fix: correct bernoulli_loader kwarg, remove duplicate import, fix mutable default

[haoyu-haoyu]

Summary

• Fix bernoulli_loader reading wrong kwarg in bindsnet/encoding/loaders.py: kwargs.get("dt", 1.0) should be kwargs.get("max_prob", 1.0). The docstring documents max_prob as the parameter for maximum spike probability per Bernoulli trial, but the code was reading dt instead.

• Remove duplicate import numpy as np in bindsnet/network/monitors.py.

• Fix mutable default argument pipeline: list = [] → pipeline: Optional[list] = None in MulticompartmentConnection.__init__ (bindsnet/network/topology.py). Mutable defaults are shared across all instances.

Test plan

• bernoulli_loader now correctly reads max_prob parameter
• No behavioral change for existing code that doesn't pass max_prob (default remains 1.0)
• Existing tests should pass without modification

[Copilot]

Pull request overview

This PR fixes a parameter handling bug in the Bernoulli encoder loader, removes a redundant import in monitors, and updates a connection constructor to avoid a shared mutable default.

Changes:

• Fix bernoulli_loader to read max_prob from kwargs instead of incorrectly reading dt.
• Remove a duplicate numpy import (and duplicated adjacent imports) in bindsnet/network/monitors.py.
• Change MulticompartmentConnection.__init__ default pipeline from [] to None to avoid mutable default sharing.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
bindsnet/encoding/loaders.py	Corrects bernoulli_loader kwarg lookup for max_prob.
bindsnet/network/monitors.py	Removes duplicated import block to clean up module imports.
bindsnet/network/topology.py	Updates pipeline default to avoid shared mutable default (but currently introduces a runtime issue when pipeline is omitted).

Comments suppressed due to low confidence (1)

bindsnet/network/topology.py:417

• Changing pipeline default to None here will pass None into AbstractMulticompartmentConnection.__init__, which currently iterates for feature in pipeline: (using the argument, not self.pipeline). That will raise a TypeError when callers omit pipeline (now the default). Consider normalizing pipeline to an empty list before calling super().__init__ (e.g., pipeline = [] if pipeline is None else pipeline) or updating the base class to iterate over self.pipeline instead of the raw argument.

        pipeline: Optional[list] = None,
        manual_update: bool = False,
        traces: bool = False,
        **kwargs,
    ) -> None:

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.