Merge pull request #129 from agimus-project/128-numpyint-deprecated-o…

…ther-type-aliases-too

128 numpyint deprecated other type aliases too

happypose/pose_estimators/cosypose/cosypose/evaluation/eval_runner/detection_eval.py

@@ -63,10 +63,10 @@ def __init__(
     @staticmethod
     def make_empty_predictions():
         infos = {
-            "view_id": np.empty(0, dtype=np.int),
-            "scene_id": np.empty(0, dtype=np.int),
+            "view_id": np.empty(0, dtype=int),
+            "scene_id": np.empty(0, dtype=int),
             "label": np.empty(0, dtype=np.object),
-            "score": np.empty(0, dtype=np.float),
+            "score": np.empty(0, dtype=float),
         }
         bboxes = torch.empty(0, 4, dtype=torch.float)
         return tc.PandasTensorCollection(infos=pd.DataFrame(infos), bboxes=bboxes)

happypose/pose_estimators/cosypose/cosypose/evaluation/eval_runner/pose_eval.py

@@ -63,10 +63,10 @@ def __init__(
     @staticmethod
     def make_empty_predictions():
         infos = {
-            "view_id": np.empty(0, dtype=np.int),
-            "scene_id": np.empty(0, dtype=np.int),
+            "view_id": np.empty(0, dtype=int),
+            "scene_id": np.empty(0, dtype=int),
             "label": np.empty(0, dtype=np.object),
-            "score": np.empty(0, dtype=np.float),
+            "score": np.empty(0, dtype=float),
         }
         poses = torch.empty(0, 4, 4, dtype=torch.float)
         return tc.PandasTensorCollection(infos=pd.DataFrame(infos), poses=poses)

happypose/pose_estimators/cosypose/cosypose/evaluation/meters/detection_meters.py

@@ -223,7 +223,7 @@ def compute_ap(label_df, label_n_gt):
             label_df = label_df.sort_values("score", ascending=False).reset_index(
                 drop=True,
             )
-            label_df["n_tp"] = np.cumsum(label_df[valid_k].values.astype(np.float))
+            label_df["n_tp"] = np.cumsum(label_df[valid_k].values.astype(float))
             label_df["prec"] = label_df["n_tp"] / (np.arange(len(label_df)) + 1)
             label_df["recall"] = label_df["n_tp"] / label_n_gt
             y_true = label_df[valid_k]

happypose/pose_estimators/cosypose/cosypose/evaluation/meters/pose_meters.py

@@ -361,7 +361,7 @@ def compute_ap(label_df, label_n_gt):
             label_df = label_df.sort_values("score", ascending=False).reset_index(
                 drop=True,
             )
-            label_df["n_tp"] = np.cumsum(label_df[valid_k].values.astype(np.float))
+            label_df["n_tp"] = np.cumsum(label_df[valid_k].values.astype(float))
             label_df["prec"] = label_df["n_tp"] / (np.arange(len(label_df)) + 1)
             label_df["recall"] = label_df["n_tp"] / label_n_gt
             y_true = label_df[valid_k]

happypose/pose_estimators/cosypose/cosypose/evaluation/meters/utils.py

@@ -25,7 +25,7 @@ def add_inst_num(
     group_keys=["scene_id", "view_id", "label"],
     key="pred_inst_num",
 ):
-    inst_num = np.empty(len(infos), dtype=np.int)
+    inst_num = np.empty(len(infos), dtype=int)
     for _group_name, group_ids in infos.groupby(group_keys).groups.items():
         inst_num[group_ids.values] = np.arange(len(group_ids))
     infos[key] = inst_num

happypose/pose_estimators/cosypose/cosypose/evaluation/pred_runner/multiview_predictions.py

