fix: IntegratorRunner dict-monitor regression + stale post-audit test updates

[chaoming0625]

Summary

Post-integration follow-up to the 16-part audit (PRs #838–#853). Fixes one
real regression surfaced by the full-suite run and re-points five stale
bug-pinning tests at the now-corrected behaviour.

Source fix

• brainpy/integrators/runner.py — IntegratorRunner dict monitors
  regressed: after _format_dict_monitors each value is a
  (var_name_or_Variable, index) tuple, but resolution tested
  isinstance(i, str) on the whole tuple (always False), so a string
  variable-name leaked through to the base runner, which then looked it up
  as an attribute of self.target and raised. Integrator state variables
  live in self.variables; resolve string names there and pass the base
  runner an already-resolved (Variable, index) pair.

Test updates (behaviour now correct, tests re-pointed)

• nvar_coverage_test — order < 2 now raises ValueError (not AssertionError).
• dnn_toolbox_fixes_test — Adan now constructs and updates cleanly.
• boost_misc_test — offline ElasticNet / Logistic regression now fit & predict.
• train_analysis_glue_fixes_test — logistic regression now fits without error.

Test plan

• 11 previously-flipped tests now pass (6 IntegratorRunner + 5 re-pointed).
• Sanity: IntegratorRunner dict monitors and DSRunner dict monitors both run green.

Summary by Sourcery

Fix IntegratorRunner dictionary monitor handling and update tests to reflect corrected optimizer and offline regression behavior.

Bug Fixes:

• Resolve IntegratorRunner dict monitor regression by correctly mapping string variable names to integrator state variables before passing them to the base runner.
• Update NVAR error expectation to match the current ValueError raised for invalid order values.
• Treat Adan optimizer update as working behavior instead of a known bug by exercising a full construct-and-update cycle without errors.
• Treat offline ElasticNet regression predict as successful behavior instead of a broken path by asserting valid predictions.
• Treat offline LogisticRegression fit as successful behavior instead of an IndexError in both general and glue tests, asserting finite fitted parameters.

Tests:

• Refresh existing tests to assert successful Adan updates, ElasticNet prediction, and LogisticRegression fitting, ensuring outputs are finite and shapes are correct.
• Adjust NVAR error tests to expect ValueError for invalid order rather than AssertionError.

[sourcery-ai]

Reviewer's Guide

Fixes an IntegratorRunner dict-monitor regression by correctly resolving string variable names to integrator state variables, and updates several tests that previously pinned buggy behavior so they now assert on the corrected, successful paths for ElasticNet/Logistic regression, Adan optimizer updates, and NVAR error typing.

Sequence diagram for IntegratorRunner dict monitor resolution fix

sequenceDiagram
  actor User
  participant IntegratorRunner
  participant Variables
  participant BaseRunner

  User->>IntegratorRunner: __init__(target, monitors)
  alt monitors is dict
    IntegratorRunner->>IntegratorRunner: _format_dict_monitors(monitors)
    loop each_monitor_entry
      IntegratorRunner->>Variables: variables[var_name]
      Variables-->>IntegratorRunner: Variable
      IntegratorRunner->>IntegratorRunner: build (Variable, index) pair
    end
  end
  IntegratorRunner->>BaseRunner: __init__(target, resolved_monitors)

Loading

File-Level Changes

Change	Details	Files
Fix IntegratorRunner dict-monitor handling so string-based monitor specs resolve to integrator state variables instead of leaking to the base runner and crashing.	Adjust dict-monitor normalization to treat values as (var_name_or_Variable, index) tuples rather than single objects. Resolve string variable names against self.variables before passing monitors to the base runner. Preserve already-resolved (Variable, index) pairs by leaving non-string-first elements untouched.	brainpy/integrators/runner.py
Update error expectation for invalid NVAR order to match new ValueError-based validation.	Change test to expect ValueError instead of AssertionError when NVAR(order<2) is constructed. Document that NVAR.init now uses check.is_integer(min_bound=2) which raises ValueError on invalid order.	brainpy/dyn/rates/nvar_coverage_test.py
Repoint offline ElasticNet and Logistic regression tests from pinning known crashes to asserting successful fit/predict behavior.	Rename ElasticNet test and update its docstring to describe the now-consistent call/predict feature handling. Replace expect-raises around ElasticNetRegression.predict with shape checks on successful predictions. Rename LogisticRegression tests and update docstrings to describe the fixed 1-D target handling. Replace expect-raises around LogisticRegression.call with assertions that fitting succeeds and parameters are finite.	brainpy/boost_misc_test.py brainpy/train_analysis_glue_fixes_test.py
Update Adan optimizer test to assert that update runs successfully and keeps parameters finite instead of pinning a TypeError from lax.cond.	Rename Adan test and update its docstring to describe the fixed lax.cond operand handling. Remove the TypeError expectation and instead perform a real update call. Assert that the updated variable contains no NaNs after the optimizer step.	brainpy/dnn/dnn_toolbox_fixes_test.py

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've left some high level feedback:

• In IntegratorRunner.__init__, the dict-monitor normalization comprehension relies on tuple indexing (v[0], v[1]) and silently passes through non-tuple values; consider unpacking with pattern matching or explicit validation (e.g., for k, (var, idx) in monitors.items()) to make the expected structure clearer and fail fast on malformed monitor specs.
• The updated test docstrings embed audit dates and process context (e.g., "Fixed in audit 2026-06-19"), which can quickly go stale; consider trimming these to describe only the expected behavior and failure mode so the tests remain accurate without historical context.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In `IntegratorRunner.__init__`, the dict-monitor normalization comprehension relies on tuple indexing (`v[0], v[1]`) and silently passes through non-tuple values; consider unpacking with pattern matching or explicit validation (e.g., `for k, (var, idx) in monitors.items()`) to make the expected structure clearer and fail fast on malformed monitor specs.
- The updated test docstrings embed audit dates and process context (e.g., "Fixed in audit 2026-06-19"), which can quickly go stale; consider trimming these to describe only the expected behavior and failure mode so the tests remain accurate without historical context.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}