broadcast(*args, dtype=torch.float32, device: str = "
     args_Nd = []
     for c in args_1d:
         if c.shape[0] != 1 and c.shape[0] != N:
-            msg = "Got non-broadcastable sizes %r" % (sizes)
+            msg = "Got non-broadcastable sizes %r" % sizes
             raise ValueError(msg)
 
         # Expand broadcast dim and keep non broadcast dims the same size

pytorch3d/structures/meshes.py

@@ -926,8 +926,8 @@ def _compute_packed(self, refresh: bool = False):
             self._num_verts_per_mesh = torch.zeros(
                 (0,), dtype=torch.int64, device=self.device
             )
-            self._faces_packed = -torch.ones(
-                (0, 3), dtype=torch.int64, device=self.device
+            self._faces_packed = -(
+                torch.ones((0, 3), dtype=torch.int64, device=self.device)
             )
             self._faces_packed_to_mesh_idx = torch.zeros(
                 (0,), dtype=torch.int64, device=self.device
@@ -977,8 +977,8 @@ def _compute_edges_packed(self, refresh: bool = False):
             return
 
         if self.isempty():
-            self._edges_packed = -torch.ones(
-                (0, 2), dtype=torch.int64, device=self.device
+            self._edges_packed = torch.full(
+                (0, 2), fill_value=-1, dtype=torch.int64, device=self.device
             )
             self._edges_packed_to_mesh_idx = torch.zeros(
                 (0,), dtype=torch.int64, device=self.device

tests/test_blending.py

@@ -261,7 +261,9 @@ def test_softmax_rgb_blend(self):
         # of the image with surrounding padded values.
         N, S, K = 1, 8, 2
         device = torch.device("cuda")
-        pix_to_face = -torch.ones((N, S, S, K), dtype=torch.int64, device=device)
+        pix_to_face = torch.full(
+            (N, S, S, K), fill_value=-1, dtype=torch.int64, device=device
+        )
         h = int(S / 2)
         pix_to_face_full = torch.randint(
             size=(N, h, h, K), low=0, high=100, device=device

tests/test_cameras.py

@@ -203,7 +203,7 @@ def test_camera_position_from_angles_torch_scalar_grads(self):
             + torch.cos(elev) * torch.cos(azim)
         )
         grad_elev = (
-            -torch.sin(elev) * torch.sin(azim)
+            -(torch.sin(elev)) * torch.sin(azim)
             + torch.cos(elev)
             - torch.sin(elev) * torch.cos(azim)
         )
@@ -260,7 +260,7 @@ def test_camera_position_from_angles_vectors_mixed_broadcast_grads(self):
             + torch.cos(elev) * torch.cos(azim)
         )
         grad_elev = (
-            -torch.sin(elev) * torch.sin(azim)
+            -(torch.sin(elev)) * torch.sin(azim)
             + torch.cos(elev)
             - torch.sin(elev) * torch.cos(azim)
         )
@@ -395,8 +395,8 @@ def init_random_cameras(cam_type: CamerasBase, batch_size: int):
                 cam_params["aspect_ratio"] = torch.rand(batch_size) * 0.5 + 0.5
             else:
                 cam_params["top"] = torch.rand(batch_size) * 0.2 + 0.9
-                cam_params["bottom"] = -torch.rand(batch_size) * 0.2 - 0.9
-                cam_params["left"] = -torch.rand(batch_size) * 0.2 - 0.9
+                cam_params["bottom"] = -(torch.rand(batch_size)) * 0.2 - 0.9
+                cam_params["left"] = -(torch.rand(batch_size)) * 0.2 - 0.9
                 cam_params["right"] = torch.rand(batch_size) * 0.2 + 0.9
         elif cam_type in (SfMOrthographicCameras, SfMPerspectiveCameras):
             cam_params["focal_length"] = torch.rand(batch_size) * 10 + 0.1
@@ -532,7 +532,7 @@ def test_perspective_mixed_inputs_broadcast(self):
         P = cameras.get_projection_transform()
         vertices = torch.tensor([1, 2, 10], dtype=torch.float32)
         z1 = 1.0  # vertices at far clipping plane so z = 1.0
-        z2 = (20.0 / (20.0 - 1.0) * 10.0 + -(20.0) / (20.0 - 1.0)) / 10.0
+        z2 = (20.0 / (20.0 - 1.0) * 10.0 + -20.0 / (20.0 - 1.0)) / 10.0
         projected_verts = torch.tensor(
             [
                 [np.sqrt(3) / 10.0, 2 * np.sqrt(3) / 10.0, z1],
@@ -660,7 +660,7 @@ def test_orthographic_mixed_inputs_broadcast(self):
         cameras = OpenGLOrthographicCameras(znear=near, zfar=far)
         P = cameras.get_projection_transform()
         vertices = torch.tensor([1.0, 2.0, 10.0], dtype=torch.float32)
-        z2 = 1.0 / (20.0 - 1.0) * 10.0 + -(1.0) / (20.0 - 1.0)
+        z2 = 1.0 / (20.0 - 1.0) * 10.0 + -1.0 / (20.0 - 1.0)
         projected_verts = torch.tensor(
             [[1.0, 2.0, 1.0], [1.0, 2.0, z2]], dtype=torch.float32
         )