@@ -137,7 +137,7 @@ def get_predictions(
                     keep_ids = np.concatenate(keep_ids)
                     batch_im_ids = np.concatenate(batch_im_ids)
                 detections_ = detections[keep_ids]
-                detections_.infos["batch_im_id"] = np.array(batch_im_ids).astype(np.int)
+                detections_.infos["batch_im_id"] = np.array(batch_im_ids).astype(int)
             else:
                 msg = "No detections"
                 raise ValueError(msg)

happypose/pose_estimators/cosypose/cosypose/integrated/icp_refiner.py

@@ -28,7 +28,7 @@ def get_normal(
         mask[depth_refine == 0] = 1
         depth_refine = depth_refine.astype(np.float32)
         depth_refine = cv2.inpaint(depth_refine, mask, 2, cv2.INPAINT_NS)
-        depth_refine = depth_refine.astype(np.float)
+        depth_refine = depth_refine.astype(float)
         depth_refine = ndimage.gaussian_filter(depth_refine, 2)
 
     uv_table = np.zeros((res_y, res_x, 2), dtype=np.int16)

happypose/pose_estimators/cosypose/cosypose/integrated/multiview_predictor.py

@@ -17,12 +17,10 @@
 
 class MultiviewScenePredictor:
     def __init__(self, mesh_db, n_sym=64, ba_aabb=True, ba_n_points=None):
-        self.mesh_db_ransac = mesh_db.batched(n_sym=n_sym, aabb=True).cuda().float()
-        self.mesh_db_ba = (
-            mesh_db.batched(aabb=ba_aabb, resample_n_points=ba_n_points, n_sym=n_sym)
-            .cuda()
-            .float()
-        )
+        self.mesh_db_ransac = mesh_db.batched(n_sym=n_sym, aabb=True).float()
+        self.mesh_db_ba = mesh_db.batched(
+            aabb=ba_aabb, resample_n_points=ba_n_points, n_sym=n_sym
+        ).float()
 
     def reproject_scene(self, objects, cameras):
         TCO_data = []

happypose/pose_estimators/cosypose/cosypose/multiview/ransac.py

@@ -102,15 +102,15 @@ def score_tmaches_batch(candidates, tmatches, TC1C2, mesh_db, bsz=4096):
 def scene_level_matching(candidates, inliers):
     cand1 = inliers["inlier_matches_cand1"]
     cand2 = inliers["inlier_matches_cand2"]
-    edges = np.ones((len(cand1)), dtype=np.int)
+    edges = np.ones((len(cand1)), dtype=int)
     n_cand = len(candidates)
     graph = csr_matrix((edges, (cand1, cand2)), shape=(n_cand, n_cand))
     n_components, ids = connected_components(graph, directed=True, connection="strong")
 
     component_size = defaultdict(lambda: 0)
     for idx in ids:
         component_size[idx] += 1
-    obj_n_cand = np.empty(len(ids), dtype=np.int)
+    obj_n_cand = np.empty(len(ids), dtype=int)
     for n, idx in enumerate(ids):
         obj_n_cand[n] = component_size[idx]
 
@@ -132,7 +132,7 @@ def scene_level_matching(candidates, inliers):
 def make_obj_infos(matched_candidates):
     scene_infos = matched_candidates.infos.loc[:, ["obj_id", "score", "label"]].copy()
     gb = scene_infos.groupby("obj_id")
-    scene_infos["n_cand"] = gb["score"].transform(len).astype(np.int)
+    scene_infos["n_cand"] = gb["score"].transform(len).astype(int)
     scene_infos["score"] = gb["score"].transform(np.sum)
     scene_infos = gb.first().reset_index(drop=False)
     return scene_infos

happypose/pose_estimators/cosypose/cosypose/recording/bop_recording_scene.py

@@ -151,7 +151,7 @@ def objects_pos_orn_rand_falling(self):
 
     def sample_camera(self):
         assert self.focal_interval.shape == (2, 2)
-        K = np.zeros((3, 3), dtype=np.float)
+        K = np.zeros((3, 3), dtype=float)
         fxfy = self.np_random.uniform(*self.focal_interval)
         W, H = max(self.resolution), min(self.resolution)
         K[0, 0] = fxfy[0]

