feat: layered precise generation for A4 report images

3-phase pipeline to solve LLM prompt overflow from too many OCR elements:
Phase 1 (generate_skeleton): compressed layout schema → skeleton JRXML
Phase 2 (refine_layout): sampled coordinates → pixel-level position tuning
Phase 3 (map_fields): OCR field names → replace $F{field_N} placeholders

Only triggered when layout_schema.total_rows > 0 on initial_generation intent.
Text requests and all other intents are unaffected (zero behavior change).

agent/state.py

@@ -47,3 +47,7 @@ class AgentState(TypedDict, total=False):
 
     # 需求8：图片批注检测（圈选/箭头标记）
     annotation_result: dict
+
+    # 需求9：分层精确生成
+    layout_schema: dict       # extract_layout_schema() 输出，列+区域结构
+    ocr_elements: list        # OCR 原始行数据（用于阶段二坐标采样）