Rename gtar file handle in Trajectory

keras_gtar/trajectory.py

@@ -22,7 +22,7 @@ class Trajectory:
     def __init__(self, filename, mode='r'):
         self.filename = filename
         self.mode = mode
-        self.traj = gtar.GTAR(filename, mode)
+        self.handle = gtar.GTAR(filename, mode)
 
     def __enter__(self):
         return self
@@ -34,25 +34,25 @@ def __len__(self):
         return len(self.frames)
 
     def close(self):
-        self.traj.close()
+        self.handle.close()
 
     @property
     def frames(self):
-        (_, frames) = self.traj.framesWithRecordsNamed('weight')
+        (_, frames) = self.handle.framesWithRecordsNamed('weight')
         return frames
 
     def get_weights(self, frame=-1):
         """Returns a list of weight arrays for a model stored at the given frame index
 
         :param frame: integer index of the step to load. Can be negative to count from the end.
         """
-        (_, frames) = self.traj.framesWithRecordsNamed('weight')
+        (_, frames) = self.handle.framesWithRecordsNamed('weight')
         frame_index = frames[frame]
 
         weight_records = collections.defaultdict(dict)
         shape_records = collections.defaultdict(dict)
         weight_pattern = re.compile(r'keras/layer/(?P<layer>\d+)/weight/(?P<weight>\d+)')
-        for rec in self.traj.getRecordTypes():
+        for rec in self.handle.getRecordTypes():
             match = weight_pattern.search(rec.getGroup())
             if not match:
                 continue
@@ -69,8 +69,8 @@ def get_weights(self, frame=-1):
             for weight_index in range(len(records)):
                 weight_rec = records[weight_index]
                 shape_rec = shape_records[i][weight_index]
-                shape = self.traj.getRecord(shape_rec, frame_index)
-                weight = self.traj.getRecord(weight_rec, frame_index)
+                shape = self.handle.getRecord(shape_rec, frame_index)
+                weight = self.handle.getRecord(weight_rec, frame_index)
                 weight = weight.reshape(shape)
                 all_weights.append(weight)
 
@@ -81,10 +81,10 @@ def load(self, frame=-1):
 
         :param frame: integer index of the step to load. Can be negative to count from the end.
         """
-        model_description = self.traj.readStr('keras/model.json')
+        model_description = self.handle.readStr('keras/model.json')
         assert model_description
 
-        extra_classes = self.traj.readBytes('keras/layer_classes.pkl')
+        extra_classes = self.handle.readBytes('keras/layer_classes.pkl')
         extra_classes = pickle.loads(extra_classes) if extra_classes else {}
 
         model = keras.models.model_from_json(model_description, extra_classes)
@@ -105,8 +105,8 @@ def save(self, model, frame=None, only_weights=False):
             layer_classes = {type(layer).__name__: type(layer) for layer in model.layers}
             layer_classes = pickle.dumps(layer_classes)
 
-            self.traj.writeStr('keras/model.json', model_json)
-            self.traj.writeBytes('keras/layer_classes.pkl', layer_classes)
+            self.handle.writeStr('keras/model.json', model_json)
+            self.handle.writeBytes('keras/layer_classes.pkl', layer_classes)
         else:
             assert frame, 'Trying to save only the weights of a model without a frame given'
 
@@ -117,8 +117,8 @@ def save(self, model, frame=None, only_weights=False):
                 for (j, weight) in enumerate(layer.get_weights()):
                     dtype_string = dtypes[weight.dtype.name]
                     group = 'keras/layer/{}/weight/{}'.format(i, j)
-                    self.traj.writePath('{}/frames/{}/weight.{}.uni'.format(group, frame, dtype_string), weight)
-                    self.traj.writePath('{}/shape.u32.uni'.format(group), weight.shape)
+                    self.handle.writePath('{}/frames/{}/weight.{}.uni'.format(group, frame, dtype_string), weight)
+                    self.handle.writePath('{}/shape.u32.uni'.format(group), weight.shape)
 
     def save_weights(self, model, frame):
         """Save (only) the current model weights.