feat: 提取凭证摘要(摘要)

- parser: 新增摘要提取逻辑，从 OCR 结果中筛选非代码/非金额文本
- 去掉 skip_keywords 中的日/期/年/月（误杀摘要文本）
- search_lines 用完整 OCR 行（不被合计行截断）
- 45张测试: 前30行中20行成功提取真实摘要

fix: remove debug prints, reject bank account/status texts

core/parser.py

@@ -135,7 +135,7 @@ class K3VoucherParser:
         skip_keywords = ['科目', '摘要', '借方', '贷方', '出纳',
                          '制单', '审核', '经办', '过账', '附件',
                          'Manager', 'HONOR', 'Magic6', '合计',
-                         '日', '期', '年', '月', '凭证', '号']
+                         '凭证', '号']
         
         for text, conf, cx, cy, box in data_lines:
             if any(k in text for k in skip_keywords):
@@ -165,6 +165,48 @@ class K3VoucherParser:
         if not codes:
             return self._parse_entries_fallback()
         
+        # === 4.5 摘要提取：从剩余 OCR 文字中找摘要 ===
+        code_y_min = min(c['y_pos'] for c in codes)
+        desc_candidates = []
+        search_lines = self.lines
+        for text, conf, cx, cy, box in search_lines:
+            if conf < 0.5:
+                continue
+            is_code = False
+            for c in codes:
+                if abs(c['y_pos'] - cy) < 20:
+                    is_code = True
+                    break
+            if is_code:
+                continue
+            if re.search(r'\d+\.\d{2}', text):
+                continue
+            if any(k in text for k in skip_keywords + ['合计', '亿', '千', '百', '拾', '万']):
+                continue
+            if re.match(r'^[\d.\-,]+$', text.strip()):
+                continue
+            if len(text.strip()) < 3:
+                continue
+            if cx < 300:
+                continue
+            if text.strip().startswith('·') or '工商银行' in text:
+                continue
+            if '支行' in text or re.search(r'\d{15,}', text):
+                continue
+            if text.strip() in ('未到账', '已到账', '未达', '已达'):
+                continue
+            if cy < code_y_min - 50:
+                continue
+            # 去重：跳过 y 距离 < 30 的重复文本
+            if any(abs(d['y'] - cy) < 30 for d in desc_candidates):
+                continue
+            desc_candidates.append({'text': text, 'y': cy, 'x': cx})
+        
+        desc_candidates.sort(key=lambda d: d['y'])
+        for i, desc in enumerate(desc_candidates):
+            if i < len(codes):
+                codes[i]['desc'] = desc['text']
+        
         # === 5. 提取金额 ===
         amounts = []
         skip_amount_keywords = ['Manager', 'HONOR', 'Magic6', 'RMB',