beatlyze-analyzer 0.1.0

Audio feature extraction: BPM, key, energy, danceability, mood, genre.

beatlyze-analyzer

Audio feature extraction library: BPM, musical key, energy, danceability, valence, LUFS loudness, mood tags, genre suggestions, and structural sections.

Installation

Light tier (librosa-only, cross-platform):

pip install "beatlyze-analyzer[light]"

Full tier (Essentia for higher-quality extraction; Linux x86_64 / macOS arm64):

pip install "beatlyze-analyzer[full]"

Usage

from beatlyze_analyzer import analyze

result = analyze("/path/to/track.mp3")
print(result["bpm"])          # 128.4
print(result["key_notation"]) # "Am"
print(result["mood_tags"])    # ["energetic", "dark"]

Output fields

Field	Type	Description
bpm	float	Tempo in beats per minute
bpm_confidence	float 0–1	Extraction confidence
key	str	Root note (e.g. "A")
scale	str	"major" or "minor"
key_notation	str	Shorthand (e.g. "Am")
key_confidence	float 0–1
energy	float 0–1	Perceived loudness / intensity
danceability	float 0–1	Beat regularity + onset density
valence	float 0–1	Musical positivity
loudness_lufs	float	Integrated loudness (LUFS R128)
mood_tags	list[str]	e.g. ["energetic", "uplifting"]
genre_suggestions	list[str]	Top 3 genre labels
duration_seconds	float	Track length
time_signature	int	Beats per bar (2, 3, 4, or 6)
sections	list[dict]	Structural segments with per-section features

License

AGPLv3 — required by the Essentia dependency (AGPLv3). Source code must remain publicly available when used in a network service.