rl 1.0.0

First stable release

The RL (Reinforcement Learning) API module reaches its first stable release. This module provides Thompson Sampling-based A/B testing for Drupal, automatically optimizing content by shifting traffic toward better-performing variants.

Changes since 1.0.0-rc1

• Full Drush CLI - New Drush commands for experiment management, configuration, analytics, and setup, achieving full parity with the admin GUI/TUI
• Claude Code skill - Built-in /rl slash command for Claude Code AI assistant, enabling natural language experiment management via Drush
• CI improvements - Updated GitHub Actions to latest major versions and migrated to Ubicloud runners

Highlights since 1.0.0-alpha1

• Thompson Sampling algorithm - Pure PHP implementation of Bayesian multi-armed bandit with sliding time window support
• Fast HTTP REST API - Optimized rl.php direct endpoint and Drupal route alternatives for recording turns and rewards
• Analytics API service - API service and Drush commands for AI-powered experiment analysis
• Interactive charts - Plotly.js-powered 2D and 3D charts with Y-axis metric toggle and max variants filter
• Administrative reports - Experiment analysis interface with date filters and snapshot storage
• Drush CLI commands - Complete command-line interface for experiment CRUD, configuration, analytics, and setup
• Claude Code AI skill - /rl slash command lets AI assistants manage experiments through natural language via Drush CLI
• Debug mode - Built-in debugging for experiment troubleshooting
• Page cache management - Service to override page cache duration for experiments
• Event logging - Optional event log for experiment activity tracking
• UX enhancements - Animated gradient border branding for RL experiments