v0.1.0

What's Changed

• feat: 实现核心数据模型、性能估算器和 PD 比率规划器 by @hlin99 in #2
• M2: 排队论性能模型 + GPU Profile 库 + 批处理效应 by @hlin99 in #4
• M3: 核心重构 — 实测数据分析器 by @hlin99 in #6
• M4: 多场景分析 — 支持多 QPS 级别的 benchmark 数据分析 by @hlin99 in #8
• M5: 敏感度分析 — P:D 比例 vs SLA 满足率曲线 by @hlin99 in #10
• M5: 敏感度分析 — P:D 比例 vs SLA 满足率曲线与悬崖检测 by @hlin99 in #11
• feat: M6 xpyd-bench integration — adapter, streaming analyzer, schema versioning by @hlin99 in #13
• feat(report): M7 — HTML report generation with inline SVG visualizations by @hlin99 in #15
• feat(cost): M8 — cost-aware optimization with budget constraints by @hlin99 in #17
• docs: update ROADMAP with M8 completion and M9 milestone by @hlin99 in #18
• feat(export): M9 — JSON/CSV export and programmatic API by @hlin99 in #20
• docs: mark M9 as completed in ROADMAP by @hlin99 in #21
• feat(capacity): M10 — capacity planning mode with scaling model by @hlin99 in #23
• docs: mark M10 as completed, add M11 milestone by @hlin99 in #24
• M10: Remove deprecated legacy estimator and planner modules by @hlin99 in #26
• feat: M11 What-If Scenario Simulation by @hlin99 in #28
• docs: update ROADMAP.md with M11 completion by @hlin99 in #29
• feat(config): Configuration Profile Support (M12) by @hlin99 in #31
• feat: configurable SLA percentile threshold (P90/P95/P99) by @hlin99 in #33
• feat: M14 — Benchmark Comparison & Regression Detection by @hlin99 in #35
• docs: mark M14 complete, add M15-M17 milestones by @hlin99 in #37
• feat(trend): M15 — Historical Trend Tracking by @hlin99 in #39
• feat(validator): M16 — Benchmark Data Validation & Outlier Detection by @hlin99 in #41
• feat(dashboard): M17 — Interactive CLI Dashboard by @hlin99 in #43
• feat: M18 Performance Interpolation Model by @hlin99 in #45
• feat: M19 Alert Rules Engine — threshold-based alerting for CI/CD by @hlin99 in #47
• feat: M20 Benchmark Annotation & Tagging system by @hlin99 in #49
• docs: mark M19 and M20 as completed in ROADMAP.md by @hlin99 in #50
• feat(pareto): M21 — Pareto Frontier Analysis by @hlin99 in #52
• feat(recommend): M22 — Recommendation Engine by @hlin99 in #54
• feat(fleet): M23 — Fleet Sizing Calculator by @hlin99 in #56
• docs: mark M23 complete, add M24-M26 milestones by @hlin99 in #57
• feat(pipeline): M24 — Batch Pipeline Runner by @hlin99 in #59
• feat(report): M25 — Markdown Report Generation by @hlin99 in #61
• docs: mark M24 complete in ROADMAP.md by @hlin99 in #62
• feat(generator): M26 — Benchmark Simulation Generator by @hlin99 in #64
• feat: M27 — SLA Budget Allocation by @hlin99 in #66
• feat: M28 — Benchmark Merge & Aggregation by @hlin99 in #68
• feat: M29 — Benchmark Filtering & Slicing by @hlin99 in #70
• refactor: split monolithic cli.py into cli/ package by @hlin99 in #72
• feat: M31 — Bootstrap Confidence Intervals for Latency Percentiles by @hlin99 in #74
• docs: mark M31 complete, add M32 roadmap by @hlin99 in #76
• feat: M32 — Multi-Model Comparison Matrix by @hlin99 in #78
• docs: mark M32 as completed in ROADMAP.md by @hlin99 in #79
• feat: M33 — A/B Test Analysis by @hlin99 in #81
