Skip to content

v1.0.1 - ORBIT - Public Facing Improvements, Code Clean Up, and Documentation

Choose a tag to compare

View all tags
@focusjordan focusjordan released this 02 Jun 13:38
· 85 commits to main since this release
Immutable release. Only release title and notes can be modified.
v1.0.1
cdc973c

What's Changed in v1.0.1

This release unifies licensing across the monorepo under Apache 2.0, delivers a comprehensive documentation overhaul for all sub-packages, and introduces critical CI/CD pipeline optimizations.

🚀 Features & Enhancements

  • Package Documentation Overhaul: Completely rewrote package-level documentation for all 5 monorepo packages to feature comprehensive API parameter details, return JSON schemas, architecture flow diagrams, and copy-pasteable Node.js usage guides.
    • @ohnrshyp/dsp (includes Krumhansl-Schmuckler musical key detection math)
    • @ohnrshyp/forensics (includes details and interpretation matrices for spectral & upsampling check routines)
    • @ohnrshyp/ledger (includes Ed25519 cryptography, CBOR serialization payloads, and pgvector schemas)
    • @ohnrshyp/metadata (includes CLAP/PANNs/wav2vec2 hybrid neural details)
    • @ohnrshyp/watermark (includes SilentCipher and Spread Spectrum parameters)

⚖️ Licensing

  • Apache-2.0 Unification: Resolved package licensing inconsistencies. Replaced leftover instances of MIT and ISC licenses inside package.json configurations (for SDK, CLI, and sub-packages) and pyproject.toml configurations (for DSP, Forensics, and Watermark packages) to guarantee a consistent legal model.
  • Updated llms.txt and root/SDK READMEs to align with Apache-2.0 guidelines.

🔧 CI/CD & Testing

  • CI Workflow Fixes: Resolved background service failures in GitHub Actions by unifying database-dependent server launches, health checks, and test runs into single workflow steps to prevent background process reaping.
  • Dynamic Audio Fixtures: Automated the generation of missing test audio assets (e.g., MP3 and WAV tones) using ffmpeg inside the CI runner on the fly.
  • Heavy Test Isolation: Moved resource-heavy neural Metadata Extractor tests to a dedicated GPU-only suite to prevent standard CPU-bound CI runners from experiencing timeouts or Out-Of-Memory (OOM) errors.
  • Added Python 3.10 and dependency installations to the ML workflow environment to enable seamless Librosa tests.
Assets 3
Loading