drumscript 0.1.4__tar.gz → 0.1.6__tar.gz

{drumscript-0.1.4 → drumscript-0.1.6}/DrumScript.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: drumscript
-Version: 0.1.4
+Version: 0.1.6
 Summary: A Python package to convert drum audio to sheet music.
 Author-email: drumscript-admin <hello.drumscript@gmail.com>
 License-Expression: Apache-2.0
@@ -50,10 +50,10 @@ Provides-Extra: dev
 Requires-Dist: ipykernel; extra == "dev"
 Requires-Dist: myst-parser; extra == "dev"
 Requires-Dist: myst-nb>=1.3.0; extra == "dev"
+Requires-Dist: mir-eval>=0.8.2; extra == "dev"
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-cov>=7.1.0; extra == "dev"
 Requires-Dist: shibuya>=2025.10.21; extra == "dev"
-Requires-Dist: fastapi>=0.128.8; extra == "dev"
 Requires-Dist: ipywidgets>=8.1.8; extra == "dev"
 Requires-Dist: requests>=2.32.5; extra == "dev"
 Requires-Dist: ruff>=0.15.11; extra == "dev"
@@ -62,14 +62,48 @@ Dynamic: license-file
 # **`DrumScript`**
 
 <!--date_created: sun-15-june-2025-->
-<!--date_edited: tues-19-may-2026--->
+<!--date_edited: thurs-18-june-2026--->
 
-**DrumScript** is an open-source Python library and CLI tool for drum audio analysis and transcription. Give it a recording — a full mix or an isolated drum stem — and it will generate PDF sheet music, MIDI files, and MusicXML output.
+**Workflow Status**
+
+[![Run Tests](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml)
+
+**Demo Notebooks**
+
+[![Try DrumScript in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/DrumScript/DrumScript/blob/main/docs/guide/interactive/drumscript_interactive_notebook.ipynb)
 
-> **[hello.drumscript@gmail.com](mailto:hello.drumscript@gmail.com)**
+**DrumScript** is an open-source Python library and CLI tool for drum audio analysis and transcription. Give it a recording — a full mix or an isolated drum stem — and it will generate PDF sheet music, MIDI files, and MusicXML output. The `DrumScript` model is a **deterministic classifier**, and doesn't use AI/machine learning. Built for drummers and by drummers, it is - and always will be - an open-source community tool.
 
 > **Python >=3.9**
 
+> **[Documentation](https://drumscript.github.io/DrumScript/)**
+
+**Public Alpha (v0.1.4+) — June to August 2026**
+
+ - We're looking for early adopters and feedback
+ - [Feedback on the classification model](https://github.com/DrumScript/DrumScript/issues), and help shape v1.0.0.
+ - In particular we are interested in hearing from everyone:: drummers (coding not required!), sound engineers and academics in Music Information Retrieval with an interest in deterministic drum/percussion classifications. 
+ - For beta release, we are planning to (amongst other things) improve the classification model, fix any user-suggested bugs, implement user-suggested feature requests and **most importantly** build a **WebGPU/ONNX/WASM UI** that will be free to use for all.
+
+> See the **[Roadmap](https://drumscript.github.io/DrumScript/guide/roadmap.html)** for what's planned.
+
+**What it looks like**
+
+<!-- TODO: Replace with a GIF showing terminal output-->
+<!-- For now, this shows the PDF transcription output -->
+
+*Input: audio recording → Output: drum notation (PDF).
+
+**Example 1: Simple groove**
+
+![DrumScript transcription output](./docs/_static/test_wav.png)
+
+**Example 2: A well-known Sabbath song**
+
+![DrumScript transcription output](./docs/_static/iron_man_1.png)
+![DrumScript transcription output](./docs/_static/iron_man_2.png)
+![DrumScript transcription output](./docs/_static/iron_man_3.png)
+
 ---
 
