feat(bench): LLM benchmarks, regression baselines, and bundle size tracking#8

Merged

SimplyLiz merged 14 commits intodevelopfrom

feature/llm-benchmark-integration

Mar 6, 2026

Commits on Feb 25, 2026

feat: auto-load .env, auto-detect Ollama, persist LLM benchmark results
SimplyLiz
committed
feat(bench): gate LLM benchmarks behind --llm flag, add scenarios and metrics
SimplyLiz
committed
feat(bench): wire --save/--check/--tolerance, organize baseline structure
SimplyLiz
committed
docs(bench): split benchmark docs into handbook + auto-generated results
SimplyLiz
committed
feat(bench): enrich benchmark report with badges, progress bars, and pie chart
SimplyLiz
committed
fix(bench): remove ugly unicode bars, fix stacked LLM comparison chart
SimplyLiz
committed
fix(bench): use paired bars for LLM comparison chart
SimplyLiz
committed
fix(bench): replace broken LLM comparison chart with summary table
SimplyLiz
committed
feat(bench): ASCII horizontal bar chart for LLM vs deterministic comparison
SimplyLiz
committed
refactor(compress): unify sync/async paths via generator (#5)
SimplyLiz
committed
docs: clarify defaultTokenCounter rationale across docs and source
SimplyLiz
committed
feat(bench): add provider summary, fuzzy dedup delta, per-provider ASCII charts
SimplyLiz
committed

Commits on Feb 26, 2026

feat(bench): track bundle size per-file with gzip in benchmark suite
SimplyLiz
committed

Commits on Mar 6, 2026

fix(bench): skip gzipBytes in regression check, run prettier
SimplyLiz
committed