Fixed the display of charts and lists

马一丁
Commit 6c643791425be26e35eda1d6626387b9375abd31 6c643791 1 parent 26d6c2c1
Showing 2 changed files with 110 additions and 2 deletions
ReportEngine/renderers/html_renderer.py
ReportEngine/renderers/markdown_renderer.py
--- a/ReportEngine/renderers/html_renderer.py
View file @6c64379
+++ b/ReportEngine/renderers/html_renderer.py
View file @6c64379
@@ -885,7 +885,12 @@ class HTMLRenderer:
         """粗略判断dict是否符合block结构"""
         if not isinstance(payload, dict):
             return False
-        if "type" in payload and isinstance(payload["type"], str):
+        block_type = payload.get("type")
+        if block_type and isinstance(block_type, str):
+            # 排除内联类型（inlineRun 等），它们不是块级元素
+            inline_types = {"inlineRun", "inline", "text"}
+            if block_type in inline_types:
+                return False
             return True
         structural_keys = {"blocks", "rows", "items", "widgetId", "widgetType", "data"}
         return any(key in payload for key in structural_keys)
@@ -896,6 +901,12 @@ class HTMLRenderer:
         if isinstance(payload, dict):
             block_list = payload.get("blocks")
             block_type = payload.get("type")
+            
+            # 排除内联类型，它们不是块级元素
+            inline_types = {"inlineRun", "inline", "text"}
+            if block_type in inline_types:
+                return collected
+            
             if isinstance(block_list, list) and not block_type:
                 for candidate in block_list:
                     collected.extend(self._collect_blocks_from_payload(candidate))
@@ -2933,6 +2944,19 @@ class HTMLRenderer:
         if not isinstance(run, dict):
             return ("" if run is None else str(run)), []
+        # 处理 inlineRun 类型：递归展开其 inlines 数组
+        if run.get("type") == "inlineRun":
+            inner_inlines = run.get("inlines") or []
+            outer_marks = run.get("marks") or []
+            # 递归合并所有内部 inlines 的文本
+            texts = []
+            all_marks = list(outer_marks)
+            for inline in inner_inlines:
+                inner_text, inner_marks = self._normalize_inline_payload(inline)
+                texts.append(inner_text)
+                all_marks.extend(inner_marks)
+            return "".join(texts), all_marks
+
         marks = list(run.get("marks") or [])
         text_value: Any = run.get("text", "")
         seen: set[int] = set()
@@ -2980,6 +3004,9 @@ class HTMLRenderer:
                     else:
                         inline_payload = self._coerce_inline_payload(payload)
                         if inline_payload:
+                            # 处理 inlineRun 类型
+                            if inline_payload.get("type") == "inlineRun":
+                                return self._normalize_inline_payload(inline_payload)
                             nested_text = inline_payload.get("text")
                             if nested_text is not None:
                                 text_value = nested_text
@@ -3073,9 +3100,12 @@ class HTMLRenderer:
         if not isinstance(payload, dict):
             return None
         inline_type = payload.get("type")
+        # 支持 inlineRun 类型：包含嵌套的 inlines 数组
+        if inline_type == "inlineRun":
+            return payload
         if inline_type and inline_type not in {"inline", "text"}:
             return None
-        if "text" not in payload and "marks" not in payload:
+        if "text" not in payload and "marks" not in payload and "inlines" not in payload:
             return None
         return payload
--- a/ReportEngine/renderers/markdown_renderer.py
View file @6c64379
+++ b/ReportEngine/renderers/markdown_renderer.py
View file @6c64379
@@ -647,11 +647,29 @@ class MarkdownRenderer:
     def _render_inline_run(self, run: Any, for_table: bool = False) -> str:
         if isinstance(run, dict):
+            # 处理 inlineRun 类型：嵌套的 inlines 数组
+            if run.get("type") == "inlineRun":
+                inner_inlines = run.get("inlines") or []
+                outer_marks = run.get("marks") or []
+                # 递归渲染内部的 inlines
+                inner_text = self._render_inlines(inner_inlines, for_table=for_table)
+                # 应用外层的 marks
+                result = inner_text
+                for mark in outer_marks:
+                    result = self._apply_mark(result, mark)
+                return result
             text = run.get("text", "")
             marks = run.get("marks") or []
         else:
             text = run if isinstance(run, str) else ""
             marks = []
+        
+        # 尝试检测并解析被错误序列化为字符串的 inlineRun JSON
+        if isinstance(text, str) and text.startswith('{"type": "inlineRun"'):
+            parsed = self._try_parse_inline_run_string(text)
+            if parsed:
+                return self._render_inline_run(parsed, for_table=for_table)
+        
         result = self._escape_text(text, for_table=for_table)
         for mark in marks:
             if not isinstance(mark, dict):
@@ -683,6 +701,66 @@ class MarkdownRenderer:
             # 颜色/字体等非通用标记直接降级为纯文本
         return result
+    def _apply_mark(self, text: str, mark: Any) -> str:
+        """
+        对文本应用单个 mark 格式。
+        
+        用于处理 inlineRun 类型的外层 marks。
+        """
+        if not isinstance(mark, dict):
+            return text
+        mtype = mark.get("type")
+        if mtype == "bold":
+            return f"**{text}**"
+        elif mtype == "italic":
+            return f"*{text}*"
+        elif mtype == "underline":
+            return f"__{text}__"
+        elif mtype == "strike":
+            return f"~~{text}~~"
+        elif mtype == "code":
+            return f"`{text}`"
+        elif mtype == "link":
+            href = mark.get("href") or mark.get("value")
+            href = str(href) if href else ""
+            return f"[{text}]({href})" if href else text
+        elif mtype == "highlight":
+            return f"=={text}=="
+        elif mtype == "subscript":
+            return f"~{text}~"
+        elif mtype == "superscript":
+            return f"^{text}^"
+        elif mtype == "math":
+            latex = self._normalize_math(mark.get("value") or text)
+            return f"${latex}$" if latex else text
+        return text
+
+    def _try_parse_inline_run_string(self, text: str) -> dict | None:
+        """
+        尝试解析被错误序列化为字符串的 inlineRun JSON。
+        
+        某些 LLM 生成的内容会将 inlineRun 结构意外地作为字符串
+        存入 text 字段，本方法尝试识别并解析这种情况。
+        
+        参数:
+            text: 可能包含 JSON 的字符串
+            
+        返回:
+            dict | None: 解析成功返回 inlineRun 字典，否则返回 None
+        """
+        if not text or not isinstance(text, str):
+            return None
+        text = text.strip()
+        if not text.startswith('{"type": "inlineRun"'):
+            return None
+        try:
+            parsed = json.loads(text)
+            if isinstance(parsed, dict) and parsed.get("type") == "inlineRun":
+                return parsed
+        except json.JSONDecodeError:
+            pass
+        return None
+
     def _is_heading_duplicate(self, block: Dict[str, Any], chapter_title: str | None) -> bool:
         """判断首个heading是否与章节标题重复"""
         if not isinstance(block, dict) or block.get("type") != "heading":