 - **[Features](#features)**
@@ -78,14 +112,7 @@ Dynamic: license-file
 - **[CLI Usage](#cli-usage)**
 - **[Contributing](#contributing)**
 - **[FAQs](#faqs)**
-
-**Workflow Status**
-
-[![Run Tests](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml)
-
-**Demo Notebooks**
-
-[![Try DrumScript in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1eDVXc3d6ezmorxINOjzldRPSC3emTl2I)
+- **[Similar projects](#similar-projects)**
 
 ---
 
@@ -115,9 +142,11 @@ DrumScript/
 │   ├── audio_processor/        # Audio loading, DSP, stem splitting
 │   ├── drum_classifier/        # Rule-based classification engine
 │   ├── notation_generator/     # Score building, PDF/MIDI/XML export
+│   ├── datasets/               # Benchmark dataset adapters (IDMT, etc.)
 │   └── utils/                  # Helpers (ffmpeg installer, research scripts)
+├── benchmarks/                 # Evaluation runners (see benchmarks/README.md)
 ├── docs/                       # Sphinx documentation
-├── tests/                      # pytest test suite
+├── tests/                      # pytest test suite (131 unit + 8 integration)
 ├── .github/workflows/          # CI/CD (tests, build, publish, docs)
 ├── pyproject.toml              # Package metadata and dependencies
 └── uv.lock                     # Pinned dependency versions
@@ -128,13 +157,13 @@ DrumScript/
 
 **For users:**
 
-```bash
+```zsh
 pip install drumscript
 ```
 
 **For developers:**
 
-```bash
+```zsh
 git clone https://github.com/DrumScript/DrumScript.git
 cd DrumScript
 uv sync # this will create a .venv
@@ -157,10 +186,20 @@ DrumScript manages all dependencies via [`pyproject.toml`](pyproject.toml) using
   - Ubuntu/Debian: `sudo apt-get install libportaudio2`
   - Windows: Usually bundled with the `sounddevice` wheel.
 
+- **git-lfs** is **only** required if you want to run the documentation notebooks locally or rebuild the docs site. Some example audio files in `docs/guide/interactive/audio/` are tracked via Git LFS to keep the main repo lightweight. `pip install drumscript` and ordinary use of the package do **not** need it. If you skip this step, `git clone` will still succeed — you'll just get small LFS pointer files in place of the example audio.
+  - macOS: `brew install git-lfs`
+  - Ubuntu/Debian: `sudo apt-get install git-lfs`
+  - Windows: [Download from git-lfs.com](https://git-lfs.com/) or install via `winget install GitHub.GitLFS`.
+  - After installing, run `git lfs install` once, then `git lfs pull` inside the cloned repo to fetch the audio.
+
 ---
 
 ## Quick Start
 
+> Please note: DrumScript assumes you are providing **drum audio-only inputs by default** 
+
+> If you are using transcription with full song use the `full_song=True` flag, ie 
+
 ### End-to-end transcription
 
 ```python
@@ -170,10 +209,11 @@ import drumscript as ds
 pdf_path = ds.transcribe("drum_audio.wav")
 
 # Transcribe a full song (separates drums automatically)
-pdf_path = ds.transcribe("full_song.mp3", full_song=True)
+pdf_path = ds.transcribe("full_song.mp3") # drum only audio
+pdf_path = ds.transcribe("full_song.mp3", full_song=True) # full song, tells DrumScript to extract the drums first and then transcribe
 
 # Get all intermediate results
-result = ds.transcribe("drum_audio.wav", full=True)
+result = ds.transcribe("drum_audio.wav", verbose=True)
 print(f"Tempo: {result['tempo']:.1f} BPM")
 print(f"Events: {len(result['events'])}")
 ```
@@ -205,11 +245,34 @@ results = ds.extract_stems(
     "full_song.mp3",
     drumless=True,
     output_format="mp3",
-    full=True,
+    verbose=True,
 )
 print(f"Backing track: {results['mix']}")
 ```
 
