2.2.0

BoxLang AI Module v2.2.0 Release Notes - Enhanced Performance Monitoring, Configuration Consolidation, and Bug Fixes

This release focuses on improving performance monitoring, consolidating configuration management, and addressing various bug fixes for better stability and reliability.

🎁 New Features

• Consolidated AI request/response logging - Execution time metrics for better performance insights

• Enhanced AI request/response metrics - Additional metrics to provide better insights into performance and potential bottlenecks

🔧 Enhancements

• Configuration Consolidation - Single source of truth for configuration options and settings, allowing for better overrides and defaults

• Stream Request Logging - Execution time metrics included for better performance monitoring and debugging insights

• Empty Chunk Handling - Skip empty chunks (keep-alive or heartbeat) when doing chat streams to prevent unnecessary processing and potential parsing errors

🐛 Bug Fixes

• aiChatStream() BIF - Fixed invalid use of request parameter, which should have been chatRequest

• AiTransformRunnable - Corrected extends declaration

• AiModel extractMessages() - Now correctly flattens messages when response has multiple choices with multiple messages into a single array

• Options Merging Order - Fixed incorrect order in aiChat() and aiChatStream() BIFs; now merges correctly: user options → module settings → default options

• Schema Builder Population - Fixed error invoking population; third argument now properly accepts array or struct, not a single value

• Provider Options Merging - Fixed bug where provider options in configuration file were not being merged into request options when creating a service instance

• aiService() API Key Detection - Fixed convention-based API key detection when options.apiKey was empty; now properly checks before applying convention key for fallback to environment variables or module settings

🚀 Upgrade Notes

This release is backward compatible. All improvements are automatic, with bug fixes addressing edge cases and configuration issues discovered in production environments.

🙏 Thank You

Thank you to all contributors and users who continue to help improve BoxLang AI!