vibego 0.2.32__py3-none-any.whl → 0.2.34__py3-none-any.whl

bot.py

@@ -69,6 +69,7 @@ from tasks.fsm import (
     TaskNoteStates,
     TaskPushStates,
 )
+from telegram_markdown import render_markdown_to_telegram
 
 # --- 简单 .env 加载 ---
 def load_env(p: str = ".env"):
@@ -470,32 +471,6 @@ def _normalize_legacy_markdown(text: str) -> str:
     return "".join(pieces)
 
 
-def _normalize_to_telegram_markdown(text: str) -> str:
-    """将标准 Markdown 格式转换为 Telegram MarkdownV2 支持的格式
-
-    转换规则：
-    - **text** → *text* （加粗）
-    - __text__ → _text_ （斜体）
-    - #### 标题 → *标题* （标题转为加粗）
-    """
-    # 先转换加粗和下划线语法
-    text = _normalize_legacy_markdown(text)
-
-    # 转换标题：移除 # 符号，将标题文本转为加粗
-    def _replace_heading(match: re.Match[str]) -> str:
-        # match.group(1) 是 # 符号
-        # match.group(2) 是标题文本
-        heading_text = match.group(2)
-        # 标题转为加粗文本（如果文本已经是加粗的，避免重复）
-        if heading_text.startswith("*") and heading_text.endswith("*"):
-            return heading_text  # 已经是加粗，保持不变
-        return f"*{heading_text}*"
-
-    text = MARKDOWN_HEADING.sub(_replace_heading, text)
-
-    return text
-
-
 # MarkdownV2 转义字符模式（用于检测已转义文本）
 _ESCAPED_MARKDOWN_PATTERN = re.compile(
     r"\\[_*\[\]()~`>#+=|{}.!:-]"  # 添加了冒号
@@ -623,12 +598,8 @@ def _unescape_if_already_escaped(text: str) -> str:
 
 def _prepare_model_payload(text: str) -> str:
     if _IS_MARKDOWN_V2:
-        # 先反转义已转义的内容，避免双重转义
         cleaned = _unescape_if_already_escaped(text)
-        # 转换标准 Markdown 格式为 Telegram 支持的格式
-        normalized = _normalize_to_telegram_markdown(cleaned)
-        # 再进行 MarkdownV2 转义
-        return _escape_markdown_v2(normalized)
+        return render_markdown_to_telegram(cleaned)
     if _IS_MARKDOWN:
         return _normalize_legacy_markdown(text)
     return text

scripts/requirements.txt

@@ -1,3 +1,4 @@
 aiogram>=3.0.0,<4.0.0
 aiohttp-socks>=0.10.0
 aiosqlite>=0.19.0
+markdown-it-py>=3.0.0,<4.0.0