+---
+
+## Benchmarking
+
+DrumScript includes a benchmarking framework for evaluating the classifier against standard ADT datasets using [`mir_eval`](https://github.com/mir-evaluation/mir_eval). Currently supports IDMT-SMT-Drums V2.
+
+```zsh
+# Install dev dependencies (includes mir_eval)
+uv sync --extra dev
+
+# Run the IDMT benchmark
+uv run --extra dev python benchmarks/run.py idmt \
+  --root /path/to/IDMT-SMT-DRUMS-V2
+
+# Run on a single subset with a limit
+uv run --extra dev python benchmarks/run.py idmt \
+  --root /path/to/IDMT-SMT-DRUMS-V2 \
+  --subset RealDrum --limit 5
+```
+
+Results are archived to `outputs/benchmarks/idmt/` with per-file metrics, summary statistics, and git commit tracking for reproducibility. See [`benchmarks/README.md`](benchmarks/README.md) for dataset setup and full usage.
+
+
 ---
 
 ## CLI Usage
@@ -218,29 +281,29 @@ DrumScript also provides a command-line interface.
 
 ### Basic transcription (isolated drum stem)
 
-```bash
+```zsh
 drumscript drum_audio.wav
 ```
 
 ### Full song transcription (auto-separates drums)
 
-```bash
-drumscript full_song.mp3 --full
+```zsh
+drumscript full_song.mp3 --full-song
 ```
 
