[NumPy] Remove references to deprecated NumPy type aliases.

This change replaces references to a number of deprecated NumPy type aliases (np.bool, np.int, np.float, np.complex, np.object, np.str) with their recommended replacement (bool, int, float, complex, object, str).

NumPy 1.24 drops the deprecated aliases, so we must remove uses before updating NumPy.

PiperOrigin-RevId: 496402381

Author: hawkinsp

research/a2n/dataset.py

@@ -36,7 +36,7 @@ def get_graph_nbrhd(train_graph, ent, exclude_tuple):
                        # er not in train_graph.reverse_kg_data[ent][nbr]]
                        (train_graph.reverse_kg_data[ent][nbr] - set([er]))]
     neighborhood += rev_nighborhood
-  neighborhood = np.array(list(set(neighborhood)), dtype=np.int)
+  neighborhood = np.array(list(set(neighborhood)), dtype=int)
   return neighborhood
 
 
@@ -55,7 +55,7 @@ def get_graph_nbrhd_with_rels(train_graph, ent, exclude_tuple):
   #                      # er not in train_graph.reverse_kg_data[ent][nbr]]
   #                      (train_graph.reverse_kg_data[ent][nbr] - set([er]))]
   #   neighborhood += rev_nighborhood
-  neighborhood = np.array(neighborhood, dtype=np.int)
+  neighborhood = np.array(neighborhood, dtype=int)
   return neighborhood
 
 
@@ -78,7 +78,7 @@ def get_graph_nbrhd_text(train_graph, ent, max_text_len):
   #                      # er not in train_graph.reverse_kg_data[ent][nbr]]
   #                      (train_graph.reverse_kg_data[ent][nbr] - set([er]))]
   #   neighborhood += rev_nighborhood
-  neighborhood = np.array(neighborhood, dtype=np.int)
+  neighborhood = np.array(neighborhood, dtype=int)
   return neighborhood
 
 
@@ -94,7 +94,7 @@ def get_graph_nbrhd_embd_text(train_graph, ent, max_text_nbrs):
   if not neighborhood:
     neighborhood = [[]]
     neighborhood_emb = [np.zeros(train_graph.embeddings[0].size)]
-  neighborhood = np.array(neighborhood, dtype=np.int)
+  neighborhood = np.array(neighborhood, dtype=int)
   neighborhood_emb = np.array(neighborhood_emb, dtype=np.float32)
   if neighborhood.shape[0] > max_text_nbrs:
     ids = np.random.choice(np.range(neighborhood.shape[0]),
@@ -162,7 +162,7 @@ def get_graph_nbrhd_paths(train_graph, ent, exclude_tuple):
     neighborhood += paths
   if not neighborhood:
     neighborhood = [[]]
-  neighborhood = np.array(neighborhood, dtype=np.int)
+  neighborhood = np.array(neighborhood, dtype=int)
   return neighborhood
 
 
@@ -223,7 +223,7 @@ def get_graph_nbrhd_paths_randwalk(train_graph, ent, exclude_tuple,
   if not neighborhood:
     neighborhood = [[]]
   # import pdb; pdb.set_trace()
-  neighborhood = np.array(neighborhood, dtype=np.int)
+  neighborhood = np.array(neighborhood, dtype=int)
   return neighborhood
 
 
@@ -354,7 +354,7 @@ def featurize_each_example(self, example_tuple):
     # else:
     #   negatives = np.array(candidate_negatives)
     negatives = sample_or_pad(
-        np.array(candidate_negatives, dtype=np.int), self.max_negatives,
+        np.array(candidate_negatives, dtype=int), self.max_negatives,
         pad_value=self.train_graph.ent_pad
     )
     # negatives is an array of shape (max_negatives)
@@ -377,8 +377,8 @@ def featurize_each_example(self, example_tuple):
       nbrhd_fn = get_graph_nbrhd
       pad_value = self.train_graph.ent_pad
     if self.model_type == "distmult":
-      nbrs_s = np.array([], dtype=np.int)
-      nbrs_candidates = np.array([], dtype=np.int)
+      nbrs_s = np.array([], dtype=int)
+      nbrs_candidates = np.array([], dtype=int)
     elif self.model_type in ["source_attention", "source_rel_attention",
                              "source_path_attention"]:
       nbrs_s = sample_or_pad(nbrhd_fn(self.train_graph, s, (s, r, t)),
@@ -396,7 +396,7 @@ def featurize_each_example(self, example_tuple):
             get_graph_nbrhd_text(self.train_graph, s, self.max_text_len),
             self.max_text_neighbors, pad_value=text_pad_value
         )
-      nbrs_candidates = np.array([], dtype=np.int)
+      nbrs_candidates = np.array([], dtype=int)
     else:
       nbrs_s = sample_or_pad(nbrhd_fn(self.train_graph, s, (s, r, t)),
                              self.max_neighbors,
@@ -417,7 +417,7 @@ def featurize_each_example(self, example_tuple):
     if self.mode != "train":
       labels = [t]
     else:
-      labels = np.zeros(candidates.shape[0], dtype=np.int)
+      labels = np.zeros(candidates.shape[0], dtype=int)
       labels[0] = 1
       idx = np.arange(candidates.shape[0])
       np.random.shuffle(idx)

research/gam/gam/data/dataset.py

@@ -669,7 +669,7 @@ def __init__(self, dataset, keep_label_proportions=False, inductive=False):
       labeled_indices = list(self.get_indices_train())
       labeled_indices_labels = dataset.get_labels(labeled_indices)
       label_counts = collections.Counter(labeled_indices_labels)
-      num_labels = np.float(len(labeled_indices))
+      num_labels = float(len(labeled_indices))
       assert num_labels > 0, 'There are no labeled samples in the dataset.'
       self.label_prop = {
           label: count / num_labels for label, count in label_counts.items()

research/gam/gam/data/loaders.py

@@ -94,7 +94,7 @@ def load_data_realistic_ssl(dataset_name, data_path, label_map_path):
   train_indices = np.asarray(train_indices)
 
   # Select the loaded train indices, and make the rest unlabeled.
-  unlabeled_mask = np.ones((train_inputs.shape[0],), dtype=np.bool)
+  unlabeled_mask = np.ones((train_inputs.shape[0],), dtype=bool)
   unlabeled_mask[train_indices] = False
   unlabeled_inputs = train_inputs[unlabeled_mask]
   unlabeled_labels = train_labels[unlabeled_mask]
@@ -159,7 +159,7 @@ def _sample_mask(idx, l):
     """Create mask."""
     mask = np.zeros(l)
     mask[idx] = 1
-    return np.array(mask, dtype=np.bool)
+    return np.array(mask, dtype=bool)
 
   def _parse_index_file(filename):
     """Parse index file."""