diff --git a/utils/markdown_to_image.py b/utils/markdown_to_image.py
index 5544a85..ca6ba30 100644
--- a/utils/markdown_to_image.py
+++ b/utils/markdown_to_image.py
@@ -7,6 +7,7 @@ import psutil
from playwright.async_api import async_playwright
import os
import asyncio
+import re
from loguru import logger
@@ -18,11 +19,11 @@ async def safe_close_browser(browser, timeout: float = 4.0) -> None:
for page in context.pages[:]:
try:
await asyncio.wait_for(page.close(), timeout=1.5)
- except:
+ except Exception:
pass
try:
await asyncio.wait_for(context.close(), timeout=timeout)
- except:
+ except Exception:
pass
try:
@@ -66,26 +67,40 @@ async def safe_close_browser(browser, timeout: float = 4.0) -> None:
logger.warning(f"force kill failed: {e}")
+def _split_hero(html_body: str):
+ title_match = re.search(r'(.*?)
', html_body, re.S | re.I)
+ hero_title = title_match.group(1).strip() if title_match else "聊天总结"
+ remain = re.sub(r'.*?
', '', html_body, count=1, flags=re.S | re.I).strip()
+
+ meta_match = re.match(r'^(.*?)
', remain, re.S | re.I)
+ hero_meta = meta_match.group(1).strip() if meta_match else "群聊总结 / 自动生成"
+ if meta_match:
+ remain = re.sub(r'^.*?
', '', remain, count=1, flags=re.S | re.I).strip()
+
+ return hero_title, hero_meta, remain
+
+
async def md_str_to_html_content(md_content):
- """将 Markdown 字符串转换为更有风格的 HTML 内容。"""
+ """将 Markdown 字符串转换为更美观的 HTML 内容。"""
html_body = markdown.markdown(md_content, extensions=['extra', 'codehilite'])
+ hero_title, hero_meta, remain_html = _split_hero(html_body)
css = """