tests/test_chamfer.py

@@ -35,6 +35,8 @@ def init_pointclouds(
         low = 0 if allow_empty else 1
         p1_lengths = torch.randint(low, P1, size=(N,), dtype=torch.int64, device=device)
         p2_lengths = torch.randint(low, P2, size=(N,), dtype=torch.int64, device=device)
+        P1 = p1_lengths.max().item()
+        P2 = p2_lengths.max().item()
         weights = torch.rand((N,), dtype=torch.float32, device=device)
 
         # list of points and normals tensors
@@ -109,9 +111,8 @@ def chamfer_distance_naive_pointclouds(p1, p2, device="cpu"):
             torch.arange(P2, device=y.device)[None] >= y_lengths[:, None]
         )  # shape [N, P2]
 
-        is_x_heterogeneous = ~(x_lengths == P1).all()
-        is_y_heterogeneous = ~(y_lengths == P2).all()
-
+        is_x_heterogeneous = (x_lengths != P1).any()
+        is_y_heterogeneous = (y_lengths != P2).any()
         # Only calculate the distances for the points which are not masked
         for n in range(N):
             for i1 in range(x_lengths[n]):

tests/test_mesh_normal_consistency.py

@@ -180,7 +180,7 @@ def test_mesh_normal_consistency_simple(self):
         # mesh1: normal consistency computation
         n0 = (verts1[1] - verts1[2]).cross(verts1[3] - verts1[2])
         n1 = (verts1[1] - verts1[2]).cross(verts1[0] - verts1[2])
-        loss1 = 1.0 - torch.cosine_similarity(n0.view(1, 3), -n1.view(1, 3))
+        loss1 = 1.0 - torch.cosine_similarity(n0.view(1, 3), -(n1.view(1, 3)))
 
         # mesh2: normal consistency computation
         # In the cube mesh, 6 edges are shared with coplanar faces (loss=0),
@@ -193,9 +193,9 @@ def test_mesh_normal_consistency_simple(self):
         n2 = (verts3[1] - verts3[2]).cross(verts3[4] - verts3[2])
         loss3 = (
             3.0
-            - torch.cosine_similarity(n0.view(1, 3), -n1.view(1, 3))
-            - torch.cosine_similarity(n0.view(1, 3), -n2.view(1, 3))
-            - torch.cosine_similarity(n1.view(1, 3), -n2.view(1, 3))
+            - torch.cosine_similarity(n0.view(1, 3), -(n1.view(1, 3)))
+            - torch.cosine_similarity(n0.view(1, 3), -(n2.view(1, 3)))
+            - torch.cosine_similarity(n1.view(1, 3), -(n2.view(1, 3)))
         )
         loss3 /= 3.0
 