 ### Extract a drumless backing track
 
-```bash
+```zsh
 drumscript full_song.mp3 --drumless
 ```
 
 ### All options
 
-```bash
+```zsh
 drumscript <audio_file> [OPTIONS]
 
 Options:
-  --full          Transcribe a full song (isolates drums first via Demucs)
+  --full-song     Transcribe a full song (isolates drums first via Demucs)
   --drumless      Extract a drumless backing track
   --mute STEM     Mute a specific stem (e.g. --mute bass). Repeatable.
   --all-stems     Export all individual stems (drums, bass, vocals, other)
@@ -251,7 +314,7 @@ Options:
 
 ### Examples
 
-```bash
+```zsh
 # Transcribe with 6/8 time signature
 drumscript drum_audio.wav --ts 6/8
 
@@ -273,6 +336,28 @@ We welcome contributions! DrumScript is intended to be a community-owned project
 
 > All bug reports and feature requests must be filed as GitHub Issues. All code changes must be submitted as Pull Requests. Keeping discussion public helps everyone.
 
+**[hello.drumscript@gmail.com](mailto:hello.drumscript@gmail.com)**
+
+## Alpha Priorities (v0.0.4 < v1.0.0) 
+The alpha phase runs between 01 June and 31 August 2026
+
+**What works today:**
+
+- End-to-end transcription pipeline: audio → onsets → classification → PDF / MIDI / MusicXML
+- Tempo detection via spectral onset envelope
+- Stem separation using Demucs (`htdemucs` 4-stem model)
+- Drumless backing track generation
+- CLI and Python API
+
+**What we're focused on during the alpha:**
+
+- Expanding test coverage across genres, kit types, and recording conditions
+- Fixing classification edge cases (deep snares vs clicky kicks, splash cymbals vs open hats)
+- Improving onset detection sensitivity for ghost notes and fast passages
+- Stabilising the public API ahead of the beta freeze
+- Community feedback collection
+
+
 ---
 
 ## FAQs
@@ -297,8 +382,9 @@ DrumScript's own classification engine is **fully deterministic** — it uses ph
 
 ## Acknowledgements
 
-- **[Demucs](https://github.com/adefossez/demucs)** — The stem splitting functionality is built upon the work of [@adefossez](https://github.com/adefossez).
-- **[librosa](https://librosa.org/)** — For foundational audio processing tools.
+1. **[Demucs](https://github.com/adefossez/demucs)** — The stem splitting functionality is built upon the work of [@adefossez](https://github.com/adefossez).
+2. **[librosa](https://librosa.org/)** — For foundational audio processing tools.
+3. **[@nanaoto](https://github.com/nanaoto)** — For building the `mir_eval` benchmarking infrastructure and IDMT-SMT-Drums V2 adapter (PR [#273](https://github.com/DrumScript/DrumScript/pull/273)).
 
 ---
 
@@ -306,6 +392,18 @@ DrumScript's own classification engine is **fully deterministic** — it uses ph
 
 [Apache License 2.0](LICENSE)
 
+---
+## Similar Projects
+
+DrumScript has no affiliation with any of the projects below. They are listed for context and reference.
+
+* **[librosa](https://librosa.org/)** — The spectral analysis library that powers DrumScript's onset detection and feature extraction.
+* **[Demucs](https://github.com/adefossez/demucs)** — The stem separation model we use for isolating drums from full mixes.
+* **[tepreece/drumscript (Golang)](https://github.com/tepreece/drumscript)** — A `(Go)lang` MIDI drum pattern scripting language by Tom Preece. Different use case (composing drum patterns via script), different technology (MIDI output rather than audio transcription). If you're looking to *write* drum patterns programmatically, check it out. Maintained by [@tepreece](https://github.com/tepreece)
+* **[basic-pitch](https://github.com/spotify/basic-pitch)** — A lightweight yet powerful audio-to-MIDI converter with pitch bend detection (better for non-percussive audio). Maintained by Spotify.
+* **[mir_eval](https://github.com/mir-evaluation/mir_eval)** — Standard evaluation metrics for music information retrieval tasks.
+* **[onset_db](https://github.com/CPJKU/onset_db)** - Provides a dataset of annotated musical onsets for tuning and evaluating audio detection algorithms. Maintained by JKU Linz.
+
 ---
 
 <!--END-->

{drumscript-0.1.4 → drumscript-0.1.6}/DrumScript.egg-info/SOURCES.txt

@@ -23,6 +23,9 @@ drumscript/audio_processor/onset_detector.py
 drumscript/audio_processor/stem_splitter.py
 drumscript/audio_processor/tempo_detector.py
 drumscript/audio_processor/tempogram.py
+drumscript/datasets/__init__.py
+drumscript/datasets/base.py
+drumscript/datasets/idmt.py
 drumscript/drum_classifier/__init__.py
 drumscript/drum_classifier/classify.py
 drumscript/notation_generator/__init__.py

{drumscript-0.1.4 → drumscript-0.1.6}/DrumScript.egg-info/requires.txt

@@ -16,10 +16,10 @@ torchaudio<2.9,>=2.4
 ipykernel
 myst-parser
 myst-nb>=1.3.0
+mir-eval>=0.8.2
 pytest
 pytest-cov>=7.1.0
 shibuya>=2025.10.21
-fastapi>=0.128.8
 ipywidgets>=8.1.8
 requests>=2.32.5
 ruff>=0.15.11

{drumscript-0.1.4 → drumscript-0.1.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: drumscript