telegram_markdown/__init__.py

@@ -0,0 +1,269 @@
+"""
+自定义 Markdown → Telegram MarkdownV2 渲染器。
+
+依赖 markdown-it-py 提供的 CommonMark 解析能力，将常见 Markdown
+语法（段落、标题、列表、引用、加粗、斜体、代码块、行内代码、链接等）
+转换为符合 Telegram MarkdownV2 语法的文本。
+
+Telegram 对 MarkdownV2 的语法要求非常严格，特殊字符必须转义，否则会直接
+返回 “can't parse entities” 错误。借助解析树，我们可以准确地区分格式化标记
+与普通文本，做到“格式保留 + 无额外反斜杠”。
+"""
+from __future__ import annotations
+
+import re
+from typing import Iterable, List, Optional, Tuple
+
+from markdown_it import MarkdownIt
+from markdown_it.token import Token
+
+_TELEGRAM_SPECIAL_CHARS = re.compile(r"([_*\[\]()~`>#+\-=|{}.!\\])")
+
+
+def _escape_plain_text(value: str) -> str:
+    """转义 Telegram MarkdownV2 的特殊字符。"""
+    if not value:
+        return ""
+    return _TELEGRAM_SPECIAL_CHARS.sub(r"\\\1", value)
+
+
+def _escape_url(value: str) -> str:
+    """URL 中统一转义括号、空格等特殊字符。"""
+    if not value:
+        return ""
+    escaped = value.replace("\\", "\\\\")
+    escaped = escaped.replace(")", r"\)").replace("(", r"\(")
+    return escaped
+
+
+def _escape_inline_code(value: str) -> str:
+    """行内代码仅需要处理反斜杠与反引号。"""
+    if not value:
+        return "``"
+    escaped = value.replace("\\", "\\\\").replace("`", r"\`")
+    return f"`{escaped}`"
+
+
+def _escape_code_block(value: str, language: str = "") -> str:
+    """多行代码块同样保留原样，只需保护反引号。"""
+    content = (value or "").rstrip("\n")
+    content = content.replace("```", r"\`\`\`")
+    header = f"```{language.strip()}\n" if language else "```\n"
+    return f"{header}{content}\n```"
+
+
+class _TelegramMarkdownRenderer:
+    """对 markdown-it 解析出的 Token 序列进行 Telegram MarkdownV2 序列化。"""
+
+    def __init__(self) -> None:
+        self._parser = MarkdownIt("commonmark", {"linkify": True})
+
+    # ---- 外部接口 ---------------------------------------------------------
+    def render(self, markdown_text: str) -> str:
+        tokens = self._parser.parse(markdown_text or "")
+        return self._render_block_tokens(tokens, 0, len(tokens)).strip()
+
+    # ---- Block 级渲染 -----------------------------------------------------
+    def _render_block_tokens(self, tokens: List[Token], start: int, end: int) -> str:
+        blocks: List[str] = []
+        index = start
+        while index < end:
+            token = tokens[index]
+            token_type = token.type
+
+            if token_type == "paragraph_open":
+                inline = tokens[index + 1]
+                blocks.append(self._render_inline(inline.children or []))
+                index += 3  # 跳过 inline + paragraph_close
+                continue
+
+            if token_type == "heading_open":
+                inline = tokens[index + 1]
+                content = self._render_inline(inline.children or [])
+                blocks.append(f"*{content}*")
+                index += 3
+                continue
+
+            if token_type == "bullet_list_open":
+                close_idx = self._find_matching(tokens, index, "bullet_list_close")
+                blocks.append(self._render_list(tokens, index + 1, close_idx, ordered=False))
+                index = close_idx + 1
+                continue
+
+            if token_type == "ordered_list_open":
+                close_idx = self._find_matching(tokens, index, "ordered_list_close")
+                start_number = int(token.attrGet("start") or "1")
+                blocks.append(
+                    self._render_list(tokens, index + 1, close_idx, ordered=True, start_number=start_number)
+                )
+                index = close_idx + 1
+                continue
+
+            if token_type == "blockquote_open":
+                close_idx = self._find_matching(tokens, index, "blockquote_close")
+                quoted = self._render_block_tokens(tokens, index + 1, close_idx)
+                lines = quoted.splitlines()
+                blocks.append("\n".join(f"> {line}" if line else ">" for line in lines))
+                index = close_idx + 1
+                continue
+
+            if token_type in {"fence", "code_block"}:
+                language = token.info or "" if token_type == "fence" else ""
+                blocks.append(_escape_code_block(token.content, language))
+                index += 1
+                continue
+
+            if token_type == "inline":
+                blocks.append(self._render_inline(token.children or []))
+                index += 1
+                continue
+
+            if token_type == "hr":
+                blocks.append("―" * 3)
+                index += 1
+                continue
+
+            # 其他 Block token（如 html_block）直接忽略或转换为原始文本
+            if token_type == "html_block":
+                blocks.append(_escape_plain_text(token.content))
+                index += 1
+                continue
+
