fix

posterbuilder/convert.py

@@ -77,7 +77,92 @@ GREEK_OR_MATH_MACROS = (
     r"Gamma|Delta|Theta|Lambda|Xi|Pi|Sigma|Upsilon|Phi|Psi|Omega"
 )
 
+MATH_INLINE_MACROS = (
+    GREEK_OR_MATH_MACROS  
+    + r"|partial|nabla|infty|cdot|times|pm|leq|geq|ldots|dots"
+)
+
+_MACRO_OUTSIDE_MATH_RE = re.compile(
+    rf"(\\(?:{MATH_INLINE_MACROS}))"          # \delta / \tau / \cdot / ...
+    rf"(?:\s*[A-Za-z])?",                     # 允许后面紧跟一个字母变量（如 \delta c）
+)
+
+_BULLET_RE = re.compile(r"•")
+
 # ===================== 基础工具 =====================
+# 覆盖的数学块（全局已有 MATH_BLOCK_RE，可以复用）
+def wrap_math_macros_outside_math(s: str) -> str:
+    """
+    目的：在“非数学环境”里遇到数学宏时，自动加上 $...$。
+    例如：\delta c  ->  $\delta c$
+          \tau      ->  $\tau$
+    已有的 $...$ / \[...\] / \(...\) 不会被二次处理（先暂存）。
+    """
+    if not s:
+        return s
+
+    # 1) 暂存已有数学块
+    stash = []
+    def _hide(m):
+        stash.append(m.group(0))
+        return f"\x00M{len(stash)-1}\x00"
+    s_hidden = MATH_BLOCK_RE.sub(_hide, s)
+
+    # 2) 把裸奔宏包进 $...$
+    def _wrap(m):
+        return f"${m.group(0)}$"
+    s_hidden = _MACRO_OUTSIDE_MATH_RE.sub(_wrap, s_hidden)
+
+    # 3) 还原数学块
+    for i, blk in enumerate(stash):
+        s_hidden = s_hidden.replace(f"\x00M{i}\x00", blk)
+
+    return s_hidden
+
+
+def wrap_math_macros_outside_math(s: str) -> str:
+    """
+    目的：在“非数学环境”里遇到数学宏时，自动加上 $...$。
+    例如：\delta c  ->  $\delta c$
+          \tau      ->  $\tau$
+    已有的 $...$ / \[...\] / \(...\) 不会被二次处理（先暂存）。
+    """
+    if not s:
+        return s
+
+    # 1) 暂存已有数学块
+    stash = []
+    def _hide(m):
+        stash.append(m.group(0))
+        return f"\x00M{len(stash)-1}\x00"
+    s_hidden = MATH_BLOCK_RE.sub(_hide, s)
+
+    # 2) 把裸奔宏包进 $...$
+    def _wrap(m):
+        return f"${m.group(0)}$"
+    s_hidden = _MACRO_OUTSIDE_MATH_RE.sub(_wrap, s_hidden)
+
+    # 3) 还原数学块
+    for i, blk in enumerate(stash):
+        s_hidden = s_hidden.replace(f"\x00M{i}\x00", blk)
+
+    return s_hidden
+
+
+_BULLET_RE = re.compile(r"•")
+
+def normalize_inline_bullets(s: str) -> str:
+    """
+    把 Unicode 的 • 统一替换为 LaTeX 的 \\textbullet{}，并确保两侧留空格。
+    """
+    if not s:
+        return s
+    s = _BULLET_RE.sub(r"\\textbullet{}", s)
+    # 若两侧无空格，补空格（避免“黏连”）
+    s = re.sub(r"(?<=\S)\\textbullet\{\}(?=\S)", r" \\textbullet{} ", s)
+    s = re.sub(r"\\textbullet\{\}(?=\S)", r"\\textbullet{} ", s)
+    s = re.sub(r"(?<=\S)\\textbullet\{\}", r" \\textbullet{}", s)
+    return s
 
 def normalize_textit_math(s: str) -> str:
     """
@@ -153,29 +238,25 @@ def fix_latex_escaped_commands(s: str) -> str:
     s = s.replace("\\}", "}")
     return s
 