-Version: 0.1.4
+Version: 0.1.6
 Summary: A Python package to convert drum audio to sheet music.
 Author-email: drumscript-admin <hello.drumscript@gmail.com>
 License-Expression: Apache-2.0
@@ -50,10 +50,10 @@ Provides-Extra: dev
 Requires-Dist: ipykernel; extra == "dev"
 Requires-Dist: myst-parser; extra == "dev"
 Requires-Dist: myst-nb>=1.3.0; extra == "dev"
+Requires-Dist: mir-eval>=0.8.2; extra == "dev"
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-cov>=7.1.0; extra == "dev"
 Requires-Dist: shibuya>=2025.10.21; extra == "dev"
-Requires-Dist: fastapi>=0.128.8; extra == "dev"
 Requires-Dist: ipywidgets>=8.1.8; extra == "dev"
 Requires-Dist: requests>=2.32.5; extra == "dev"
 Requires-Dist: ruff>=0.15.11; extra == "dev"
@@ -62,14 +62,48 @@ Dynamic: license-file
 # **`DrumScript`**
 
 <!--date_created: sun-15-june-2025-->
-<!--date_edited: tues-19-may-2026--->
+<!--date_edited: thurs-18-june-2026--->
 
-**DrumScript** is an open-source Python library and CLI tool for drum audio analysis and transcription. Give it a recording — a full mix or an isolated drum stem — and it will generate PDF sheet music, MIDI files, and MusicXML output.
+**Workflow Status**
+
+[![Run Tests](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml)
+
+**Demo Notebooks**
+
+[![Try DrumScript in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/DrumScript/DrumScript/blob/main/docs/guide/interactive/drumscript_interactive_notebook.ipynb)
 
-> **[hello.drumscript@gmail.com](mailto:hello.drumscript@gmail.com)**
+**DrumScript** is an open-source Python library and CLI tool for drum audio analysis and transcription. Give it a recording — a full mix or an isolated drum stem — and it will generate PDF sheet music, MIDI files, and MusicXML output. The `DrumScript` model is a **deterministic classifier**, and doesn't use AI/machine learning. Built for drummers and by drummers, it is - and always will be - an open-source community tool.
 
 > **Python >=3.9**
 
+> **[Documentation](https://drumscript.github.io/DrumScript/)**
+
+**Public Alpha (v0.1.4+) — June to August 2026**
+
+ - We're looking for early adopters and feedback
+ - [Feedback on the classification model](https://github.com/DrumScript/DrumScript/issues), and help shape v1.0.0.
+ - In particular we are interested in hearing from everyone:: drummers (coding not required!), sound engineers and academics in Music Information Retrieval with an interest in deterministic drum/percussion classifications. 
+ - For beta release, we are planning to (amongst other things) improve the classification model, fix any user-suggested bugs, implement user-suggested feature requests and **most importantly** build a **WebGPU/ONNX/WASM UI** that will be free to use for all.
+
+> See the **[Roadmap](https://drumscript.github.io/DrumScript/guide/roadmap.html)** for what's planned.
+
+**What it looks like**
+
+<!-- TODO: Replace with a GIF showing terminal output-->
+<!-- For now, this shows the PDF transcription output -->
+
+*Input: audio recording → Output: drum notation (PDF).
+
+**Example 1: Simple groove**
+
+![DrumScript transcription output](./docs/_static/test_wav.png)
+
+**Example 2: A well-known Sabbath song**
+
+![DrumScript transcription output](./docs/_static/iron_man_1.png)
+![DrumScript transcription output](./docs/_static/iron_man_2.png)
+![DrumScript transcription output](./docs/_static/iron_man_3.png)
+
 ---
 
 - **[Features](#features)**
@@ -78,14 +112,7 @@ Dynamic: license-file
 - **[CLI Usage](#cli-usage)**
 - **[Contributing](#contributing)**
 - **[FAQs](#faqs)**
-
-**Workflow Status**
-
-[![Run Tests](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/DrumScript/DrumScript/actions/workflows/tests.yml)
-
-**Demo Notebooks**
-
-[![Try DrumScript in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1eDVXc3d6ezmorxINOjzldRPSC3emTl2I)
+- **[Similar projects](#similar-projects)**
 
 ---
 
@@ -115,9 +142,11 @@ DrumScript/
 │   ├── audio_processor/        # Audio loading, DSP, stem splitting
 │   ├── drum_classifier/        # Rule-based classification engine
 │   ├── notation_generator/     # Score building, PDF/MIDI/XML export
+│   ├── datasets/               # Benchmark dataset adapters (IDMT, etc.)
 │   └── utils/                  # Helpers (ffmpeg installer, research scripts)
+├── benchmarks/                 # Evaluation runners (see benchmarks/README.md)
 ├── docs/                       # Sphinx documentation
-├── tests/                      # pytest test suite
+├── tests/                      # pytest test suite (131 unit + 8 integration)
 ├── .github/workflows/          # CI/CD (tests, build, publish, docs)
 ├── pyproject.toml              # Package metadata and dependencies
 └── uv.lock                     # Pinned dependency versions
@@ -128,13 +157,13 @@ DrumScript/
 
 **For users:**
 
-```bash
+```zsh
 pip install drumscript
 ```
 