+            index += 1
+
+        return "\n\n".join(part for part in blocks if part)
+
+    def _render_list(
+        self,
+        tokens: List[Token],
+        start: int,
+        end: int,
+        *,
+        ordered: bool,
+        start_number: int = 1,
+    ) -> str:
+        parts: List[str] = []
+        index = start
+        counter = start_number
+        while index < end:
+            token = tokens[index]
+            if token.type != "list_item_open":
+                index += 1
+                continue
+
+            close_idx = self._find_matching(tokens, index, "list_item_close")
+            body = self._render_block_tokens(tokens, index + 1, close_idx)
+            lines = body.splitlines()
+            prefix = f"{counter}. " if ordered else "- "
+            if lines:
+                rendered = [prefix + lines[0]]
+                rendered.extend(("  " if ordered else "  ") + line for line in lines[1:])
+                parts.append("\n".join(rendered))
+            else:
+                parts.append(prefix.rstrip())
+
+            if ordered:
+                counter += 1
+            index = close_idx + 1
+
+        return "\n".join(parts)
+
+    # ---- Inline 渲染 ------------------------------------------------------
+    def _render_inline(self, tokens: Iterable[Token]) -> str:
+        rendered, _ = self._render_inline_segment(list(tokens), 0, None)
+        return rendered
+
+    def _render_inline_segment(
+        self,
+        tokens: List[Token],
+        start: int,
+        closing_type: Optional[str],
+    ) -> Tuple[str, int]:
+        parts: List[str] = []
+        index = start
+
+        while index < len(tokens):
+            token = tokens[index]
+            token_type = token.type
+
+            if closing_type and token_type == closing_type:
+                return "".join(parts), index + 1
+
+            if token_type == "text":
+                parts.append(_escape_plain_text(token.content))
+                index += 1
+                continue
+
+            if token_type in {"softbreak", "hardbreak"}:
+                parts.append("\n")
+                index += 1
+                continue
+
+            if token_type == "code_inline":
+                parts.append(_escape_inline_code(token.content))
+                index += 1
+                continue
+
+            if token_type == "strong_open":
+                inner, next_index = self._render_inline_segment(tokens, index + 1, "strong_close")
+                parts.append(f"*{inner}*")
+                index = next_index
+                continue
+
+            if token_type == "em_open":
+                inner, next_index = self._render_inline_segment(tokens, index + 1, "em_close")
+                parts.append(f"_{inner}_")
+                index = next_index
+                continue
+
+            if token_type == "link_open":
+                href = token.attrGet("href") or ""
+                inner, next_index = self._render_inline_segment(tokens, index + 1, "link_close")
+                parts.append(f"[{inner}]({_escape_url(href)})")
+                index = next_index
+                continue
+
+            if token_type == "image":
+                alt_text = self._render_inline(token.children or [])
+                parts.append(f"[{alt_text}]")
+                index += 1
+                continue
+
+            if token_type == "html_inline":
+                parts.append(_escape_plain_text(token.content))
+                index += 1
+                continue
+
+            if token_type == "strikethrough_open":
+                inner, next_index = self._render_inline_segment(tokens, index + 1, "strikethrough_close")
+                parts.append(f"~{inner}~")
+                index = next_index
+                continue
+
+            index += 1
+
+        return "".join(parts), len(tokens)
+
+    # ---- 工具方法 ---------------------------------------------------------
+    @staticmethod
+    def _find_matching(tokens: List[Token], start_idx: int, closing_type: str) -> int:
+        """使用 nesting 深度寻找匹配的关闭 token。"""
+        open_type = tokens[start_idx].type
+        depth = 0
+        for idx in range(start_idx, len(tokens)):
+            token = tokens[idx]
+            if token.type == open_type and token.nesting == 1:
+                depth += 1
+            elif token.type == closing_type and token.nesting == -1:
+                depth -= 1
+                if depth == 0:
+                    return idx
+        raise ValueError(f"未找到匹配的 {closing_type}")
+
+
+_RENDERER = _TelegramMarkdownRenderer()
+
+
+def render_markdown_to_telegram(markdown_text: str) -> str:
+    """将 Markdown 文本转换为 Telegram MarkdownV2。"""
+    return _RENDERER.render(markdown_text)

