-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Nick edited this page Mar 10, 2026
·
21 revisions
Pattern-Adaptive Transmodal Anti-Spam System
PATAS is an autonomous pattern discovery and rule management system for anti-spam operations. It analyzes historical message logs, automatically discovers spam patterns, generates safe blocking rules, and evaluates their effectiveness before deployment.
Key Characteristics:
- Signal engine, not enforcement - PATAS provides patterns and metrics that inform anti-spam decisions
- On-premise deployment - Designed for deployment within your infrastructure
- Two-stage processing - Fast scanning + deep analysis for 70-90% cost reduction
- Deterministic and rule-based - Core engine is deterministic; ML/LLM is optional
- Safety-first design - Multiple safety profiles with clear risk boundaries
Core Workflow:
- Ingest - Load historical message logs into PATAS
- Discover - Automatically identify recurring spam patterns (two-stage: fast scan + deep analysis)
- Generate - Create safe SQL rules from discovered patterns
- Evaluate - Test rules on historical data (shadow mode)
- Promote - Activate rules that meet safety thresholds
- Monitor - Track rule performance and deprecate underperforming rules
New to PATAS? Start here:
- Quick Start Guide - Installation, configuration, and basic usage (10-15 minutes)
- Architecture - Understand system design and components
- API Quickstart - Quick start guide for API integration
- FAQ - Frequently asked questions
- Product PRD - Complete Product Requirements Document
- Safety Profiles - Conservative, Balanced, and Aggressive profiles
- Custom Profiles - Define custom aggressiveness profiles with specific thresholds
- LLM Usage - LLM integration and privacy guarantees
- Threshold Calibration Guide - How to calibrate thresholds for your use case
- Incremental Mining - Process only new messages for faster, cheaper operations
- Performance and Cost - Real-world metrics, throughput, and cost estimates
- On-Premise Deployment - Local LLM/embedding models, air-gapped deployment
- Scaling & Cost Design - How PATAS handles millions of logs efficiently
- Scaling Guide - Horizontal scaling strategies
- API Reference - Complete API endpoint documentation (includes filtering, explanations, risk assessment)
- API Quickstart - Quick start guide for API integration
- Engineering Notes - Technical overview for engineering teams
- Code Overview - Code structure and navigation guide
-
Legacy Public Wiki Import - Imported reference pages preserved from the archived
PATAS-publicwiki
- Configuration - Configuration reference and environment variables
- Configuration Examples - Example configurations for different deployment scenarios
- Production Deployment Guide - Production deployment and maintenance
- Migration Guide - Database migration instructions
- Performance and Cost - Real-world metrics, throughput, and cost estimates
- Horizontal Scaling - Scaling to 10M+ messages with multiple instances
- Scaling & Cost Design - How PATAS handles millions of logs efficiently
- Scaling Guide - Horizontal scaling strategies
- Distributed Locks - Multi-instance coordination
- Checkpointing - Resume pattern mining from checkpoints
- Performance Guide - Performance benchmarks and optimization
- Load Testing Guide - Load testing tools and procedures
- Monitoring Setup - Grafana and Prometheus setup guide
- Alerting - Alert rules and AlertManager configuration
- Safety Guide - Safety profiles and enforcement model
- Security - Security considerations and best practices
- Security Testing - Security testing procedures and tools
- Security Audit Checklist - Comprehensive security checklist
- Secret Management - Best practices for managing secrets
- Privacy and Data Protection - Privacy modes and data handling
- Roadmap - Planned enhancements after successful pilot
- GitHub: KikuAI-Lab/PATAS