Tags: vim89/toon4s
Tags
feat: apache spark integration feat: apache spark integration # feat: apache spark integration ### Added - Apache Spark integration module (toon4s-spark) for DataFrame/Dataset TOON encoding - Schema alignment analyzer with benchmark-based scoring for production safety - Adaptive chunking to optimize prompt tax based on dataset size - Delta Lake CDC integration for real-time streaming - Iceberg time travel support for historical snapshot analysis - Production monitoring with health checks and telemetry - LLM client abstraction compatible with llm4s patterns ### Changed - CI workflows updated for multi-module publishing - Unified versioning across toon4s-core and toon4s-spark ### Removed - Old toon4s-compare module ### Bug fixes - escape dollar signs in scaladoc code examples
Merge pull request #43 from vim89/perf/apply-optimization-opportunities perf: performance optimization - inspired from PR #42 by @rorygraves . Thank you Rory 🙏 Primitives.scala: Hoisted structuralChars Set to object level + eliminated .trim allocations using Character.isWhitespace() (5-10% improvement) Encoders.scala: Pre-allocated StringBuilder capacity in 5 hot path locations (10-15% improvement) StringifyVisitor.scala: Added Primitives.quoteAndEscape() method and eliminated duplicate quoting logic (5% improvement) DelimitedValuesParser.scala: Eliminated .trim allocations by trimming during building with while loops (15-20% improvement) SimpleJson.stringify: Rewrote array and object serialization using while loops instead of .map().mkString() (20-30% improvement) SimpleJson.quote: Converted from foldLeft to while loop for single-pass quoting (10% improvement) Encoders.extractTabularHeader: Added early exit using iterator.forall for short-circuit evaluation (3-5% improvement)
PreviousNext