Files
agent_jrxml/prompts/skeleton_generation.md
T
panda bd5bfbac2d fix: band-level windowed refine_layout + programmatic map_fields to prevent 91.5% content loss
Root cause: LLM receiving full 34k-char JRXML would regenerate from scratch
instead of modifying coordinates in-place, shrinking output to ~3k chars.

Solution (programmatic node control, not prompt engineering):

- New agent/jrxml_windower.py: decompose JRXML into header (never sent to
  LLM) + individual bands. Split bands >4000 chars at element boundaries.
  Reassemble with element count validation (>10% change = rollback).

- Rewrite refine_layout: per-band windowed LLM processing (~2-4k chars
  each). LLM cannot "reimagine" the entire report.

- Rewrite map_fields: 100% programmatic regex $F{field_N} -> real name
  replacement. Zero LLM calls, zero content loss.

- _sanitize_field_name: non-ASCII chars escaped to _uXXXX_ format for
  valid JRXML identifiers.

- Tests: 48 new unit tests (windower 28 + map_fields 20). All passing.
  Full suite 385 tests, zero regressions.
2026-05-24 08:55:38 +08:00

1.1 KiB
Raw Blame History

你是一位资深 JasperReports 工程师。根据以下报表布局描述和用户需求,生成一个完整的骨架 JRXML 文件。

关键规则:

  • 只输出 JRXML 代码,不要解释,不要 markdown 标记。
  • 使用 $F{{field_1}}, $F{{field_2}}, ... 作为占位字段名,并在 部分声明它们。
  • 报表结构必须正确(title, pageHeader, columnHeader, detail, pageFooter 等 band)。
  • 元素位置使用近似值即可,后续会精确调整。
  • 根元素为 ,命名空间和 schemaLocation 必须精确,不可使用其他 URL(如 jaspersoft.com)。
  • 包含 ,在 中放置占位 SQLSELECT * FROM table_name)。
  • 确保 JRXML 兼容 JasperReports 7.0.6。

报表布局描述: {layout_schema}

参考模板和组件: {context}

{template_context}

用户需求: {user_request}