fix(math/sparse): coo_to_csr bounds check + sparse/event/delay regression suite

[chaoming0625]

Fresh review of brainpy/math sparse/event/jitconn operators + delay variables.

• Medium — coo_to_csr silently produced a corrupt CSR (indptr[-1] != nse) on out-of-range pre_ids; now raises a clear ValueError.
• Added a thorough regression suite verifying CSR/COO/event matvec & matmat and jit-connectivity ops against dense references (including transpose=True and autodiff), and TimeDelay/LengthDelay ring-buffer modulo-wrap behavior. Prior-audit findings (csrmm transpose, TimeDelay modulo, coomv) verified already-fixed and now locked by tests.

In-scope: 58 passed. Findings: docs/issues-found-20260619-math-sparse-event.md.

Summary by Sourcery

Tighten sparse COO-to-CSR conversion error handling and add regression coverage for sparse/event operations and delay variables.

Bug Fixes:

• Raise a clear ValueError in coo_to_csr when pre_ids fall outside [0, num_row) to avoid generating structurally invalid CSR matrices.

Enhancements:

• Document coo_to_csr behaviour and constraints, including its eager-only nature due to jnp.unique.

Documentation:

• Document the results of the math/sparse, math/event, jit connectivity, and delay variable audit in a new issues-found report.

Tests:

• Add regression tests for coo_to_csr, csr_to_coo, and csr_to_dense against dense references.
• Add regression tests for CSR matrix-matrix multiplication, COO matrix-vector multiplication, and event-driven CSR matmat to match dense and autodiff behaviour, including transpose variants.
• Add regression tests for pre_syn_post helpers to validate event routing and per-group reductions.
• Add regression tests for TimeDelay and LengthDelay ring-buffer wrap-around behaviour to lock in correct modulo/rotation semantics.

[sourcery-ai]

Reviewer's Guide

Adds an eager bounds check to sparse.coo_to_csr to prevent structurally invalid CSR output on out-of-range pre_ids, and introduces a focused regression test suite covering sparse COO/CSR helpers, event-driven CSR matvec/matmat paths (including transpose and autodiff), pre_syn_post routing utilities, and TimeDelay/LengthDelay ring-buffer wrap behaviour. The PR also records the sparse/event audit findings in docs/issues-found-20260619-math-sparse-event.md.

File-Level Changes

Change	Details	Files
Harden coo_to_csr against out-of-range pre_ids and improve its documentation.	Expand coo_to_csr docstring with detailed argument/return semantics, error conditions, and a note that it is an eager (non-jittable) helper. Convert pre_ids/post_ids to JAX arrays up front, then compute min/max of pre_ids when non-empty and raise a ValueError if any pre_id is negative or >= num_row. Keep the existing sorting/CSR construction logic intact so that valid inputs now yield CSR with indptr[-1] == len(indices) and int32 row pointers.	brainpy/math/sparse/utils.py
Add regression tests for sparse COO/CSR utilities and matvec/matmat paths, including transpose and autodiff.	Introduce utils_test covering coo_to_csr roundtrips (including unsorted COO, empty rows, and out-of-range/negative pre_ids), csr_to_coo roundtrips, and csr_to_dense vs dense numpy references. Add csr_mm_test to validate csrmm forward results for transpose=False/True against dense references and to check gradients of the transpose=True path w.r.t. CSR values. Add coo_mv_test to exercise coomv on unsorted COO triples for both orientations, scalar-weight broadcasting, and gradient of the scalar-weight path. Add event csr_matmat tests to compare event.csrmm (binary CSR matmat) against dense references for both orientations and consistency with the float CSR path.	brainpy/math/sparse/utils_test.py brainpy/math/sparse/csr_mm_test.py brainpy/math/sparse/coo_mv_test.py brainpy/math/event/csr_matmat_test.py
Add regression tests for pre_syn_post routing and delay variable ring-buffer behaviour.	Add pre_syn_post_test to cover pre2post_event_sum (scalar and per-synapse values, and dense transpose reference), syn2post_sum, syn2post_mean handling of empty groups (no NaNs), and syn2post_softmax per-group normalization. Extend delayvars_coverage_test with ring-buffer regression tests: TimeDelay modulo-based wrap-around on the true-branch (exact-step, no interpolation) and LengthDelay ramp behaviour for both ROTATE_UPDATE and CONCAT_UPDATE update methods.	brainpy/math/pre_syn_post_test.py brainpy/math/delayvars_coverage_test.py
Record the sparse/event/delay audit scope, findings, and status in documentation.	Add an issues-found audit document describing the reviewed modules, summarizing correctness of sparse/event/jitconn operators against dense references, detailing the Medium coo_to_csr bounds bug fix, and listing remaining low-priority observations with their status. Cross-reference prior audit IDs and mark them as already fixed and re-verified in this branch.	docs/issues-found-20260619-math-sparse-event.md

