from __future__ import annotations
|
|
import json
|
import sys
|
import tempfile
|
import unittest
|
from pathlib import Path
|
|
|
MODULE_DIR = Path(__file__).resolve().parents[1] / "meeting_minutes_tool"
|
sys.path.insert(0, str(MODULE_DIR))
|
|
import meeting_minutes_tool as tool # noqa: E402
|
|
|
class TranscriptIndexTests(unittest.TestCase):
|
def test_index_extracts_navigation_and_flags_regression(self) -> None:
|
with tempfile.TemporaryDirectory() as temporary:
|
root = Path(temporary)
|
transcript = root / "sample.txt"
|
transcript.write_text(
|
"\n".join(
|
[
|
"甲 共享音频(00:00:10): 正式开始，政治局政策与算力网投资约21.4%。",
|
"甲 共享音频(00:00:20): 请问华为和半导体公司怎么看？",
|
"甲 共享音频(00:00:05): 这个数字大概需要核对，不构成投资建议。",
|
"甲 共享音频(00:00:30): 一段足够长并且会重复出现的测试内容用于重复检测。",
|
"甲 共享音频(00:00:40): 一段足够长并且会重复出现的测试内容用于重复检测。",
|
]
|
),
|
encoding="utf-8",
|
)
|
result = tool.build_transcript_index(transcript, root / "index", max_items=20)
|
|
self.assertEqual(result["source"]["line_count"], 5)
|
self.assertEqual(result["sequence_check"]["timestamp_regression_count"], 1)
|
self.assertEqual(
|
result["sequence_check"]["exact_normalized_duplicate_group_count"], 1
|
)
|
self.assertGreater(
|
result["category_total_counts"]["numbers_and_time_windows"], 0
|
)
|
self.assertTrue((root / "index" / "transcript-index.json").is_file())
|
self.assertTrue((root / "index" / "transcript-index.md").is_file())
|
|
def test_missing_transcript_fails_fast(self) -> None:
|
with tempfile.TemporaryDirectory() as temporary:
|
root = Path(temporary)
|
with self.assertRaises(FileNotFoundError):
|
tool.build_transcript_index(root / "missing.txt", root / "index")
|
|
|
class PdfPipelineTests(unittest.TestCase):
|
SAMPLE_MARKDOWN = """# 2026年度超长标题会议纪要【第6场：政策与产业链深度解读和后续跟踪安排】
|
|
> **免责声明：内容为会议观点记录，不构成投资建议。**
|
|
## 一、核心结论
|
|
1. 这是第一条核心结论。
|
2. 这是第二条核心结论。
|
|
## 二、主体内容
|
|
### 2.1 政策与产业
|
|
- 算力网与新型电网需要结合订单验证。
|
- 无法确认的数据需要标记为待核对。
|
|
| 行业 | 观点 | 风险 |
|
|---|---|---|
|
| 算力 | 关注订单和交付 | 估值与技术迭代 |
|
| 电网 | 关注招标与资本开支 | 项目节奏 |
|
|
## 三、问答环节重点
|
|
### 3.1 如何判断？
|
|
- 回答：结合基本面、价格和风险承受能力判断。
|
|
## 四、后续跟踪清单
|
|
1. 跟踪订单。
|
2. 跟踪正式政策。
|
"""
|
|
def test_pdf_and_qa_roundtrip(self) -> None:
|
with tempfile.TemporaryDirectory() as temporary:
|
root = Path(temporary)
|
markdown = root / "minutes.md"
|
pdf = root / "minutes.pdf"
|
qa_dir = root / "qa"
|
markdown.write_text(self.SAMPLE_MARKDOWN, encoding="utf-8")
|
|
pdf_result = tool.render_markdown_pdf(markdown, pdf)
|
qa_result = tool.run_pdf_qa(markdown, pdf, qa_dir, dpi=96)
|
|
self.assertGreater(pdf_result["pdf_bytes"], 10_000)
|
self.assertTrue(pdf.is_file())
|
self.assertTrue(qa_result["automated_checks_passed"])
|
self.assertTrue(qa_result["manual_visual_review_required"])
|
self.assertEqual(
|
qa_result["pdf"]["page_count"],
|
len(qa_result["render"]["page_images"]),
|
)
|
self.assertTrue((qa_dir / "qa-report.json").is_file())
|
self.assertTrue((qa_dir / "qa-report.md").is_file())
|
payload = json.loads((qa_dir / "qa-report.json").read_text(encoding="utf-8"))
|
self.assertEqual(payload["schema_version"], "meeting-minutes-pdf-qa.v1")
|
|
def test_invalid_markdown_does_not_leave_pdf(self) -> None:
|
with tempfile.TemporaryDirectory() as temporary:
|
root = Path(temporary)
|
markdown = root / "invalid.md"
|
pdf = root / "invalid.pdf"
|
markdown.write_text("没有一级标题。", encoding="utf-8")
|
with self.assertRaises(ValueError):
|
tool.render_markdown_pdf(markdown, pdf)
|
self.assertFalse(pdf.exists())
|
|
|
if __name__ == "__main__":
|
unittest.main()
|
