v0.1.6 (Alpha)

Release Date: 2026-07-19

Important Note

Final v0.1.x release. After this, the next release jumps to v0.2.0 to signal the deprecation shim put in place for --full flag, which will be replaced by --verbose in a later version

What’s Changed

Highlights

Benchmarking mir_eval runner added

  • PR #273 by nanaoto — IDMT-SMT-Drums V2 benchmark runner with mir_eval scaffolding:

    • benchmarks/run.py entrypoint with dataset adapter dispatch, evaluation loop, CSV/JSON archive with git commit tracking

    • drumscript/datasets/ package: BenchmarkItem dataclass and IDMT adapter (XML/SVL annotation parsing)

    • benchmarks/README.md documenting conventions, dataset setup, and planned dataset coverage

    • Unit tests for benchmark runner (test_benchmarks_run.py) and IDMT dataset adapter (test_idmt_dataset.py)

    • mir_eval added as a dev dependency

  • New unit test file tests/unit/test_deprecation_warnings.py (13 tests)

  • New unit test file tests/unit/test_cli_args.py (4 tests)

Fixed

  • Deprecation shim for full parameter on the Python API:

    • New verbose parameter added to transcribe(), extract_stems(), and detect_tempo() as the canonical replacement for full

    • Passing full=True (or full=False) continues to work but emits a DeprecationWarning directing users to verbose

    • Warning explicitly names the v1.0.0 (beta) removal target so users have a clear migration timeline

    • Passing both full and verbose together raises TypeError (ambiguous, almost certainly a bug)

    • Internal helper _resolve_verbose_flag() centralises the resolution logic so it cannot drift between wrappers

    • Docstrings updated to mark verbose as primary and full as deprecated

    • Flag inconsistency between drumscript/main.py and drumscript/__init__.py resolved: argparse in main block updated so the CLI flag is now --full-song (hyphenated, consistent with --all-stems). Auto-converts to args.full_song matching the Python API parameter name.

Contributors

  • @victoria-mckinney

  • @nanaoto

A special thank you to @nanaoto for integrating mir_eval into DrumScript! 🥁