fix: preserve empty markdown table cells (#1806)

docs/CHANGELOG.md

@@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/).
 
 ## [Unreleased]
 
+- [修复] 修复通知 Markdown 表格转换在空单元格后将后续内容错配到错误表头的问题。
 - [修复] 将 Docker 可安装的 Longbridge SDK 版本固定为 0.2.75，避免 `longbridge>=0.2.77` 从包索引消失后导致 docker-build 失败。
 
 - [改进] Web 设置页新增首次启动配置检查卡，串联基础配置状态、自选股入口、模型配置入口和一次简短试跑。

src/formatters.py

@@ -458,8 +458,12 @@ def _is_markdown_table_separator(row: str) -> bool:
 
 
 def _parse_markdown_table_row(row: str) -> List[str]:
-    cells = [c.strip() for c in row.strip().strip('|').split('|')]
-    return [c for c in cells if c]
+    stripped = row.strip()
+    if stripped.startswith('|'):
+        stripped = stripped[1:]
+    if stripped.endswith('|'):
+        stripped = stripped[:-1]
+    return [c.strip() for c in stripped.split('|')]
 
 
 def _strip_inline_markdown(text: str) -> str:
@@ -500,6 +504,8 @@ def _flush_table_as_key_value_rows(buffer: List[str], output: List[str], *, bull
     header = rows[0]
     data_rows = rows[1:] if len(rows) > 1 else []
     for row in data_rows:
+        if len(row) < len(header):
+            row = row + [""] * (len(header) - len(row))
         if len(header) == 2 and len(row) >= 2:
             output.append(f"{bullet} {_format_two_column_table_row(header, row)}")
             continue
@@ -772,8 +778,7 @@ def format_feishu_markdown(content: str) -> str:
 
         def _parse_row(row: str) -> List[str]:
             """解析表格行，提取单元格"""
-            cells = [c.strip() for c in row.strip().strip('|').split('|')]
-            return [c for c in cells if c]
+            return _parse_markdown_table_row(row)
 
         rows = []
         for raw in buffer:
@@ -790,6 +795,8 @@ def format_feishu_markdown(content: str) -> str:
         header = rows[0]
         data_rows = rows[1:] if len(rows) > 1 else []
         for row in data_rows:
+            if len(row) < len(header):
+                row = row + [""] * (len(header) - len(row))
             pairs = []
             for idx, cell in enumerate(row):
                 key = header[idx] if idx < len(header) else f"列{idx + 1}"

tests/test_formatters.py

@@ -342,6 +342,21 @@ class TestNotificationMarkdownFormatters(unittest.TestCase):
         self.assertNotIn("价格指标：MA5", result)
         self.assertNotIn("类型：N/A", result)
 
+    def test_markdown_tables_to_key_value_rows_preserves_empty_cells(self):
+        text = (
+            "| Header1 | Header2 | Header3 |\n"
+            "|---------|---------|---------|\n"
+            "| Value1  |         | Value3  |\n"
+            "| Tail1   | Tail2   |         |"
+        )
+        colon = "\uff1a"
+
+        result = markdown_tables_to_key_value_rows(text)
+
+        self.assertIn(f"Header1{colon}Value1 | Header2{colon} | Header3{colon}Value3", result)
+        self.assertIn(f"Header1{colon}Tail1 | Header2{colon}Tail2 | Header3{colon}", result)
+        self.assertNotIn(f"Header2{colon}Value3", result)
+
     def test_markdown_tables_to_key_value_rows_keeps_fenced_code_tables(self):
         text = (
             "```markdown\n"
@@ -372,6 +387,19 @@ class TestNotificationMarkdownFormatters(unittest.TestCase):
         self.assertIn("• 股票：600519 | 信号：强势", result)
         self.assertIn("• 关注量能", result)
 
+    def test_feishu_formatter_preserves_empty_table_cells(self):
+        text = (
+            "| Header1 | Header2 | Header3 |\n"
+            "|---------|---------|---------|\n"
+            "| Value1  |         | Value3  |"
+        )
+        colon = "\uff1a"
+
+        result = format_feishu_markdown(text)
+
+        self.assertIn(f"Header1{colon}Value1 | Header2{colon} | Header3{colon}Value3", result)
+        self.assertNotIn(f"Header2{colon}Value3", result)
+
     def test_telegram_formatter_uses_supported_markdown(self):
         text = "## 日报\n\n| 股票 | 信号 |\n| --- | --- |\n| 600519 | 强势 |\n\n[详情](https://example.com/report)"