Merge pull request #1527 from SpiNNakerManchester/test

Fix some new typing errors

spynnaker/plot_utils.py

@@ -17,6 +17,8 @@
 from types import ModuleType
 from typing import Optional
 import numpy as np
+from numpy.typing import NDArray
+
 # pylint: disable=invalid-name
 plt: Optional[ModuleType]
 try:
@@ -113,6 +115,7 @@ def heat_plot(data_sets, ylabel=None, title=None):
 
 
 def _get_colour():
+    """ Yields a colour"""
     yield "b."
     yield "g."
     yield "r."
@@ -184,7 +187,7 @@ def plot_spikes(spikes, title="spikes"):
 if __name__ == "__main__":
     spike_data = np.loadtxt("spikes.csv", delimiter=',')
     plot_spikes(spike_data)
-    doubled_spike_data = np.loadtxt("spikes.csv", delimiter=',')
+    doubled_spike_data: NDArray = np.loadtxt("spikes.csv", delimiter=',')
     for _i, doubled_spike_data_i in enumerate(doubled_spike_data):
         doubled_spike_data_i[0] = doubled_spike_data[_i][0] + 5
     plot_spikes([spike_data, doubled_spike_data])

spynnaker/pyNN/models/neural_projections/connectors/all_but_me_connector.py

@@ -246,6 +246,7 @@ def gen_connector_params(
                            synapse_info.n_post_neurons)
         has_weights = int(self.__weights is not None)
         params = numpy.array([n_values, has_weights], dtype=uint32)
+        weights: NDArray[uint32]
         if self.__weights is None:
             weights = numpy.zeros(0, dtype=uint32)
         else:

spynnaker/pyNN/models/neuron/abstract_population_vertex.py

@@ -1252,6 +1252,7 @@ def get_connections_from_machine(
             return self.__connection_cache[app_edge, synapse_info]
 
         # Start with something in the list so that concatenate works
+        connections: List[ConnectionsArray]
         connections = [numpy.zeros(0, dtype=NUMPY_CONNECTORS_DTYPE)]
         progress = ProgressBar(
             len(self.machine_vertices),

spynnaker/pyNN/models/neuron/local_only/local_only_convolution.py

@@ -320,7 +320,7 @@ def get_mean_positive_weight(
         pos_weights = conn.kernel_weights[conn.kernel_weights > 0]
         if len(pos_weights) == 0:
             return 0
-        return numpy.mean(pos_weights)
+        return float(numpy.mean(pos_weights))
 
     @overrides(AbstractSupportsSignedWeights.get_mean_negative_weight)
     def get_mean_negative_weight(
@@ -329,7 +329,7 @@ def get_mean_negative_weight(
         neg_weights = conn.kernel_weights[conn.kernel_weights < 0]
         if len(neg_weights) == 0:
             return 0
-        return numpy.mean(neg_weights)
+        return float(numpy.mean(neg_weights))
 
     @overrides(AbstractSupportsSignedWeights.get_variance_positive_weight)
     def get_variance_positive_weight(
@@ -338,7 +338,7 @@ def get_variance_positive_weight(
         pos_weights = conn.kernel_weights[conn.kernel_weights > 0]
         if len(pos_weights) == 0:
             return 0
-        return numpy.var(pos_weights)
+        return float(numpy.var(pos_weights))
 
     @overrides(AbstractSupportsSignedWeights.get_variance_negative_weight)
     def get_variance_negative_weight(
@@ -347,4 +347,4 @@ def get_variance_negative_weight(
         neg_weights = conn.kernel_weights[conn.kernel_weights < 0]
         if len(neg_weights) == 0:
             return 0
-        return numpy.var(neg_weights)
+        return float(numpy.var(neg_weights))

spynnaker/pyNN/models/neuron/population_machine_neurons.py

@@ -317,14 +317,15 @@ def _write_neuron_core_parameters(
         # Write the keys
         spec.write_array(keys)
 