 **For developers:**
 
-```bash
+```zsh
 git clone https://github.com/DrumScript/DrumScript.git
 cd DrumScript
 uv sync # this will create a .venv
@@ -157,10 +186,20 @@ DrumScript manages all dependencies via [`pyproject.toml`](pyproject.toml) using
   - Ubuntu/Debian: `sudo apt-get install libportaudio2`
   - Windows: Usually bundled with the `sounddevice` wheel.
 
+- **git-lfs** is **only** required if you want to run the documentation notebooks locally or rebuild the docs site. Some example audio files in `docs/guide/interactive/audio/` are tracked via Git LFS to keep the main repo lightweight. `pip install drumscript` and ordinary use of the package do **not** need it. If you skip this step, `git clone` will still succeed — you'll just get small LFS pointer files in place of the example audio.
+  - macOS: `brew install git-lfs`
+  - Ubuntu/Debian: `sudo apt-get install git-lfs`
+  - Windows: [Download from git-lfs.com](https://git-lfs.com/) or install via `winget install GitHub.GitLFS`.
+  - After installing, run `git lfs install` once, then `git lfs pull` inside the cloned repo to fetch the audio.
+
 ---
 
 ## Quick Start
 
+> Please note: DrumScript assumes you are providing **drum audio-only inputs by default** 
+
+> If you are using transcription with full song use the `full_song=True` flag, ie 
+
 ### End-to-end transcription
 
 ```python
@@ -170,10 +209,11 @@ import drumscript as ds
 pdf_path = ds.transcribe("drum_audio.wav")
 
 # Transcribe a full song (separates drums automatically)
-pdf_path = ds.transcribe("full_song.mp3", full_song=True)
+pdf_path = ds.transcribe("full_song.mp3") # drum only audio
+pdf_path = ds.transcribe("full_song.mp3", full_song=True) # full song, tells DrumScript to extract the drums first and then transcribe
 
 # Get all intermediate results
-result = ds.transcribe("drum_audio.wav", full=True)
+result = ds.transcribe("drum_audio.wav", verbose=True)
 print(f"Tempo: {result['tempo']:.1f} BPM")
 print(f"Events: {len(result['events'])}")
 ```
@@ -205,11 +245,34 @@ results = ds.extract_stems(
     "full_song.mp3",
     drumless=True,
     output_format="mp3",
-    full=True,
+    verbose=True,
 )
 print(f"Backing track: {results['mix']}")
 ```
 
+---
+
+## Benchmarking
+
+DrumScript includes a benchmarking framework for evaluating the classifier against standard ADT datasets using [`mir_eval`](https://github.com/mir-evaluation/mir_eval). Currently supports IDMT-SMT-Drums V2.
+
+```zsh
+# Install dev dependencies (includes mir_eval)
+uv sync --extra dev
+
+# Run the IDMT benchmark
+uv run --extra dev python benchmarks/run.py idmt \
+  --root /path/to/IDMT-SMT-DRUMS-V2
+
+# Run on a single subset with a limit
+uv run --extra dev python benchmarks/run.py idmt \
+  --root /path/to/IDMT-SMT-DRUMS-V2 \
+  --subset RealDrum --limit 5
+```
+
+Results are archived to `outputs/benchmarks/idmt/` with per-file metrics, summary statistics, and git commit tracking for reproducibility. See [`benchmarks/README.md`](benchmarks/README.md) for dataset setup and full usage.
+
+
 ---
 
 ## CLI Usage
@@ -218,29 +281,29 @@ DrumScript also provides a command-line interface.
 
 ### Basic transcription (isolated drum stem)
 
-```bash
+```zsh
 drumscript drum_audio.wav
 ```
 
 ### Full song transcription (auto-separates drums)
 
-```bash
-drumscript full_song.mp3 --full
+```zsh
+drumscript full_song.mp3 --full-song
 ```
 
 ### Extract a drumless backing track
 
-```bash
+```zsh
 drumscript full_song.mp3 --drumless
 ```
 
 ### All options
 
-```bash
+```zsh
 drumscript <audio_file> [OPTIONS]
 
