123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229 |
- # -*- coding: utf-8 -*-
- import lxml.etree
- from lxml import html
- from searx.testing import SearxTestCase
- from searx.exceptions import SearxXPathSyntaxException, SearxEngineXPathException
- from searx import utils
- class TestUtils(SearxTestCase):
- def test_gen_useragent(self):
- self.assertIsInstance(utils.gen_useragent(), str)
- self.assertIsNotNone(utils.gen_useragent())
- self.assertTrue(utils.gen_useragent().startswith('Mozilla'))
- def test_searx_useragent(self):
- self.assertIsInstance(utils.searx_useragent(), str)
- self.assertIsNotNone(utils.searx_useragent())
- self.assertTrue(utils.searx_useragent().startswith('searx'))
- def test_html_to_text(self):
- html_str = """
- <a href="/testlink" class="link_access_account">
- <style>
- .toto {
- color: red;
- }
- </style>
- <span class="toto">
- <span>
- <img src="test.jpg" />
- </span>
- </span>
- <span class="titi">
- Test text
- </span>
- <script>value='dummy';</script>
- </a>
- """
- self.assertIsInstance(utils.html_to_text(html_str), str)
- self.assertIsNotNone(utils.html_to_text(html_str))
- self.assertEqual(utils.html_to_text(html_str), "Test text")
- def test_extract_text(self):
- html_str = """
- <a href="/testlink" class="link_access_account">
- <span class="toto">
- <span>
- <img src="test.jpg" />
- </span>
- </span>
- <span class="titi">
- Test text
- </span>
- </a>
- """
- dom = html.fromstring(html_str)
- self.assertEqual(utils.extract_text(dom), 'Test text')
- self.assertEqual(utils.extract_text(dom.xpath('//span')), 'Test text')
- self.assertEqual(utils.extract_text(dom.xpath('//span/text()')), 'Test text')
- self.assertEqual(utils.extract_text(dom.xpath('count(//span)')), '3.0')
- self.assertEqual(utils.extract_text(dom.xpath('boolean(//span)')), 'True')
- self.assertEqual(utils.extract_text(dom.xpath('//img/@src')), 'test.jpg')
- self.assertEqual(utils.extract_text(dom.xpath('//unexistingtag')), '')
- self.assertEqual(utils.extract_text(None, allow_none=True), None)
- with self.assertRaises(ValueError):
- utils.extract_text(None)
- with self.assertRaises(ValueError):
- utils.extract_text({})
- def test_extract_url(self):
- def f(html_str, search_url):
- return utils.extract_url(html.fromstring(html_str), search_url)
- self.assertEqual(f('<span id="42">https://example.com</span>', 'http://example.com/'), 'https://example.com/')
- self.assertEqual(f('https://example.com', 'http://example.com/'), 'https://example.com/')
- self.assertEqual(f('//example.com', 'http://example.com/'), 'http://example.com/')
- self.assertEqual(f('//example.com', 'https://example.com/'), 'https://example.com/')
- self.assertEqual(f('/path?a=1', 'https://example.com'), 'https://example.com/path?a=1')
- with self.assertRaises(lxml.etree.ParserError):
- f('', 'https://example.com')
- with self.assertRaises(Exception):
- utils.extract_url([], 'https://example.com')
- def test_html_to_text_invalid(self):
- html = '<p><b>Lorem ipsum</i>dolor sit amet</p>'
- self.assertEqual(utils.html_to_text(html), "Lorem ipsum")
- def test_match_language(self):
- self.assertEqual(utils.match_language('es', ['es']), 'es')
- self.assertEqual(utils.match_language('es', [], fallback='fallback'), 'fallback')
- self.assertEqual(utils.match_language('ja', ['jp'], {'ja': 'jp'}), 'jp')
- aliases = {'en-GB': 'en-UK', 'he': 'iw'}
- # guess country
- self.assertEqual(utils.match_language('de-DE', ['de']), 'de')
- self.assertEqual(utils.match_language('de', ['de-DE']), 'de-DE')
- self.assertEqual(utils.match_language('es-CO', ['es-AR', 'es-ES', 'es-MX']), 'es-ES')
- self.assertEqual(utils.match_language('es-CO', ['es-MX']), 'es-MX')
- self.assertEqual(utils.match_language('en-UK', ['en-AU', 'en-GB', 'en-US']), 'en-GB')
- self.assertEqual(utils.match_language('en-GB', ['en-AU', 'en-UK', 'en-US'], aliases), 'en-UK')
