vibego 0.2.33__tar.gz → 0.2.35__tar.gz

{vibego-0.2.33 → vibego-0.2.35}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vibego
-Version: 0.2.33
+Version: 0.2.35
 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-0.2.33 → vibego-0.2.35}/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
@@ -723,12 +694,20 @@ async def _send_with_markdown_guard(
         if raw_sender is None:
             raise
 
+        # Markdown 彻底失败时退回纯文本发送，需要先清理 MarkdownV2 转义符号
+        fallback_payload = _unescape_if_already_escaped(text)
+        if fallback_payload != text:
+            worker_log.debug(
+                "Markdown 降级为纯文本发送，已移除转义符号",
+                extra={"original_length": len(text), "fallback_length": len(fallback_payload)},
+            )
+
         worker_log.warning(
             "Markdown 解析仍失败，将以纯文本发送",
             extra={"length": len(text)},
         )
-        await raw_sender(text)
-        return text
+        await raw_sender(fallback_payload)
+        return fallback_payload
 
 
 async def _notify_send_failure_message(chat_id: int) -> None:

{vibego-0.2.33 → vibego-0.2.35}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "vibego"
-version = "0.2.33"
+version = "0.2.35"
 description = "vibego CLI：用于初始化与管理 Telegram Master Bot 的工具"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -16,6 +16,7 @@ dependencies = [
     "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",
 ]
 classifiers = [
     "Programming Language :: Python :: 3",
@@ -34,14 +35,14 @@ include-package-data = true
 py-modules = ["bot", "master", "logging_setup", "project_repository"]
 
 [tool.setuptools.packages.find]
-include = ["vibego_cli", "tasks", "scripts"]
+include = ["vibego_cli", "tasks", "scripts", "telegram_markdown"]
 
 [tool.setuptools.package-data]
 "vibego_cli" = ["data/*.txt"]
 "scripts" = ["*.sh", "models/*.sh"]
 
 [tool.bumpversion]
-current_version = "0.2.33"
+current_version = "0.2.35"
 parse = "(?P<major>\\d+)\\.(?P<minor>\\d+)\\.(?P<patch>\\d+)"
 serialize = ["{major}.{minor}.{patch}"]
 search = "{current_version}"

{vibego-0.2.33 → vibego-0.2.35}/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

vibego-0.2.35/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-0.2.35/tests/test_markdown_guard.py

@@ -0,0 +1,43 @@
+"""验证 Markdown 降级为纯文本时的反转义逻辑。"""
+
+from __future__ import annotations
+
+import pytest
+from aiogram.exceptions import TelegramBadRequest
+from aiogram.methods.base import TelegramMethod
+
+import bot
+
+
+@pytest.mark.asyncio()
+async def test_markdown_guard_plain_fallback_removes_escapes() -> None:
+    """确保回退纯文本时不会残留 MarkdownV2 转义反斜杠。"""
+
+    delivered: list[str] = []
+
+    class _DummyMethod(TelegramMethod):
+        """构造最小化的 TelegramMethod 以便创建异常对象。"""
+
+        __returning__ = bool
+
+        @property
+        def __api_method__(self) -> str:
+            return "testMethod"
+
+        def build_response(self, data, bot):  # pragma: no cover - 测试不会触发
+            return True
+
+        def build_request(self, bot):  # pragma: no cover - 测试不会触发
+            return "testMethod", {}
+
+    async def _failing_sender(_: str) -> None:
+        raise TelegramBadRequest(_DummyMethod(), "Bad Request: can't parse entities")
+
+    async def _raw_sender(payload: str) -> None:
+        delivered.append(payload)
+
+    original = "测试 \\_Markdown\\_ 转义"
+    result = await bot._send_with_markdown_guard(original, _failing_sender, raw_sender=_raw_sender)
+
+    assert delivered and delivered[0] == "测试 _Markdown_ 转义"
+    assert result == "测试 _Markdown_ 转义"