tests/test_obj_io.py

@@ -52,11 +52,11 @@ def test_load_obj_simple(self):
         )
         self.assertTrue(torch.all(verts == expected_verts))
         self.assertTrue(torch.all(faces.verts_idx == expected_faces))
-        padded_vals = -torch.ones_like(faces.verts_idx)
+        padded_vals = -(torch.ones_like(faces.verts_idx))
         self.assertTrue(torch.all(faces.normals_idx == padded_vals))
         self.assertTrue(torch.all(faces.textures_idx == padded_vals))
         self.assertTrue(
-            torch.all(faces.materials_idx == -torch.ones(len(expected_faces)))
+            torch.all(faces.materials_idx == -(torch.ones(len(expected_faces))))
         )
         self.assertTrue(normals is None)
         self.assertTrue(textures is None)
@@ -124,10 +124,12 @@ def test_load_obj_complex(self):
             [[0.749279, 0.501284], [0.999110, 0.501077], [0.999455, 0.750380]],
             dtype=torch.float32,
         )
-        expected_faces_normals_idx = -torch.ones_like(expected_faces, dtype=torch.int64)
+        expected_faces_normals_idx = -(
+            torch.ones_like(expected_faces, dtype=torch.int64)
+        )
         expected_faces_normals_idx[4, :] = torch.tensor([1, 1, 1], dtype=torch.int64)
-        expected_faces_textures_idx = -torch.ones_like(
-            expected_faces, dtype=torch.int64
+        expected_faces_textures_idx = -(
+            torch.ones_like(expected_faces, dtype=torch.int64)
         )
         expected_faces_textures_idx[4, :] = torch.tensor([0, 0, 1], dtype=torch.int64)
 
@@ -207,7 +209,7 @@ def test_load_obj_textures_only(self):
         self.assertClose(expected_textures, textures)
         self.assertClose(expected_verts, verts)
         self.assertTrue(
-            torch.all(faces.normals_idx == -torch.ones_like(faces.textures_idx))
+            torch.all(faces.normals_idx == -(torch.ones_like(faces.textures_idx)))
         )
         self.assertTrue(normals is None)
         self.assertTrue(materials is None)

tests/test_point_mesh_distance.py

@@ -431,7 +431,7 @@ def test_point_mesh_edge_distance(self):
         # Naive implementation: forward & backward
         edges_packed = meshes.edges_packed()
         edges_list = packed_to_list(edges_packed, meshes.num_edges_per_mesh().tolist())
-        loss_naive = torch.zeros((N), dtype=torch.float32, device=device)
+        loss_naive = torch.zeros(N, dtype=torch.float32, device=device)
         for i in range(N):
             points = pcls.points_list()[i]
             verts = meshes.verts_list()[i]
@@ -461,7 +461,7 @@ def test_point_mesh_edge_distance(self):
         self.assertClose(loss_op, loss_naive)
 
         # Compare backward pass
-        rand_val = torch.rand((1)).item()
+        rand_val = torch.rand(1).item()
         grad_dist = torch.tensor(rand_val, dtype=torch.float32, device=device)
 
         loss_naive.backward(grad_dist)
@@ -707,7 +707,7 @@ def test_point_mesh_face_distance(self):
         pcls_op = Pointclouds(points_op)
 
         # naive implementation
-        loss_naive = torch.zeros((N), dtype=torch.float32, device=device)
+        loss_naive = torch.zeros(N, dtype=torch.float32, device=device)
         for i in range(N):
             points = pcls.points_list()[i]
             verts = meshes.verts_list()[i]
@@ -735,7 +735,7 @@ def test_point_mesh_face_distance(self):
         self.assertClose(loss_op, loss_naive)
 
         # Compare backward pass
-        rand_val = torch.rand((1)).item()
+        rand_val = torch.rand(1).item()
         grad_dist = torch.tensor(rand_val, dtype=torch.float32, device=device)
 
         loss_naive.backward(grad_dist)

tests/test_points_alignment.py

@@ -112,11 +112,13 @@ def test_init_transformation(self, batch_size=10):
                     ]
 
