feat: Add OneChart score reconstruction research files and validation results

- Introduced new JSON files for normalized and raw inputs, source metadata, and public inputs.
- Added findings document detailing the methodology and results of the score reconstruction.
- Included member differences and worked examples for clarity on data discrepancies.
- Implemented a Python script for reproducing scores based on public inputs.
- Created SQL for raw reaggregation of data.
- Added task metadata for tracking the research completion and validation metrics.

.trellis/tasks/archive/2026-09/09-21-onechart-score-reconstruction/design.md

@@ -0,0 +1,9 @@
+# 研究设计
+
+这是只读分析任务，不进入产品实现阶段。
+
+公开证据链：页面展示 → 实际引用脚本 → 实际请求的公开数据 → 评分与排名字段。数据库证据链：库表目录 → 字段及单位 → 重叠交易日原始值 → 候选公式复算 → 与公开评分比较。
+
+优先检验可解释的低自由度公式。分开检验资金比率、横截面排序/归一化、时间窗口聚合；用多日和不同板块类型验证，避免单点拟合。识别数据修订、单位换算、成分股聚合与板块原始数据的口径差异。
+
+数据库连接强制 default_transaction_read_only，设置查询超时。仅在本机保留任务所需数据；公开资料可以缓存供复核，凭据不落盘。报告不将无法唯一识别的参数写成确定结论。