{vibego-0.2.33 → vibego-0.2.35}/tests/test_task_list_entry.py

@@ -419,8 +419,11 @@ def test_task_list_search_flow(monkeypatch):
         user_message = DummyMessage(text="登录")
         await bot.on_task_list_search_keyword(user_message, state)
         assert await state.get_state() is None
-        # 在 MarkdownV2 模式下 * 会被转义
-        assert message.edits and ("*任务搜索结果*" in message.edits[-1]["text"] or "\\*任务搜索结果\\*" in message.edits[-1]["text"])
+        # 在 MarkdownV2 模式下会出现 * 或 _ 的格式化占位
+        assert message.edits
+        header_text = message.edits[-1]["text"]
+        expected_headers = ("*任务搜索结果*", "\\*任务搜索结果\\*", "_任务搜索结果_")
+        assert any(header in header_text for header in expected_headers)
         assert "- 🛠️ 修复登录问题" not in message.edits[-1]["text"]
         assert "- ⚪ 修复登录问题" not in message.edits[-1]["text"]
         assert user_message.calls and "搜索完成" in user_message.calls[-1]["text"]
@@ -457,8 +460,11 @@ def test_task_list_search_cancel_restores_list(monkeypatch):
         cancel_message = DummyMessage(text="取消")
         await bot.on_task_list_search_keyword(cancel_message, state)
         assert await state.get_state() is None
-        # 在 MarkdownV2 模式下 * 会被转义
-        assert message.edits and ("*任务列表*" in message.edits[-1]["text"] or "\\*任务列表\\*" in message.edits[-1]["text"])
+        # 在 MarkdownV2 模式下会出现 * 或 _ 的格式化占位
+        assert message.edits
+        header_text = message.edits[-1]["text"]
+        expected_headers = ("*任务列表*", "\\*任务列表\\*", "_任务列表_")
+        assert any(header in header_text for header in expected_headers)
         assert cancel_message.calls and "已返回任务列表" in cancel_message.calls[-1]["text"]
 
     asyncio.run(_scenario())

vibego-0.2.35/tests/test_telegram_markdown_renderer.py

@@ -0,0 +1,58 @@
+from pathlib import Path
+import sys
+
+sys.path.append(str(Path(__file__).resolve().parents[1]))
+
+import pytest
+
+from telegram_markdown import render_markdown_to_telegram
+
+
+@pytest.mark.parametrize(
+    "source, expected",
+    [
+        ("**bold** text", "*bold* text"),
+        ("__italic__", "*italic*"),
+        ("`code_snippet`", "`code_snippet`"),
+        ("普通文本 (test)", "普通文本 \\(test\\)"),
+        ("raw *literal* star", "raw _literal_ star"),
+        ("带感叹号! 以及加号+", "带感叹号\\! 以及加号\\+"),
+        ("multiline\nline2", "multiline\nline2"),
+    ],
+)
+def test_render_inline_markdown(source: str, expected: str) -> None:
+    assert render_markdown_to_telegram(source) == expected
+
+
+def test_render_code_block_with_language() -> None:
+    markdown = "```python\nprint('hello')\n```"
+    rendered = render_markdown_to_telegram(markdown)
+    assert rendered == "```python\nprint('hello')\n```"
+
+
+def test_render_bullet_list() -> None:
+    markdown = "- item 1\n- item 2"
+    rendered = render_markdown_to_telegram(markdown)
+    assert rendered.splitlines() == ["- item 1", "- item 2"]
+
+
+def test_render_ordered_list() -> None:
+    markdown = "1. first\n2. second"
+    rendered = render_markdown_to_telegram(markdown)
+    assert rendered.splitlines() == ["1. first", "2. second"]
+
+
+def test_render_heading_and_paragraph_spacing() -> None:
+    markdown = "# Title\n\nParagraph line."
+    rendered = render_markdown_to_telegram(markdown)
+    assert rendered == "*Title*\n\nParagraph line\\."
+
+
+def test_render_blockquote() -> None:
+    markdown = "> quote"
+    assert render_markdown_to_telegram(markdown) == "> quote"
+
+
+def test_render_link() -> None:
+    markdown = "[link](https://example.com/path?a=1&b=2)"
+    assert render_markdown_to_telegram(markdown) == "[link](https://example.com/path?a=1&b=2)"

{vibego-0.2.33 → vibego-0.2.35}/vibego.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vibego
-Version: 0.2.33
+Version: 0.2.35
 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-0.2.33 → vibego-0.2.35}/vibego.egg-info/SOURCES.txt

@@ -430,10 +430,12 @@ tasks/constants.py
 tasks/fsm.py
 tasks/models.py
 tasks/service.py
+telegram_markdown/__init__.py
 tests/test_auto_compact.py
 tests/test_chat_menu_buttons.py
 tests/test_claudecode_jsonl.py
 tests/test_long_poll_mechanism.py
+tests/test_markdown_guard.py
 tests/test_master_project_management.py
 tests/test_master_restart.py
 tests/test_numbered_reply_inputs.py
@@ -442,6 +444,7 @@ tests/test_task_create_flow.py
 tests/test_task_description.py
 tests/test_task_detail_back.py
 tests/test_task_list_entry.py
+tests/test_telegram_markdown_renderer.py
 tests/test_unescape_markdown.py
 tests/test_version_display.py
 vibego.egg-info/PKG-INFO

{vibego-0.2.33 → vibego-0.2.35}/vibego.egg-info/requires.txt

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

{vibego-0.2.33 → vibego-0.2.35}/vibego.egg-info/top_level.txt

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

{vibego-0.2.33 → vibego-0.2.35}/vibego_cli/__init__.py

@@ -7,6 +7,6 @@ from __future__ import annotations
 
 __all__ = ["main", "__version__"]
 
-__version__ = "0.2.33"
+__version__ = "0.2.35"
 
 from .main import main  # noqa: E402