fix/ 修复AI聊天时回复markdown导致聊天气泡布局问题
This commit is contained in:
@@ -5,6 +5,44 @@ from __future__ import annotations
|
||||
import re
|
||||
|
||||
|
||||
def strip_markdown_for_chat(text: str) -> str:
|
||||
"""
|
||||
将模型偶然输出的常见 Markdown 剥成纯文本,供 App 聊天气泡展示。
|
||||
保留换行与字面量 [SPLIT];不做完整 MD 解析,以简单可预测为主。
|
||||
"""
|
||||
if not text:
|
||||
return text
|
||||
s = text
|
||||
# 围栏代码块(含首行语言标记):整段替换为块内正文,去掉栅栏
|
||||
s = re.sub(
|
||||
r"```(?:[^\n`]*)\n([\s\S]*?)```",
|
||||
r"\1",
|
||||
s,
|
||||
flags=re.MULTILINE,
|
||||
)
|
||||
s = s.replace("```", "")
|
||||
# 图片  → alt;链接 [label](url) → label
|
||||
s = re.sub(r"!\[([^\]]*)\]\([^)]*\)", r"\1", s)
|
||||
s = re.sub(r"\[([^\]]*)\]\([^)]*\)", r"\1", s)
|
||||
# ATX 标题
|
||||
s = re.sub(r"(?m)^#{1,6}\s+", "", s)
|
||||
# 无序列表行首(仅限行首减号/星号/+ 后接空格,避免误判「—」)
|
||||
s = re.sub(r"(?m)^\s*[-*+]\s+", "", s)
|
||||
# 有序列表「数字. 」仅行首
|
||||
s = re.sub(r"(?m)^\s*\d+\.\s+", "", s)
|
||||
# 粗体/删除线常见标记
|
||||
s = s.replace("**", "").replace("__", "")
|
||||
s = s.replace("~~", "")
|
||||
# 行内反引号
|
||||
s = s.replace("`", "")
|
||||
# 孤立 emphasis:*词* 或 _词_(不含跨行)
|
||||
s = re.sub(r"(?<![*])\*([^*\n]+)\*(?![*])", r"\1", s)
|
||||
s = re.sub(r"(?<![_])_([^_\n]+)_(?![_])", r"\1", s)
|
||||
# 分割线
|
||||
s = re.sub(r"(?m)^\s*---+\s*$", "", s)
|
||||
return s
|
||||
|
||||
|
||||
def segments_from_llm_response(
|
||||
response_text: str,
|
||||
*,
|
||||
@@ -15,7 +53,7 @@ def segments_from_llm_response(
|
||||
优先按字面 [SPLIT] 拆段;若模型只输出一段、但用空行写了多段,再按段落拆。
|
||||
解决「两段话 + 换行」却未写 [SPLIT] 时仍要拆气泡 / 多段 TTS 的情况。
|
||||
"""
|
||||
text = (response_text or "").strip()
|
||||
text = strip_markdown_for_chat((response_text or "").strip())
|
||||
if not text:
|
||||
return []
|
||||
primary = [p.strip() for p in text.split("[SPLIT]") if p.strip()]
|
||||
|
||||
Reference in New Issue
Block a user