-import re
 
 def escape_text(s: str) -> str:
     if not s:
         return ""
 
-    # --- 1) 捕获 math: $...$ 或 \( ... \) ----
+    # --- 1) 捕获所有数学块（沿用全局 MATH_BLOCK_RE）----
     math_blocks = []
     def store_math(m):
         math_blocks.append(m.group(0))
         return f"\0{len(math_blocks)-1}\0"
 
-    # 识别 $...$ 和 \( ... \)
-    s = re.sub(r"\${1,2}.*?\${1,2}|\\\(.+?\\\)", store_math, s)
+    s = MATH_BLOCK_RE.sub(store_math, s)
 
     # --- 2) 转义文本字符（不碰 math） ----
     rep = {
         "&": r"\&", "%": r"\%", "$": r"\$", "#": r"\#",
         "_": r"\_", "{": r"\{", "}": r"\}",
-        "~": r"~{}",  # 保持 LaTeX 不换行空格
-        "^": r"\^{}",
+        "~": r"~{}", "^": r"\^{}",
     }
-
     for k, v in rep.items():
         s = s.replace(k, v)
 
@@ -187,6 +268,7 @@ def escape_text(s: str) -> str:
 
 
 
+
 def soft_wrap_title_for_logo(title: str, first_limit=68, next_limit=72) -> str:
     if not title or len(title) <= first_limit: return title
     def break_at(s: str, limit: int):
@@ -230,17 +312,25 @@ def format_content_to_latex(content: str) -> str:
     """格式化正文内容，自动修复 LaTeX 命令"""
     if not content:
         return ""
-    # 🔧 新增：先修复由于 \t 被误解析的 LaTeX 命令
+
+    # 1) 先修复 \t 造成的命令断头
     content = fix_latex_escaped_commands(content)
-    # NEW: 规范 \textit{...} 中的“伪数学”写法
+
+    # 2) 规范 \textit{...} 里的“伪数学”
     content = normalize_textit_math(content)
 
+    # 3) **把非数学环境的数学宏包进 $...$**  ← NEW（修正 \delta c）
+    content = wrap_math_macros_outside_math(content)
+
+    # 之后再进行 itemize 的识别与转义
     lines = [ln.strip() for ln in content.splitlines() if ln.strip()]
     if lines and all(ln.startswith(("-", "•")) for ln in lines):
         items = [escape_text(ln.lstrip("-• ").strip()) for ln in lines]
         return "\n".join(["\\begin{itemize}"] + [f"\\item {it}" for it in items] + ["\\end{itemize}"])
+
     return escape_text(" ".join(lines))
 
+
 def make_block(title: str, content: str, figures_tex: str = "") -> str:
     body = format_content_to_latex(content or "")
     if figures_tex: body = (body + "\n\n" if body else "") + figures_tex
@@ -581,8 +671,8 @@ def build():
     cleaned_tex = cleaned_tex.replace(r"\{", "{")
     cleaned_tex = cleaned_tex.replace(r"\}", "}")
     # 注意：要先处理上面的大括号再处理反斜杠，否则会提前破坏结构
-    cleaned_tex = cleaned_tex.replace(r"\\\\", r"\\")  # 避免双转义干扰
-    cleaned_tex = cleaned_tex.replace(r"\\", "\\")      # 最终将 \\ → \
+    # cleaned_tex = cleaned_tex.replace(r"\\\\", r"\\")  # 避免双转义干扰
+    # cleaned_tex = cleaned_tex.replace(r"\\", "\\")      # 最终将 \\ → \
     cleaned_tex = cleaned_tex.replace(r"\t\t", "\\t")
     cleaned_tex = strip_stray_t(cleaned_tex)