Skip to content

Conversation

@swernerx
Copy link
Member

@swernerx swernerx commented Jan 17, 2026

🤖 I have created a release beep boop

0.3.0 (2026-01-20)

⚠ BREAKING CHANGES

  • The library now uses native C++ bindings for all array operations. The pure TypeScript implementation has been removed.

Features

  • add 18 new real-world benchmark scenarios (f99bf82)
  • add 20 new scenarios to NumPy Python benchmark (7d5345a)
  • add boolean array support and NumPy conformity testing (4c86a50)
  • add completeness check for NumPy conformity tests (29b1139)
  • add fused arithmetic operations for reduced N-API overhead (d65a182)
  • add fused C++ operations for performance optimization (01d04e9)
  • add in-place arithmetic operations (6ca7dc1)
  • add keepdims parameter to reduction operations (0d1bad6)
  • add main entry point with np namespace (fcfa5df)
  • add matrix-vector dot, norm variants, statistics, and NDArray methods (a6eeaf4)
  • add NumPy vs numpy-node benchmark comparison (b834dd7)
  • add platform-specific package distribution (60f005f)
  • add release-please for automated releases (79f0962)
  • auto-generate NumPy reference on pnpm install (21caf0e)
  • core: add NDArray class and type system (3dbc6eb)
  • creation: add array creation functions (5f62ff5)
  • export linalg module from main entry point (87a9e38)
  • implement einsum (Einstein summation) (91e6dca)
  • implement FFT module with fft, ifft, rfft, irfft (06aca8c)
  • implement NumPy-style views for transpose and reshape (25e7d2b)
  • implement Tier 3+4 NumPy functions (67ba81a)
  • linalg: add linear algebra module (2d36f0a)
  • linalg: implement eigvals, eig, svd, qr, cholesky, matrix_rank (f4cf8f9)
  • math: implement axis-based reductions for sum, mean, std (15c4b0c)
  • math: implement NumPy-style broadcasting for binary operations (86969bc)
  • native: add C++ N-API module scaffolding (1f7492d)
  • ops: add array operations (15a393f)
  • stats: add statistical functions (81f961b)
  • switch to npm trusted publishing with provenance (549fe9e)
  • use local .venv for NumPy installation (3e8e259)

Bug Fixes

  • add id-token permission to release-please workflow (2abc41b)
  • add missing cmath include and extend CI matrix (6010e20)
  • add repository field for npm provenance validation (fcf96c1)
  • also trigger publish on manual releases (a0d42d4)
  • ci: install lapack-reference separately on Windows (4c7717b)
  • ci: trigger publish directly via release event, not workflow_call (deafae0)
  • ci: upgrade npm for OIDC trusted publishing support (d0534be)
  • ci: use Node.js 24 for publish job (includes npm v11) (dc7948e)
  • correct softmax cross-entropy benchmark (d20f1c9)
  • define _USE_MATH_DEFINES for Windows M_PI (b5cb4e8)
  • install openblas with lapack feature on Windows (c78f631)
  • lockfile (c329a45)
  • M_PI must be defined before any includes on MSVC (8ac036a)
  • native: configure cmake-js and fix build errors (2929e6c)
  • ops: fix reduceAxis keepdims indexing bug (88a4c02)
  • remove darwin-x64, use macos-15 for arm64 (77072cf)
  • stats: fix average and variance functions (00e5ff4)
  • update CI to use Node 22 and refresh lockfile (15b5176)
  • update macOS runner for x64 builds (2ffb447)
  • use @numpy-node scope for platform packages (94a563f)
  • use PAT for release-please (265123a)

Performance Improvements

  • add batch_matmul for reduced N-API overhead (3ddc8fd)
  • add broadcasting optimization and fused operations (8057cae)
  • add fused ML operations and optimize benchmarks to 70% wins (9543d95)
  • add fused native operations for numerical methods (fc84fac)
  • add native percentile with O(n) quickselect algorithm (7b96aec)
  • add native percentile with optimized sort (14e94a4)
  • add vDSP-accelerated 2D-1D broadcasting for add and multiply (b6fb482)
  • add vDSP/vecLib optimizations for macOS (7c04bea)

Code Refactoring

  • replace pure TypeScript with native C++ bindings (a8fa1ce)

This PR was generated with Release Please. See documentation.

@codecov
Copy link

codecov bot commented Jan 17, 2026

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

Thanks for integrating Codecov - We've got you covered ☂️

@swernerx swernerx force-pushed the release-please--branches--main--components--numpy-node branch 15 times, most recently from 730c556 to db64270 Compare January 19, 2026 18:12
@swernerx swernerx force-pushed the release-please--branches--main--components--numpy-node branch from db64270 to 02b3d20 Compare January 20, 2026 08:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants