| name | hlzd-pipeline-viz |
| description | B2B Skill 链路可视化 —— 消费各 Skill trace JSON，输出 8 KPI 仪表盘 + 漏斗图 + 评级分布 + 决策分布 + halt 警报 self-contained HTML（纯 stdlib + inline SVG，零依赖）。Use when 用户说'链路可视化'、'Pipeline 仪表盘'、'B2B 转化漏斗'、'Funnel Dashboard'、'pipeline metrics'、'trace aggregate'、'B2B analytics'。 |
| license | MIT |
| metadata | {"author":"HLZD (海联智达 / 海良数科)","version":"0.1.0","industry":"cross-border-b2b","category":"analytics","triggered_by":["链路可视化","仪表盘","漏斗","funnel dashboard","pipeline metrics","trace aggregate","B2B analytics"]} |
HLZD Pipeline Viz
把多个 Skill 的 trace JSON 累计成一份 self-contained HTML dashboard。
8 KPI 卡片
| KPI | 来源 |
|---|
| Traces | 输入 trace 数 |
| Halts | compliance/fraud 拦截数 |
| Follow-up actions due | followup-sequencer 输出 |
| Buyers evaluated | diligence 输出 |
| Quote total (USD) | quotation-gen 累计 |
| Avg margin | quotation-gen ratio |
| Knowledge entities | knowledge-graph 累计 |
| Graph nodes/edges | knowledge-graph 累计 |
4 图表
- Pipeline Funnel (按 stage 数量降序 SVG trapezoid)
- Grade Distribution (横向条形图 A/B/C/D)
- Clearance 分布 (CLEARED/PENDING_REVIEW/BLOCKED)
- Negotiation Decisions (accept_round_2 / counter / ...)
用法
import json
import lib
traces = [json.loads(p) for p in ["trace1.json", "trace2.json"]]
result = lib.run_pipeline(traces, title="HLZD Pipeline")
html = result["html_dashboard"]
open("dashboard.html", "w").write(html)
Output
result["html_dashboard"] —— 直接浏览器打开result["aggregation"] —— 累计指标 JSON (for further processing)
Related Skills
- 所有 18 个 Skill (上游) —— 提供 trace 输入
- 自身是 v1.3 的下游 — 完成后可以做执行控制台
Versioning
| 版本 | 说明 |
|---|
| 0.1.0 | 首版：8 KPI + 4 图表 + halt 警报 + SVG 自包含 |
