fix: route txt/md KB uploads to TextParser and clarify missing markitdown

astrbot/core/knowledge_base/parsers/util.py

@@ -1,10 +1,32 @@
+from astrbot.core.exceptions import KnowledgeBaseUploadError
+
 from .base import BaseParser
 
+# Formats that are plain text by nature and need no external dependency.
+_TEXT_PARSER_EXTS = {".md", ".txt", ".markdown"}
+# Formats handled by markitdown-no-magika (an optional heavy dependency).
+_MARKITDOWN_EXTS = {".rst", ".adoc", ".xlsx", ".docx", ".xls"}
+
+_MARKITDOWN_MISSING_HINT = (
+    "文档解析失败：处理该格式需要 markitdown-no-magika 依赖，"
+    "请先安装：pip install 'markitdown-no-magika[docx,xls,xlsx]'"
+)
+
 
 async def select_parser(ext: str) -> BaseParser:
-    if ext in {".md", ".txt", ".markdown", ".rst", ".adoc", ".xlsx", ".docx", ".xls"}:
-        from .markitdown_parser import MarkitdownParser
+    if ext in _TEXT_PARSER_EXTS:
+        from .text_parser import TextParser
 
+        return TextParser()
+    if ext in _MARKITDOWN_EXTS:
+        try:
+            from .markitdown_parser import MarkitdownParser
+        except ImportError as exc:
+            raise KnowledgeBaseUploadError(
+                stage="parsing",
+                user_message=_MARKITDOWN_MISSING_HINT,
+                details={"file_ext": ext},
+            ) from exc
         return MarkitdownParser()
     if ext == ".epub":
         from .epub_parser import EpubParser

tests/test_kb_select_parser.py

@@ -0,0 +1,58 @@
+from __future__ import annotations
+
+import sys
+
+import pytest
+
+from astrbot.core.exceptions import KnowledgeBaseUploadError
+from astrbot.core.knowledge_base.parsers.markitdown_parser import MarkitdownParser
+from astrbot.core.knowledge_base.parsers.text_parser import TextParser
+from astrbot.core.knowledge_base.parsers.util import select_parser
+
+
+@pytest.mark.parametrize("ext", [".txt", ".md", ".markdown"])
+@pytest.mark.asyncio
+@pytest.mark.asyncio
+async def test_text_formats_use_text_parser(ext: str) -> None:
+    parser = await select_parser(ext)
+
+    assert isinstance(parser, TextParser)
+
+
+@pytest.mark.parametrize("ext", [".rst", ".adoc", ".xlsx", ".docx", ".xls"])
+@pytest.mark.asyncio
+@pytest.mark.asyncio
+async def test_office_formats_use_markitdown_parser(ext: str) -> None:
+    parser = await select_parser(ext)
+
+    assert isinstance(parser, MarkitdownParser)
+
+
+@pytest.mark.asyncio
+async def test_epub_and_pdf_routing() -> None:
+    from astrbot.core.knowledge_base.parsers.epub_parser import EpubParser
+    from astrbot.core.knowledge_base.parsers.pdf_parser import PDFParser
+
+    assert isinstance(await select_parser(".epub"), EpubParser)
+    assert isinstance(await select_parser(".pdf"), PDFParser)
+
+
+@pytest.mark.asyncio
+async def test_unsupported_ext_raises() -> None:
+    with pytest.raises(ValueError, match="暂时不支持的文件格式"):
+        await select_parser(".exe")
+
+
+@pytest.mark.asyncio
+async def test_missing_markitdown_dependency_raises_clear_error(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    # Drop cached modules so the lazy import runs again, then make
+    # `import markitdown_no_magika` fail as if the package were absent.
+    for name in list(sys.modules):
+        if name == "markitdown_no_magika" or name.endswith("markitdown_parser"):
+            monkeypatch.delitem(sys.modules, name)
+    monkeypatch.setitem(sys.modules, "markitdown_no_magika", None)
+
+    with pytest.raises(KnowledgeBaseUploadError, match="markitdown-no-magika"):
+        await select_parser(".docx")