fix: close adversarial review findings on document parsing

Malformed download URLs are rejected before the second request, error bodies are read only up to a bound, the telemetry summary keeps structural dispatch for duck-typed results, and the BMP magic now requires a valid DIB header. Docs and design are aligned with the implementation, and new test wording is neutral.
2026-10-22 02:09:06 -04:00
parent bb621cad4d
commit cb1e08f1f3
10 changed files with 180 additions and 25 deletions
@@ -1,8 +1,8 @@
 """OCR 文档结果包的合成 fixtures(#34)。

 全部**合成**: 保留 2026-10-21 实测的结构骨架(para_blocks 块类型、table/image 的嵌套
-`blocks`、span 字段),一切识别文本替换为 `SYNTH-*` 占位——第三方论文内容许可不明、
-医疗数据不入库(零业务假设 + P5),真实 ZIP 不提交。
+`blocks`、span 字段),一切识别文本替换为 `SYNTH-*` 占位——第三方文档内容许可不明、
+敏感数据不入库(零业务假设 + P5),真实 ZIP 不提交。
 """

 import io