UnisKB/apps/system_manage/test_document_export.py

90 lines
4.5 KiB
Python

"""No database or external model is needed for these export contract checks."""
import base64
import io
from pathlib import Path
import tempfile
import unittest
from unittest.mock import patch
import zipfile
from docx import Document
from PIL import Image
from system_manage.document_export import export_document, ExportError
class DocumentExportTests(unittest.TestCase):
def test_real_docx_preserves_structure_and_merged_cells(self):
image = io.BytesIO()
Image.new('RGB', (40, 20), 'blue').save(image, format='PNG')
html = '<h1>中文报告</h1><p><strong>加粗</strong><em>斜体</em><sup>2</sup></p>'
html += '<ol start="3"><li>第三项<ul><li>子项</li></ul></li></ol>'
html += '<p><a href="https://example.com">可编辑链接</a><img alt="测试图" src="data:image/png;base64,' + base64.b64encode(image.getvalue()).decode() + '"></p>'
html += '<table><tr><th colspan="2">表头</th></tr><tr><td rowspan="2">合并</td><td>A</td></tr><tr><td>B</td></tr></table><p>结束标记</p>'
raw, mime = export_document(html)
self.assertTrue(raw.startswith(b'PK'))
self.assertIn('wordprocessingml', mime)
doc = Document(io.BytesIO(raw))
self.assertEqual(len(doc.inline_shapes), 1)
self.assertEqual(len(doc.tables[0].rows), 3)
self.assertEqual(doc.tables[0].cell(1, 0).text, doc.tables[0].cell(2, 0).text)
with zipfile.ZipFile(io.BytesIO(raw)) as archive:
xml = archive.read('word/document.xml').decode()
for expected in ('中文报告', '结束标记', 'w:numPr', 'w:hyperlink', 'w:tblHeader', 'w:vMerge', 'w:gridSpan', 'w:vertAlign'):
self.assertIn(expected, xml)
def test_image_resolution_never_fetches_external_urls(self):
for src in ('https://example.com/private', 'file:///etc/passwd', 'data:image/png;base64,AAAA'):
with self.assertRaises(ExportError):
export_document('<p><img src="' + src + '"></p>')
def test_unsupported_format_and_paper_are_rejected(self):
for kwargs in ({'format_name': 'html'}, {'paper': 'unknown'}, {'paper': 'Letter'}):
with self.assertRaises(ExportError): export_document('<p>正文</p>', **kwargs)
def test_long_table_grid_is_not_rebuilt_per_cell(self):
from docx.table import Table
read_grid = Table._cells.fget
calls = []
def counted(table):
calls.append(1)
return read_grid(table)
html = '<table><tr><th>编号</th><th>内容</th></tr>' + ''.join(
f'<tr><td>{i}</td><td>完整内容 {i}</td></tr>' for i in range(400)) + '</table>'
with patch.object(Table, '_cells', property(counted)):
raw, _ = export_document(html)
self.assertEqual(len(calls), 1)
doc = Document(io.BytesIO(raw))
self.assertEqual(len(doc.tables[0].rows), 401)
self.assertEqual(doc.tables[0].cell(400, 1).text, '完整内容 399')
section = doc.sections[0]
self.assertAlmostEqual(section.page_width.mm, 210, places=1)
self.assertAlmostEqual(section.page_height.mm, 297, places=1)
def test_bad_table_cannot_overflow_column_limit(self):
with self.assertRaises(ExportError):
export_document('<table><tr>' + '<td>x</td>' * 31 + '</tr></table>')
def test_conversion_failure_is_explicit(self):
with patch('system_manage.document_export.shutil.which', return_value=None):
with self.assertRaisesRegex(ExportError, '转换组件不可用'):
export_document('<p>正文</p>', 'pdf')
def test_conversion_uses_private_profile_and_removes_temporary_files(self):
directories = []
def convert(args, **kwargs):
root = Path(args[args.index('--outdir') + 1]); directories.append(root)
self.assertTrue(any(arg.startswith('-env:UserInstallation=file:') for arg in args))
self.assertEqual(kwargs['timeout'], 90)
self.assertNotIn('shell', kwargs)
(root / 'document.pdf').write_bytes(b'%PDF-1.7\ncontract-test')
return type('Result', (), {'returncode': 0})()
with patch('system_manage.document_export.shutil.which', return_value='soffice'), patch('system_manage.document_export.subprocess.run', side_effect=convert):
raw, mime = export_document('<p>正文</p>', 'pdf')
self.assertEqual(mime, 'application/pdf')
self.assertTrue(raw.startswith(b'%PDF'))
self.assertTrue(all(not path.exists() for path in directories))
if __name__ == '__main__':
unittest.main()