flux-tensor-midi 0.1.0

PLATO rooms as musicians — flux tensor MIDI ensemble coordination

FLUX-Tensor-MIDI ⚒️

PLATO rooms as musicians. Each room has a T-0 clock, produces timestamped events (tiles=notes), listens to other rooms, snaps to rhythm via Eisenstein lattice, and sends side-channels (nods/smiles/frowns) for ensemble coordination.

Zero external dependencies. Pure Python 3.10+.

Musical Analogy Table

PLATO Concept	Musical Equivalent	MIDI Mapping
Room	Musician	MIDI Channel
T-0 Clock	Metronome / Conductor	MIDI Clock (0xF8)
Tile (event)	Note	Note On/Off
FluxVector (9-ch)	Chord / Harmonic Spectrum	9 MIDI notes (C4–D5)
Salience	Note Velocity	Velocity (0–127)
Tolerance	Pitch Bend / Jitter	Pitch Bend range
Eisenstein Snap	Rhythmic Quantization	Grid snapping
Nod	Nod (agreement)	CC #1 (Mod Wheel)
Smile	Smile (approval)	CC #2 (Breath)
Frown	Frown (disagreement)	CC #3 (Expression)
Listening (room→room)	Musician's ear	MIDI Channel routing
Conductor	Band leader / click	Master MIDI Clock
Ensemble	Band / Orchestra	Multi-channel MIDI output
Jaccard Harmony	Chord similarity	Chromatic interval match
Rhythmic Role	Groove part	Program change (patch)

Rhythm Ratios (Eisenstein Lattice)

Role	Ratio	Period (at 120 BPM)	Musical Feel
Root	1:1	500 ms	Downbeat / quarter
Halftime	2:1	1000 ms	Half speed
Triplet	3:2	750 ms	Swung feel
Waltz	3:1	1500 ms	Waltz time
Compound	4:3	~667 ms	Compound meter
Doubletime	1:2	250 ms	Double speed
Offset	1:1	500 ms + 120° phase	Syncopated
Quintuple	5:4	625 ms	Quintuple meter
Septuple	7:4	875 ms	Septuple meter

Covering radius: 1/√3 ≈ 0.577 — optimal hexagonal tiling.

Quick Start

pip install flux-tensor-midi

from flux_tensor_midi import FluxVector, TZeroClock, RoomMusician, EisensteinSnap
from flux_tensor_midi.core.snap import RhythmicRole
from flux_tensor_midi.ensemble.band import Band

# Create musicians with different rhythmic roles
piano = RoomMusician("Piano", role=RhythmicRole.ROOT)
bass = RoomMusician("Bass", role=RhythmicRole.HALFTIME)
drums = RoomMusician("Drums", role=RhythmicRole.TRIPLET)

# Form a band with a conductor
band = Band("Trio", conductor=piano, bpm=120.0)
band.add_musician(bass)
band.add_musician(drums)
band.everyone_listens_to_everyone()

# Set some initial state
piano.update_state(FluxVector([0.8, 0.0, 0.6, 0.0, 0.7, 0.0, 0.0, 0.0, 0.0]))
bass.update_state(FluxVector([0.0, 0.9, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]))  # simple root

# Tick the band
results = band.tick_all()
print(f"Piano event at {results['Piano'][0]:.1f}ms")

# Check harmony
hs = band.harmony()
print(f"Chord quality: {hs.quality()}, consonance: {hs.consonance():.3f}")

# Side-channel comms
piano.send_nod(bass)
print(f"Bass received nods from: {bass.receive_sidechannels()['nods']}")

Package Structure

flux_tensor_midi/
├── __init__.py
├── core/
│   ├── __init__.py
│   ├── flux.py       — FluxVector (9-ch tensor with salience+tolerance)
│   ├── clock.py      — TZeroClock (EWMA-adaptive metronome)
│   ├── room.py       — RoomMusician (PLATO room as musician)
│   └── snap.py       — EisensteinSnap (rhythmic quantization lattice)
├── midi/
│   ├── __init__.py
│   ├── events.py     — MidiEvent, NoteName, flux→MIDI conversion
│   ├── clock.py      — MidiClock (24 PPQN, software clock)
│   └── channel.py    — MidiChannel mapping by rhythmic role
├── sidechannel/
│   ├── __init__.py
│   ├── nod.py        — Nod gesture (agreement/acknowledgment)
│   ├── smile.py      — Smile gesture (positive affect)
│   └── frown.py      — Frown gesture (disagreement/concern)
├── harmony/
│   ├── __init__.py
│   ├── jaccard.py    — Jaccard similarity for FluxVector sets
│   ├── spectrum.py   — Spectral analysis (centroid, flux, autocorr)
│   └── chord.py      — HarmonyState (consonance, quality, voice leading)
└── ensemble/
    ├── __init__.py
    ├── band.py       — Band (ensemble with conductor)
    └── score.py      — Score (recorded performance, export)

Running Tests

cd python
pip install -e ".[dev]"
pytest -v

License

MIT