{vibego-0.2.32.dist-info → vibego-0.2.34.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vibego
-Version: 0.2.32
+Version: 0.2.34
 Summary: vibego CLI：用于初始化与管理 Telegram Master Bot 的工具
 Author: Hypha
 License-Expression: LicenseRef-Proprietary
@@ -15,6 +15,7 @@ Description-Content-Type: text/markdown
 Requires-Dist: aiogram<4.0.0,>=3.0.0
 Requires-Dist: aiohttp-socks>=0.10.0
 Requires-Dist: aiosqlite>=0.19.0
+Requires-Dist: markdown-it-py<4.0.0,>=3.0.0
 
 # vibe-bot（Telegram → Mac CLI → Telegram 回推）
 

{vibego-0.2.32.dist-info → vibego-0.2.34.dist-info}/RECORD

@@ -1,4 +1,4 @@
-bot.py,sha256=uJcmMJmj5SlpwUXSgtnIEeNiUGkBg-lG5g6kM9BobP4,272046
+bot.py,sha256=fOlusQvWmjL8z6w-QY9IQ3RxbsEgCL5weaYSiDl6Tdw,270925
 logging_setup.py,sha256=gvxHi8mUwK3IhXJrsGNTDo-DR6ngkyav1X-tvlBF_IE,4613
 master.py,sha256=ZW4A3Gh0MUKFnfZX-VJ7OCNnBzlEcOWkPKYH92OfyKA,112967
 project_repository.py,sha256=UcthtSGOJK0cTE5bQCneo3xkomRG-kyc1N1QVqxeHIs,17577
@@ -7,7 +7,7 @@ scripts/bump_version.sh,sha256=a4uB8V8Y5LPsoqTCdzQKsEE8HhwpBmqRaQInG52LDig,4089
 scripts/log_writer.py,sha256=8euoMlRo7cbtHApbcEoJnwzLABxti-ovJWFLRN1oDQw,3843
 scripts/master_healthcheck.py,sha256=-X0VVsZ0AXaOb7izxTO_oyu23g_1jsirNdGIcP8nrSI,8321
 scripts/publish.sh,sha256=ehLfMedcXuGKJ87jpZy3kuiFszG9Cpavp3zXPfR4h-g,3511
-scripts/requirements.txt,sha256=QSt30DSSSHtfucTFPpc7twk9kLS5rVLNTcvDiagxrZg,62
+scripts/requirements.txt,sha256=ukJbFLJyzqnQYMz6j07O-IOrG87IwXg0oikmn1nfJ9M,91
 scripts/run_bot.sh,sha256=rN4K1nz041XBaUJmnBBKHS2cHmQf11vPNX8wf1hbVR4,4596
 scripts/start.sh,sha256=w1Q35NB-9FRZAez5I5veqgYIJ8XnVukGt8TTE_ad248,13608
 scripts/start_tmux_codex.sh,sha256=xyLv29p924q-ysxvZYAP3T6VrqLPBPMBWo9QP7cuL50,4438
@@ -426,14 +426,15 @@ tasks/constants.py,sha256=tS1kZxBIUm3JJUMHm25XI-KHNUZl5NhbbuzjzL_rF-c,299
 tasks/fsm.py,sha256=rKXXLEieQQU4r2z_CZUvn1_70FXiZXBBugF40gpe_tQ,1476
 tasks/models.py,sha256=N_qqRBo9xMSV0vbn4k6bLBXT8C_dp_oTFUxvdx16ZQM,2459
 tasks/service.py,sha256=w_S_aWiVqRXzXEpimLDsuCCCX2lB5uDkff9aKThBw9c,41916
-vibego_cli/__init__.py,sha256=yIxuD2p_Dd2qwaOZwbuf-21aP6wKVdMFcPi5wJfYlos,311
+telegram_markdown/__init__.py,sha256=bG3H9fWn5GfTqC6xvd49xbVdYWfSFeaX2nefweOYcWY,9757
+vibego_cli/__init__.py,sha256=YiGf4wB8hakiJXIr-AwQ78eW9g7s9QwQN7KTX7AWhAA,311
 vibego_cli/__main__.py,sha256=qqTrYmRRLe4361fMzbI3-CqpZ7AhTofIHmfp4ykrrBY,158
 vibego_cli/config.py,sha256=VxkPJMq01tA3h3cOkH-z_tiP7pMgfSGGicRvUnCWkhI,3054
 vibego_cli/deps.py,sha256=1nRXI7Dd-S1hYE8DligzK5fIluQWETRUj4_OKL0DikQ,1419
 vibego_cli/main.py,sha256=X__NXwZnIDIFbdKSTbNyZgZHKcPlN0DQz9sqTI1aQ9E,12158
 vibego_cli/data/worker_requirements.txt,sha256=QSt30DSSSHtfucTFPpc7twk9kLS5rVLNTcvDiagxrZg,62
-vibego-0.2.32.dist-info/METADATA,sha256=Dm8z912xz2fEhlZLca0kzY-tsZEzWqinWmsEB2ycaGQ,10475
-vibego-0.2.32.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-vibego-0.2.32.dist-info/entry_points.txt,sha256=Lsy_zm-dlyxt8-9DL9blBReIwU2k22c8-kifr46ND1M,48
-vibego-0.2.32.dist-info/top_level.txt,sha256=R56CT3nW5H5v3ce0l3QDN4-C4qxTrNWzRTwrxnkDX4U,69
-vibego-0.2.32.dist-info/RECORD,,
+vibego-0.2.34.dist-info/METADATA,sha256=eFQD8C0DvvCHPbvuvAI5h56kh2iqp8Ty5LN-ILJG2aY,10519
+vibego-0.2.34.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+vibego-0.2.34.dist-info/entry_points.txt,sha256=Lsy_zm-dlyxt8-9DL9blBReIwU2k22c8-kifr46ND1M,48
+vibego-0.2.34.dist-info/top_level.txt,sha256=rWDj9KERtbJL6Lar9Xa0O6dthaFSY_jc1WNpQgUrXCM,87
+vibego-0.2.34.dist-info/RECORD,,

{vibego-0.2.32.dist-info → vibego-0.2.34.dist-info}/top_level.txt

@@ -4,4 +4,5 @@ master
 project_repository
 scripts
 tasks
+telegram_markdown
 vibego_cli

vibego_cli/__init__.py

@@ -7,6 +7,6 @@ from __future__ import annotations
 
 __all__ = ["main", "__version__"]
 
-__version__ = "0.2.32"
+__version__ = "0.2.34"
 
 from .main import main  # noqa: E402