Slots

xesn/cost.py

@@ -19,6 +19,9 @@
 from .psd import psd
 
 class CostFunction():
+
+    __slots__ = ("ESN", "train_data", "macro_data", "config")
+
     def __init__(self, ESN, train_data, macro_data, config):
 
         self.ESN        = ESN

xesn/driver.py

@@ -30,17 +30,18 @@ class Driver():
         config (str or dict): either a path to a yaml file or dict containing experiment parameters
         output_directory (str, optional): directory to save results and write logs to
     """
-    name                    = "driver"
-    config                  = None
-    output_directory        = None
-    walltime                = None
-    localtime               = None
+
+    __slots__ = (
+        "config", "output_directory",
+        "walltime", "localtime",
+        "esn_name", "ESN",
+        "logfile", "logname", "logger",
+    )
 
     def __init__(self,
                  config,
                  output_directory=None):
 
-
         self._make_output_directory(output_directory)
         self._create_logger()
         self.set_config(config)
@@ -277,7 +278,7 @@ def get_sample_indices(self, data_length, n_samples, n_steps, n_spinup, random_s
 
     def _make_output_directory(self, out_dir):
         """Make provided output directory. If none given, make a unique directory:
-        output-{self.name}-XX, where XX is 00->99
+        output-driver-XX, where XX is 00->99
 
         Args:
             out_dir (str or None): path to dump output, or None for default
@@ -289,11 +290,11 @@ def _make_output_directory(self, out_dir):
 
             # make a unique default directory
             i=0
-            out_dir = f"output-{self.name}-{i:02d}"
+            out_dir = f"output-driver-{i:02d}"
             while os.path.isdir(out_dir):
                 if i>99:
                     raise ValueError("Hit max number of default output directories...")
-                out_dir = f"output-{self.name}-{i:02d}"
+                out_dir = f"output-driver-{i:02d}"
                 i = i+1
             os.makedirs(out_dir)
 
@@ -317,7 +318,7 @@ def _create_logger(self):
         self.logfile = os.path.join(self.output_directory, 'stdout.log')
 
         # create a logger
-        self.logname = f'{self.name}_logger'
+        self.logname = f'driver_logger'
         self.logger = logging.getLogger(self.logname)
         self.logger.setLevel(logging.DEBUG)
 

xesn/esn.py

@@ -18,12 +18,12 @@ class ESN():
     """A classic ESN architecture, with no distribution or parallelism.
     It is assumed that all data used with this architecture can fit into memory.
     """
-    W                   = None
-    Win                 = None
-    Wout                = None
-    input_kwargs        = None
-    adjacency_kwargs    = None
-    bias_kwargs         = None
+    __slots__ = (
+        "W", "Win", "Wout",
+        "n_input", "n_output", "n_reservoir",
+        "leak_rate", "tikhonov_parameter", "bias_vector",
+        "input_kwargs", "adjacency_kwargs", "bias_kwargs",
+    )
 
     @property
     def input_factor(self):

xesn/lazyesn.py

@@ -23,6 +23,9 @@ class LazyESN(ESN):
         2. Non-global axes, i.e., axes which is chunked up or made up of patches, are first
         3. Can handle multi-dimensional data, but only 2D chunking
     """
+    __slots__ = (
+        "esn_chunks", "overlap", "persist", "boundary"
+    )
 
     @property
     def output_chunks(self):

xesn/matrix.py

@@ -71,18 +71,11 @@ class RandomMatrix():
         random_seed (int, optional): used to control the RNG for matrix generation
     """
 
-    # Set by user
-    n_rows          = None
-    n_cols          = None
-
-    distribution    = None
-    normalization   = "multiply"
-    factor          = 1.0
-
-    random_seed     = None
-
-    # Set automatically
-    dist_kw         = None
+    __slots__ = (
+        "n_rows", "n_cols",
+        "distribution", "normalization", "factor",
+        "random_seed", "random_state",
+    )
 
     def __init__(
             self,
@@ -225,8 +218,9 @@ class SparseRandomMatrix(RandomMatrix):
         random_seed (int, optional): used to control the RNG for matrix generation
     """
 
-    density         = None
-    format          = "coo" # scipy's default
+    __slots__ = (
+        "density", "format",
+    )
 
     def __init__(
             self,

xesn/xdata.py

@@ -8,6 +8,11 @@ class XData():
     See :meth:`setup` for the main usage.
     """
 
+    __slots__ = (
+        "field_name", "zstore_path",
+        "dimensions", "subsampling", "normalization",
+    )
+
     def __init__(self,
                  field_name,
                  zstore_path,