feat: 新增 OCR 单据字段精确提取模块

- 新增 backend/ocr_extractor.py: 两阶段提取流水线 (文档分析 + 字段提取)
- 四种提取策略: 精确KV匹配/模糊KV匹配/正则模式/表格结构匹配
- agent/state.py: 新增 ocr_extraction_result 和 uploaded_file_path 字段
- agent/nodes.py: process_input() 中自动触发 OCR 提取钩子
- app.py: 文件上传时保留图片路径, 总结卡片中展示提取结果
- .env.example: 新增 OCR_USE_GPU / OCR_CONFIDENCE_THRESHOLD 配置项
- tests/test_ocr_extraction.py: 48 个单元测试全部通过

agent/nodes.py

@@ -7,6 +7,7 @@ import os
 import re
 import time
 from datetime import datetime, timezone
+from pathlib import Path
 from typing import Dict
 
 from dotenv import load_dotenv
@@ -114,6 +115,30 @@ def process_input(state: AgentState) -> Dict:
     conv_history.append({"role": "user", "content": user_input})
     state["conversation_history"] = conv_history
 
+    # OCR 单据字段精确提取（处理上传的图片文件）
+    uploaded_path = state.get("uploaded_file_path", "")
+    if uploaded_path and Path(uploaded_path).is_file():
+        suffix = Path(uploaded_path).suffix.lower()
+        if suffix in (".png", ".jpg", ".jpeg", ".bmp", ".webp"):
+            try:
+                from backend.ocr_extractor import OcrExtractor
+                extractor = OcrExtractor()
+                ocr_result = extractor.extract(uploaded_path, [])
+                if ocr_result.get("ocr_available"):
+                    state["ocr_extraction_result"] = ocr_result
+                    _node_log.info(
+                        "OCR 字段提取完成",
+                        extra={
+                            "file": uploaded_path,
+                            "elements": ocr_result.get("total_elements", 0),
+                            "fields": len(ocr_result.get("fields", [])),
+                        },
+                    )
+            except Exception as e:
+                _node_log.warning(f"OCR 字段提取失败: {e}")
+                state["ocr_extraction_result"] = {"error": str(e)}
+        state["uploaded_file_path"] = ""
+
     # 重置本轮请求字段
     state["retry_count"] = 0
     state["user_modification_request"] = user_input

agent/state.py

@@ -40,3 +40,7 @@ class AgentState(TypedDict, total=False):
 
     # 需求6：失败上下文传递 — 重试耗尽后暂存失败信息，下次用户输入时自动注入
     pending_failure_context: dict
+
+    # 需求7：OCR 单据字段精确提取结果
+    ocr_extraction_result: dict
+    uploaded_file_path: str