Release 0.2.4

[chaoming0625]

Bumps the version to 0.2.4 and finalizes the release changelog.

What's in 0.2.4

• Control-flow online learning (Phases 0–4). ETP operations inside vmap,
  cond, scan / for_loop, and weight-free while bodies now participate in
  eligibility-trace learning, plus structured scan descent for long inner
  scans (compile size independent of loop length).
• Three new ETP operators — grouped_matmul, embedding, and einsum —
  each with a matching braintrace.nn layer (GroupedLinear, Embedding) and
  single-step BPTT-oracle coverage.
• Performance: chunk-factorized multi-step D-RTRL trace update, 2.4–4.5×
  faster on multi-step windows (chunked_trace knob).
• Compiler: deterministic hidden-group ordering across processes and
  transparent user-jax.jit inlining; directly-fed fan-out and perturbation
  fixes.
• Correctness: _op / _algorithm audit closes 24 findings (4 Critical,
  6 High, 6 Medium, 8 Minor).
• Breaking: the YW_TO_W ETP rule is renamed to DT_TO_T.
• Internal: mypy CI gate repaired; signature cleanups.

See changelog.md for the full, itemized notes.

Changes

• braintrace/_version.py: 0.2.3 → 0.2.4
• changelog.md: UNRELEASED → Version 0.2.4, with a summary intro and
  entries for every PR shipped since v0.2.3.

Summary by Sourcery

Release version 0.2.4 with expanded eligibility-trace online learning through JAX control flow and updated release notes.

New Features:

• Document new ETP operators grouped_matmul, embedding, and einsum, including their associated braintrace.nn layers.
• Document structured scan descent for long ETP scans to support Phase 4 online learning.
• Document user jax.jit inlining and deterministic compilation behavior.
• Document performance improvements from chunk-factorized multi-step D-RTRL trace updates.
• Document the ETP rule rename from YW_TO_W to DT_TO_T.

Enhancements:

• Add detailed release notes summarizing compiler robustness improvements, fan-out handling, and perturbation behavior.
• Describe internal type-checking and signature cleanup work as part of the release.

Build:

• Bump library version from 0.2.3 to 0.2.4 in braintrace/_version.py.

Documentation:

• Finalize changelog.md for Version 0.2.4 with a comprehensive summary of new features, performance, correctness fixes, breaking changes, and internal maintenance.

[sourcery-ai]

Reviewer's Guide

Release 0.2.4 bumps the library version and promotes the unreleased changelog section to a finalized 0.2.4 release entry, with detailed notes on new ETP operators, control-flow online learning, compiler determinism, performance improvements, correctness fixes, a breaking ETP rule rename, and internal maintenance work.

State diagram for changelog promotion to Version 0.2.4

stateDiagram-v2
    [*] --> UNRELEASED
    UNRELEASED --> Version_0_2_4
    Version_0_2_4 --> [*]

Loading

File-Level Changes

Change	Details	Files
Version constant updated to 0.2.4.	Change version string from 0.2.3 to 0.2.4. Keep version_info derivation logic unchanged.	braintrace/_version.py
Changelog entry for 0.2.4 finalized and moved from UNRELEASED to a structured release section.	Rename the top changelog section from UNRELEASED to Version 0.2.4. Add an introductory paragraph summarizing control-flow ETP learning, new operators, performance, compiler and correctness work, and the breaking rule rename. Add detailed subsections describing new ETP operators (grouped_matmul, embedding, einsum) and their associated layers and rules. Document structured scan descent behavior and its impact on long scans and control-flow learning phases. Describe compiler changes including user jax.jit inlining, deterministic hidden-group ordering, fan-out fixes, and perturbation robustness. Document performance improvements via chunk-factorized multi-step D-RTRL trace updates and the chunked_trace knob. Summarize results of the ETP _op / _algorithm correctness audit and associated fixes and test additions. Call out the breaking rename of the ETP rule YW_TO_W to DT_TO_T and its impact on custom primitives. Note internal maintenance: mypy CI gate repair and signature cleanups.	changelog.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 reviewed your changes and they look great!

---

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.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!