 Options:
-  --full          Transcribe a full song (isolates drums first via Demucs)
+  --full-song     Transcribe a full song (isolates drums first via Demucs)
   --drumless      Extract a drumless backing track
   --mute STEM     Mute a specific stem (e.g. --mute bass). Repeatable.
   --all-stems     Export all individual stems (drums, bass, vocals, other)
@@ -251,7 +314,7 @@ Options:
 
 ### Examples
 
-```bash
+```zsh
 # Transcribe with 6/8 time signature
 drumscript drum_audio.wav --ts 6/8
 
@@ -273,6 +336,28 @@ We welcome contributions! DrumScript is intended to be a community-owned project
 
 > All bug reports and feature requests must be filed as GitHub Issues. All code changes must be submitted as Pull Requests. Keeping discussion public helps everyone.
 
+**[hello.drumscript@gmail.com](mailto:hello.drumscript@gmail.com)**
+
+## Alpha Priorities (v0.0.4 < v1.0.0) 
+The alpha phase runs between 01 June and 31 August 2026
+
+**What works today:**
+
+- End-to-end transcription pipeline: audio → onsets → classification → PDF / MIDI / MusicXML
+- Tempo detection via spectral onset envelope
+- Stem separation using Demucs (`htdemucs` 4-stem model)
+- Drumless backing track generation
+- CLI and Python API
+
+**What we're focused on during the alpha:**
+
+- Expanding test coverage across genres, kit types, and recording conditions
+- Fixing classification edge cases (deep snares vs clicky kicks, splash cymbals vs open hats)
+- Improving onset detection sensitivity for ghost notes and fast passages
+- Stabilising the public API ahead of the beta freeze
+- Community feedback collection
+
+
 ---
 
 ## FAQs
@@ -297,8 +382,9 @@ DrumScript's own classification engine is **fully deterministic** — it uses ph
 
 ## Acknowledgements
 
-- **[Demucs](https://github.com/adefossez/demucs)** — The stem splitting functionality is built upon the work of [@adefossez](https://github.com/adefossez).
-- **[librosa](https://librosa.org/)** — For foundational audio processing tools.
+1. **[Demucs](https://github.com/adefossez/demucs)** — The stem splitting functionality is built upon the work of [@adefossez](https://github.com/adefossez).
+2. **[librosa](https://librosa.org/)** — For foundational audio processing tools.
+3. **[@nanaoto](https://github.com/nanaoto)** — For building the `mir_eval` benchmarking infrastructure and IDMT-SMT-Drums V2 adapter (PR [#273](https://github.com/DrumScript/DrumScript/pull/273)).
 
 ---
 
@@ -306,6 +392,18 @@ DrumScript's own classification engine is **fully deterministic** — it uses ph
 
 [Apache License 2.0](LICENSE)
 
+---
+## Similar Projects
+
+DrumScript has no affiliation with any of the projects below. They are listed for context and reference.
+
+* **[librosa](https://librosa.org/)** — The spectral analysis library that powers DrumScript's onset detection and feature extraction.
+* **[Demucs](https://github.com/adefossez/demucs)** — The stem separation model we use for isolating drums from full mixes.
+* **[tepreece/drumscript (Golang)](https://github.com/tepreece/drumscript)** — A `(Go)lang` MIDI drum pattern scripting language by Tom Preece. Different use case (composing drum patterns via script), different technology (MIDI output rather than audio transcription). If you're looking to *write* drum patterns programmatically, check it out. Maintained by [@tepreece](https://github.com/tepreece)
+* **[basic-pitch](https://github.com/spotify/basic-pitch)** — A lightweight yet powerful audio-to-MIDI converter with pitch bend detection (better for non-percussive audio). Maintained by Spotify.
+* **[mir_eval](https://github.com/mir-evaluation/mir_eval)** — Standard evaluation metrics for music information retrieval tasks.
+* **[onset_db](https://github.com/CPJKU/onset_db)** - Provides a dataset of annotated musical onsets for tuning and evaluating audio detection algorithms. Maintained by JKU Linz.
+
 ---
 
 <!--END-->