-    def __in_selector(self, n: int, selector: Selector) -> bool:
+    def __in_selector(
+            self, n: Union[int, numpy.integer], selector: Selector) -> bool:
         if isinstance(selector, Container):
             return n in selector
         return n == selector
 
     def _write_current_source_parameters(
             self, spec: DataSpecificationBase):
-        n_atoms = self._vertex_slice.n_atoms
+        n_atoms: int = self._vertex_slice.n_atoms
 
         spec.comment(
             f"\nWriting Current Source Parameters for {n_atoms} Neurons:\n")
@@ -350,7 +351,7 @@ def _write_current_source_parameters(
         if current_sources:
             # Array to keep track of the number of each type of current source
             # (there are four, but they are numbered 1 to 4, so five elements)
-            cs_index_array = [0, 0, 0, 0, 0]
+            cs_index_array: List[int] = [0, 0, 0, 0, 0]
 
             # Data sent to the machine will be current sources per neuron
             # This array will have the first entry indicating the number of
@@ -363,9 +364,10 @@ def _write_current_source_parameters(
                 cs_id = current_source.current_source_id
 
                 # Only use IDs that are on this core
-                for i, n in enumerate(self._vertex_slice.get_raster_ids()):
+                for i, raster_id in enumerate(
+                        self._vertex_slice.get_raster_ids()):
                     if self.__in_selector(
-                            n, current_source_id_list[current_source]):
+                            raster_id, current_source_id_list[current_source]):
                         # I think this is now right, but test it more...
                         neuron_current_sources[i][0] += 1
                         neuron_current_sources[i].append(cs_id)
@@ -378,16 +380,16 @@ def _write_current_source_parameters(
 
             # Now loop over the neurons on this core and write the current
             # source ID and index for sources attached to each neuron
-            for n in range(n_atoms):
-                n_current_sources = neuron_current_sources[n][0]
+            for atom in range(n_atoms):
+                n_current_sources = neuron_current_sources[atom][0]
                 spec.write_value(n_current_sources)
                 if n_current_sources != 0:
                     for csid in range(n_current_sources * 2):
-                        spec.write_value(neuron_current_sources[n][csid+1])
+                        spec.write_value(neuron_current_sources[atom][csid+1])
 
             # Write the number of each type of current source
-            for n in range(1, len(cs_index_array)):
-                spec.write_value(cs_index_array[n])
+            for cs_index in range(1, len(cs_index_array)):
+                spec.write_value(cs_index_array[cs_index])
 
             # Now loop over the current sources and write the data required
             # for each type of current source

spynnaker/pyNN/models/neuron/synapse_io.py

@@ -352,7 +352,9 @@ def _get_row_data(
         # Blank the plastic data
         fp_data = numpy.zeros((n_rows, 0), dtype=uint32)
         pp_data = numpy.zeros((n_rows, 0), dtype=uint32)
+        fp_size:  NDArray[uint32]
         fp_size = numpy.zeros((n_rows, 1), dtype=uint32)
+        pp_size:  NDArray[uint32]
         pp_size = numpy.zeros((n_rows, 1), dtype=uint32)
     else:
         assert isinstance(synapse_dynamics, AbstractPlasticSynapseDynamics)

spynnaker/pyNN/models/spike_source/spike_source_poisson_vertex.py

@@ -292,7 +292,8 @@ def __init__(
         self.__spike_recorder = MultiSpikeRecorder()
 
         if max_rate is None:
-            all_rates = list(_flatten(self.__data["rates"]))
+            all_rates: List[numpy.floating] = list(
+                _flatten(self.__data["rates"]))
             self.__max_rate = numpy.amax(all_rates) if all_rates else 0
         else:
             self.__max_rate = max_rate

spynnaker/pyNN/utilities/neo_buffer_database.py

@@ -20,8 +20,8 @@
 import re
 import struct
 from typing import (
-    Any, Collection, Dict, Iterable, List, Optional, Sequence, Tuple, Union,
-    TYPE_CHECKING)
+    Any, Collection, Dict, Iterable, List, Optional, Sequence, Set, Tuple,
+    TYPE_CHECKING, Union)
 
 import numpy
 from numpy import floating, integer, uint8, uint32
@@ -785,7 +785,7 @@ def __combine_indexes(
         data_set = set(data_indexes)
         indexes = [i for i in view_indexes if i in data_set]
         # check for missing and report
-        view_set = set(view_indexes)
+        view_set: Set[int] = set(view_indexes)
         missing = view_set.difference(data_indexes)
         if missing:
             logger.warning("No {} available for neurons {}",