                     # run full icp
-                    converged, _, Xt, (
-                        R,
-                        T,
-                        s,
-                    ), t_hist = points_alignment.iterative_closest_point(
+                    (
+                        converged,
+                        _,
+                        Xt,
+                        (R, T, s),
+                        t_hist,
+                    ) = points_alignment.iterative_closest_point(
                         X,
                         Y,
                         estimate_scale=False,
@@ -130,11 +132,13 @@ def test_init_transformation(self, batch_size=10):
                     t_init = t_hist[min(2, len(t_hist) - 1)]
 
                     # rerun the ICP
-                    converged_init, _, Xt_init, (
-                        R_init,
-                        T_init,
-                        s_init,
-                    ), t_hist_init = points_alignment.iterative_closest_point(
+                    (
+                        converged_init,
+                        _,
+                        Xt_init,
+                        (R_init, T_init, s_init),
+                        t_hist_init,
+                    ) = points_alignment.iterative_closest_point(
                         X,
                         Y,
                         init_transform=t_init,
@@ -182,11 +186,13 @@ def test_heterogenous_inputs(self, batch_size=10):
                 n_points_Y = Y_pcl.num_points_per_cloud()
 
                 # run icp with Pointlouds inputs
-                _, _, Xt_pcl, (
-                    R_pcl,
-                    T_pcl,
-                    s_pcl,
-                ), _ = points_alignment.iterative_closest_point(
+                (
+                    _,
+                    _,
+                    Xt_pcl,
+                    (R_pcl, T_pcl, s_pcl),
+                    _,
+                ) = points_alignment.iterative_closest_point(
                     X_pcl,
                     Y_pcl,
                     estimate_scale=estimate_scale,
@@ -263,11 +269,13 @@ def _compare_with_trimesh(
         ]
 
         # run the icp algorithm
-        converged, _, _, (
-            R_ours,
-            T_ours,
-            s_ours,
-        ), _ = points_alignment.iterative_closest_point(
+        (
+            converged,
+            _,
+            _,
+            (R_ours, T_ours, s_ours),
+            _,
+        ) = points_alignment.iterative_closest_point(
             X,
             Y,
             estimate_scale=estimate_scale,

tests/test_points_normals.py

@@ -74,7 +74,10 @@ def test_pcl_normals(self, batch_size=3, num_points=300, neighborhood_size=50):
 
                 # check for both disambiguation options
                 for disambiguate_directions in (True, False):
-                    curvatures, local_coord_frames = estimate_pointcloud_local_coord_frames(
+                    (
+                        curvatures,
+                        local_coord_frames,
+                    ) = estimate_pointcloud_local_coord_frames(
                         pcl,
                         neighborhood_size=neighborhood_size,
                         disambiguate_directions=disambiguate_directions,

tests/test_rasterize_meshes.py

@@ -448,7 +448,7 @@ def _test_back_face_culling(self, rasterize_meshes_fn, device, bin_size):
         ], dtype=torch.int64, device=device)
         # fmt: on
 
-        pix_to_face_padded = -torch.ones_like(pix_to_face_frontface)
+        pix_to_face_padded = -(torch.ones_like(pix_to_face_frontface))
         # Run with and without culling
         # Without culling, for k=0, the front face (i.e. face 2) is
         # rasterized and for k=1, the back face (i.e. face 3) is

tests/test_sample_points_from_meshes.py

@@ -95,7 +95,7 @@ def test_sampling_output(self):
         x, y, z = samples[1, :].unbind(1)
         radius = torch.sqrt(x ** 2 + y ** 2 + z ** 2)
 
-        self.assertClose(radius, torch.ones((num_samples)))
+        self.assertClose(radius, torch.ones(num_samples))
 
         # Pyramid: points shoudl lie on one of the faces.
         pyramid_verts = samples[2, :]