• feat(workload): Workload Characterization & Clustering (M34) by @hlin99 in #83
• feat(scaling): throughput scaling analysis (M35) by @hlin99 in #85
• feat(metrics): Prometheus/OpenMetrics export (M36) by @hlin99 in #87
• feat(correlation): request-latency correlation analysis (M37) by @hlin99 in #89
• feat(discovery): batch benchmark discovery & auto-loading (M38) by @hlin99 in #91
• feat(heatmap): latency heatmap data generation (M39) by @hlin99 in #93
• docs: mark M39 as completed in ROADMAP.md by @hlin99 in #94
• feat(timeline): request timeline analysis (M40) by @hlin99 in #96
• feat(drift): distribution drift detection using KS test (M41) by @hlin99 in #98
• feat(root-cause): anomaly root cause analysis (M42) by @hlin99 in #100
• docs: mark M42 as completed in ROADMAP.md by @hlin99 in #101
• feat: M43 — Tail Latency Analysis by @hlin99 in #103
• feat: add efficiency scorecard (M44) by @hlin99 in #105
• feat: M45 Benchmark Plan Generator by @hlin99 in #107
• feat: M46 SLA Threshold Tuning Advisor by @hlin99 in #109
• feat: M47 Capacity Forecasting by @hlin99 in #111
• feat: M48 Multi-SLA Tier Analysis by @hlin99 in #113
• feat: M49 Saturation Point Detection (#114) by @hlin99 in #115
• docs: update M49 as completed, add M50 by @hlin99 in #116
• feat: M50 Latency Decomposition Analysis by @hlin99 in #118
• docs: comprehensive README rewrite and Phase 2 roadmap (M51) by @hlin99 in #120
• feat: plugin architecture with entry-point discovery (M52) by @hlin99 in #122
• feat: M53 — End-to-End Integration Tests by @hlin99 in #124
• fix: M54 — Fix Pre-Existing CLI Bugs (scorecard, pareto, threshold-advisor) by @hlin99 in #126
• feat: M55 Request Fairness Analysis by @hlin99 in #128
• feat(summary): add benchmark quick summary subcommand (M56) by @hlin99 in #130
• feat: M57 Outlier Impact Analysis by @hlin99 in #132
• feat: M58 — Percentile Convergence Analysis by @hlin99 in #134
• feat: M59 — Load Profile Classification by @hlin99 in #136
• feat: M60 — Throughput Percentile Analysis by @hlin99 in #138
• feat: M61 — Token Efficiency Analysis by @hlin99 in #140
• docs: mark M61 complete, add Phase 2 milestones (M62-M66) by @hlin99 in #142
• feat: M62 Request Queuing Time Analysis by @hlin99 in #144
• feat: M63 Batch Size Impact Analysis by @hlin99 in #146
• feat: M64 Multi-Benchmark Statistical Summary by @hlin99 in #148
• feat: M65 Cost Projection & ROI Calculator by @hlin99 in #150
• feat: Benchmark Schema Versioning & Migration (M66) by @hlin99 in #152
• feat: Request Replay Schedule Generator (M67) by @hlin99 in #154
• feat: M68 — Latency-Token Regression Analysis by @hlin99 in #156
• feat: M69 — Latency CDF Export by @hlin99 in #158
• feat: M70 — Benchmark Health Check by @hlin99 in #160
• feat: M71 — Warm-up Exclusion Filter by @hlin99 in #162
• feat: M72 — Request Arrival Pattern Analysis by @hlin99 in #164
• feat: M73 Benchmark Sampling & Downsampling by @hlin99 in #166
• feat: M74 — Request Size Distribution Analysis by @hlin99 in #168
• feat: M75 — Concurrency Utilization Analysis by @hlin99 in #170
• feat: M76 — Benchmark Reproducibility Score ...