"""测试 MarkItDown DOC Reader 的解析功能。"""

import pytest
import os
from readers.doc import markitdown


class TestMarkitdownDocReaderParse:
    """测试 MarkItDown DOC Reader 的 parse 方法。"""

    def test_module_importable(self):
        """测试模块可以正确导入。"""
        assert markitdown is not None
        assert hasattr(markitdown, 'parse')
        assert callable(markitdown.parse)

    def test_file_not_exists(self, tmp_path):
        """测试文件不存在的情况。"""
        non_existent_file = str(tmp_path / "non_existent.doc")

        content, error = markitdown.parse(non_existent_file)

        # 验证返回 None 和错误信息
        assert content is None
        assert error is not None

    def test_parse_simple_doc(self, simple_doc_path):
        """测试解析简单 DOC 文件。"""
        content, error = markitdown.parse(simple_doc_path)

        # 只要不崩溃即可，不强制要求成功解析
        if content is not None:
            assert len(content.strip()) > 0

    def test_parse_with_headings_doc(self, with_headings_doc_path):
        """测试解析带标题的 DOC 文件。"""
        content, error = markitdown.parse(with_headings_doc_path)

        # 只要不崩溃即可
        if content is not None:
            assert len(content.strip()) > 0

    def test_parse_with_table_doc(self, with_table_doc_path):
        """测试解析带表格的 DOC 文件。"""
        content, error = markitdown.parse(with_table_doc_path)

        # 只要不崩溃即可
        if content is not None:
            assert len(content.strip()) > 0