PAXECT SelfTune Plugin v1.0.0 — Initial Public Release

PAXECT SelfTune Plugin v1.0.0 — Initial Public Release

Adaptive self-learning performance tuner for deterministic pipelines.
Monitors execution metrics, latency, and system health to dynamically adjust runtime parameters — fully offline and reproducible.

Key capabilities

• Adaptive tuning based on observed metrics
• Deterministic adjustment of blocksize, compression, and parallelism
• Auto-learn mode with JSON telemetry output
• Zero network dependency — 100% offline
• Integrates seamlessly with PAXECT Core
• Provides human-readable summary and JSONL logs

System requirements

• Python 3.9 – 3.12
• psutil package recommended for metrics
• Works standalone or with PAXECT Core

Quick start

Run the enterprise demo scripts from the repository root:

python demos/selftune_enterprise_demo_01_standalone.py       # Baseline self-tuning [OK]
python demos/selftune_enterprise_demo_02_safety_throttle.py  # Adaptive throttle & safety check [OK]
python demos/selftune_enterprise_demo_03_adaptive_benchmark.py # Performance benchmark cycle [OK]

License

MIT License (see LICENSE in repo)