|
|
--- |
|
|
license: mit |
|
|
datasets: |
|
|
- Raiff1982/Nexus-signal-engine |
|
|
language: |
|
|
- en |
|
|
base_model: |
|
|
- Raiff1982/CodetteV5 |
|
|
--- |
|
|
|
|
|
# Nexis Signal Engine |
|
|
|
|
|
**Author**: Jonathan Harrison (a.k.a. Raiff1982) |
|
|
**License**: MIT |
|
|
**Version**: 1.0.0 |
|
|
**Repository**: [GitHub](https://github.com/Raiff1982/Nexus-signal-engine) |
|
|
**Hugging Face Space**: [nexus-signal-engine](https://huggingface.co/Raiff1982/nexus-signal-engine) |
|
|
|
|
|
--- |
|
|
|
|
|
## 🔍 Overview |
|
|
|
|
|
The **Nexis Signal Engine** is a production-grade AI analysis framework designed for: |
|
|
|
|
|
- **Ethical signal profiling** |
|
|
- **Entropy detection** |
|
|
- **Adversarial input resilience** |
|
|
- **Multi-perspective reasoning** |
|
|
- **Secure memory with audit trails** |
|
|
|
|
|
It integrates natural language lemmatization, fuzzy matching, FFT-based harmony profiling, and multiple ethical reasoning lenses (utilitarian, deontological, virtue, systems). |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧠 Key Features |
|
|
|
|
|
- 🧭 **Ethical Risk Scoring**: Uses configurable fuzzy-matched terms to detect intent, corruption risk, and ethical misalignment. |
|
|
- 🔁 **Perspective Agents**: Includes *Colleen* (vector), *Luke* (entropy), and *Kellyanne* (harmonics) to simulate multi-agent cognition. |
|
|
- 🔐 **Integrity-Sealed Memory**: SQLite + FTS5 + per-record salted SHA-256 integrity checks + rotation + pruning. |
|
|
- 🔍 **Resonance Equation**: Applies FFT over alphabetic patterns to derive harmonic volatility. |
|
|
- 🛡️ **Adaptive Intervention**: Flags and reframes high-risk signals in real-time. |
|
|
- 📈 **Built-In Metrics**: Tracks average processing time and internal error rates. |
|
|
- 🧪 **Comprehensive Unit Tests**: >30 robust tests for edge cases, adversarial tokens, and DB state. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🛠 Installation |
|
|
|
|
|
To install locally: |
|
|
|
|
|
```bash |
|
|
pip install . |