happypose/pose_estimators/cosypose/cosypose/utils/xarray.py

@@ -24,7 +24,7 @@ def xr_merge(ds1, ds2, on, how="left", dim1="dim_0", dim2="dim_0", fill_value=np
     mask = np.isfinite(idx2)
     # assert mask.sum() == ds2.dims[dim1]
     idx1 = idx1[mask]
-    idx2 = idx2[mask].astype(np.int)
+    idx2 = idx2[mask].astype(int)
 
     for k, data_var in ds2.data_vars.items():
         array = data_var.values

happypose/pose_estimators/cosypose/cosypose/visualization/multiview.py

@@ -179,9 +179,9 @@ def mark_inliers(cand_inputs, cand_matched):
         on=["scene_id", "view_id", "label", "cand_id"],
         how="left",
     )
-    infos["is_inlier"] = infos["is_inlier"].astype(np.float)
-    infos.loc[~np.isfinite(infos.loc[:, "is_inlier"].astype(np.float)), "is_inlier"] = 0
-    infos["is_inlier"] = infos["is_inlier"].astype(np.bool)
+    infos["is_inlier"] = infos["is_inlier"].astype(float)
+    infos.loc[~np.isfinite(infos.loc[:, "is_inlier"].astype(float)), "is_inlier"] = 0
+    infos["is_inlier"] = infos["is_inlier"].astype(bool)
     cand_inputs.infos = infos
     return cand_inputs
 

happypose/pose_estimators/megapose/evaluation/evaluation_runner.py

@@ -74,10 +74,10 @@ def __init__(
     @staticmethod
     def make_empty_predictions():
         infos = {
-            "view_id": np.empty(0, dtype=np.int),
-            "scene_id": np.empty(0, dtype=np.int),
+            "view_id": np.empty(0, dtype=int),
+            "scene_id": np.empty(0, dtype=int),
             "label": np.empty(0, dtype=np.object),
-            "score": np.empty(0, dtype=np.float),
+            "score": np.empty(0, dtype=float),
         }
         poses = torch.empty(0, 4, 4, dtype=torch.float)
         return tc.PandasTensorCollection(infos=pd.DataFrame(infos), poses=poses)

happypose/pose_estimators/megapose/inference/types.py

@@ -172,8 +172,8 @@ def from_numpy(
         Args:
         ----
             rgb: [H,W,3] np.uint8
-            depth: [H,W] np.float
-            K: [3,3] np.float
+            depth: [H,W] float
+            K: [3,3] float
 
         """
         assert rgb.dtype == np.uint8

happypose/toolbox/inference/types.py

@@ -172,8 +172,8 @@ def from_numpy(
         Args:
         ----
             rgb: [H,W,3] np.uint8
-            depth: [H,W] np.float
-            K: [3,3] np.float
+            depth: [H,W] float
+            K: [3,3] float
 
         """
         assert rgb.dtype == np.uint8

happypose/toolbox/visualization/meshcat_utils.py

@@ -88,7 +88,7 @@ def visualize_mesh(vis, mesh, transform=None, color=None, texture_png=None):
             color_hex = rgb2hex(tuple(color))
             material = meshcat.geometry.MeshPhongMaterial(color=color_hex)
         else:  # color is np.ndarray, e.g. [1,0,0]
-            if not np.issubdtype(color.dtype, np.int):
+            if not np.issubdtype(color.dtype, int):
                 color = (color * 255).astype(np.int32)
             color_hex = rgb2hex(tuple(color))
             material = meshcat.geometry.MeshPhongMaterial(color=color_hex)
@@ -112,7 +112,7 @@ def visualize_scene(vis, object_dict, randomize_color=True):
                 color = data["color"]
 
                 # if it's not an integer, convert it to [0,255]
-                if not np.issubdtype(color.dtype, np.int):
+                if not np.issubdtype(color.dtype, int):
                     color = (color * 255).astype(np.int32)
             else:
                 color = np.random.randint(low=0, high=256, size=3)