---

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 found 2 issues

Prompt for AI Agents

Please address the comments from this code review:

## Individual Comments

### Comment 1
<location path="brainpy/math/sparse/utils_test.py" line_range="21" />
<code_context>
+# coo_to_csr
+# ---------------------------------------------------------------------------
+
+def test_coo_to_csr_valid_roundtrip():
+    # 3 x 4 matrix:
+    #   [[0, 2, 0, 4],
</code_context>
<issue_to_address>
**suggestion (testing):** Add an explicit test for the empty-COO case in `coo_to_csr`.

There’s no test case for empty `pre_ids`/`post_ids` (e.g. `num_row > 0` with `nse == 0`), which is easy to implement incorrectly (e.g. `indptr` length/dtype or `indptr[-1]` not zero). Please add a test (e.g. `test_coo_to_csr_empty_matrix`) using `pre = post = np.array([], dtype=int)`, `num_row = 3`, and asserting that `indices` is empty and `indptr` is `np.array([0, 0, 0, 0], dtype=np.int32)` to cover structurally empty matrices.
</issue_to_address>

### Comment 2
<location path="brainpy/math/sparse/utils_test.py" line_range="85" />
<code_context>
+    return np.asarray(indptr), np.asarray(indices), np.asarray(order)
+
+
+def test_csr_to_coo_roundtrip():
+    rows = np.array([0, 0, 1, 2, 2])
+    cols = np.array([1, 3, 0, 1, 3])
</code_context>
<issue_to_address>
**suggestion (testing):** Consider adding a test for `csr_to_coo` with a completely empty CSR (no non-zeros).

You already cover a standard roundtrip and CSR with empty rows. It would be helpful to add a case where the CSR represents an all-zero matrix, e.g. `indptr = np.zeros(num_rows + 1, dtype=int)` and empty `indices`. A test like `test_csr_to_coo_all_zero_matrix` can assert `csr_to_coo` returns empty `rows`/`cols` with the correct dtype/shape and handles this corner case without errors.
</issue_to_address>

---

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.}

[sourcery-ai]

suggestion (testing): Add an explicit test for the empty-COO case in coo_to_csr.

There’s no test case for empty pre_ids/post_ids (e.g. num_row > 0 with nse == 0), which is easy to implement incorrectly (e.g. indptr length/dtype or indptr[-1] not zero). Please add a test (e.g. test_coo_to_csr_empty_matrix) using pre = post = np.array([], dtype=int), num_row = 3, and asserting that indices is empty and indptr is np.array([0, 0, 0, 0], dtype=np.int32) to cover structurally empty matrices.

[sourcery-ai]

suggestion (testing): Consider adding a test for csr_to_coo with a completely empty CSR (no non-zeros).

You already cover a standard roundtrip and CSR with empty rows. It would be helpful to add a case where the CSR represents an all-zero matrix, e.g. indptr = np.zeros(num_rows + 1, dtype=int) and empty indices. A test like test_csr_to_coo_all_zero_matrix can assert csr_to_coo returns empty rows/cols with the correct dtype/shape and handles this corner case without errors.