Add real-time pipeline, biofeedback fusion, embodied interaction, and evaluation modules

[Copilot]

The repository needed to demonstrate capability for WP1 MishMash doctoral research in real-time human-AI co-creation for musical performance. The existing system had strong foundations (hybrid agent, DDSP synths, EEG mapper) but lacked the streaming architecture, multimodal biofeedback integration, embodied interaction, and research evaluation infrastructure that WP1 specifically requires.

New modules

• performance_system/realtime/ — Event-driven streaming pipeline: EventBus (thread-safe pub/sub with latency tracking), StreamingPipeline (configurable stage chain with per-stage metrics), PerformanceState (session state with human-AI synchrony computation)

• performance_system/biofeedback/ — Multimodal physiological signal fusion: MultimodalFusion (weighted fusion with staleness detection and confidence decay), HRVProcessor (SDNN/RMSSD from IBI), GSRProcessor (tonic/phasic decomposition), PhysiologicalMapper (direct, affective circumplex, adaptive strategies)

• performance_system/embodied/ — Gesture and spatial interaction: GestureController (classifies sustained/percussive/sweeping/tremolo from 3D position), SpatialAudioMapper (position→pan/reverb/brightness), InteractionModeManager (responsive/collaborative/autonomous/mirroring with smooth blending)

• performance_system/evaluation/ — Research metrics: InteractionMetrics (synchrony via Pearson correlation, latency percentiles, adaptation rate, coherence), SessionLogger (timestamped frames with JSON export)

Usage

from performance_system.realtime import EventBus, StreamingPipeline, PipelineStage
from performance_system.biofeedback import MultimodalFusion, ModalityConfig, PhysiologicalMapper
from performance_system.embodied import InteractionModeManager, InteractionMode
from performance_system.evaluation import InteractionMetrics, SessionLogger

fusion = MultimodalFusion()
fusion.add_modality(ModalityConfig(name='eeg', weight=1.0, n_features=2))
fusion.add_modality(ModalityConfig(name='hrv', weight=0.8, n_features=1))

fusion.update('eeg', {'arousal': 0.7, 'focus': 0.5})
fusion.update('hrv', {'relaxation': 0.6})
state = fusion.get_fused_state()  # weighted, staleness-aware fusion

Tests

59 new tests across 4 files (70 total). All passing. Version bumped to 0.2.0.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.