- # language aliases
- self.assertEqual(utils.match_language('iw', ['he']), 'he')
- self.assertEqual(utils.match_language('he', ['iw'], aliases), 'iw')
- self.assertEqual(utils.match_language('iw-IL', ['he']), 'he')
- self.assertEqual(utils.match_language('he-IL', ['iw'], aliases), 'iw')
- self.assertEqual(utils.match_language('iw', ['he-IL']), 'he-IL')
- self.assertEqual(utils.match_language('he', ['iw-IL'], aliases), 'iw-IL')
- self.assertEqual(utils.match_language('iw-IL', ['he-IL']), 'he-IL')
- self.assertEqual(utils.match_language('he-IL', ['iw-IL'], aliases), 'iw-IL')
- def test_ecma_unscape(self):
- self.assertEqual(utils.ecma_unescape('text%20with%20space'), 'text with space')
- self.assertEqual(utils.ecma_unescape('text using %xx: %F3'),
- 'text using %xx: ó')
- self.assertEqual(utils.ecma_unescape('text using %u: %u5409, %u4E16%u754c'),
- 'text using %u: 吉, 世界')
- class TestHTMLTextExtractor(SearxTestCase):
- def setUp(self):
- self.html_text_extractor = utils.HTMLTextExtractor()
- def test__init__(self):
- self.assertEqual(self.html_text_extractor.result, [])
- def test_handle_charref(self):
- self.html_text_extractor.handle_charref('xF')
- self.assertIn('\x0f', self.html_text_extractor.result)
- self.html_text_extractor.handle_charref('XF')
- self.assertIn('\x0f', self.html_text_extractor.result)
- self.html_text_extractor.handle_charref('97')
- self.assertIn('a', self.html_text_extractor.result)
- def test_handle_entityref(self):
- entity = 'test'
- self.html_text_extractor.handle_entityref(entity)
- self.assertIn(entity, self.html_text_extractor.result)
- def test_invalid_html(self):
- text = '<p><b>Lorem ipsum</i>dolor sit amet</p>'
- with self.assertRaises(utils.HTMLTextExtractorException):
- self.html_text_extractor.feed(text)
- class TestXPathUtils(SearxTestCase):
- TEST_DOC = """<ul>
- <li>Text in <b>bold</b> and <i>italic</i> </li>
- <li>Another <b>text</b> <img src="data:image/gif;base64,R0lGODlhAQABAIAAAAUEBAAAACwAAAAAAQABAAACAkQBADs="></li>
- </ul>"""
- def test_get_xpath_cache(self):
- xp1 = utils.get_xpath('//a')
- xp2 = utils.get_xpath('//div')
- xp3 = utils.get_xpath('//a')
- self.assertEqual(id(xp1), id(xp3))
- self.assertNotEqual(id(xp1), id(xp2))
- def test_get_xpath_type(self):
- utils.get_xpath(lxml.etree.XPath('//a'))
- with self.assertRaises(TypeError):
- utils.get_xpath([])
- def test_get_xpath_invalid(self):
- invalid_xpath = '//a[0].text'
- with self.assertRaises(SearxXPathSyntaxException) as context:
- utils.get_xpath(invalid_xpath)
- self.assertEqual(context.exception.message, 'Invalid expression')
- self.assertEqual(context.exception.xpath_str, invalid_xpath)
- def test_eval_xpath_unregistered_function(self):
- doc = html.fromstring(TestXPathUtils.TEST_DOC)
- invalid_function_xpath = 'int(//a)'
- with self.assertRaises(SearxEngineXPathException) as context:
- utils.eval_xpath(doc, invalid_function_xpath)
- self.assertEqual(context.exception.message, 'Unregistered function')
- self.assertEqual(context.exception.xpath_str, invalid_function_xpath)
- def test_eval_xpath(self):
- doc = html.fromstring(TestXPathUtils.TEST_DOC)
- self.assertEqual(utils.eval_xpath(doc, '//p'), [])
- self.assertEqual(utils.eval_xpath(doc, '//i/text()'), ['italic'])
- self.assertEqual(utils.eval_xpath(doc, 'count(//i)'), 1.0)
- def test_eval_xpath_list(self):
- doc = html.fromstring(TestXPathUtils.TEST_DOC)
- # check a not empty list
- self.assertEqual(utils.eval_xpath_list(doc, '//i/text()'), ['italic'])
- # check min_len parameter
- with self.assertRaises(SearxEngineXPathException) as context:
- utils.eval_xpath_list(doc, '//p', min_len=1)
- self.assertEqual(context.exception.message, 'len(xpath_str) < 1')
- self.assertEqual(context.exception.xpath_str, '//p')
- def test_eval_xpath_getindex(self):
- doc = html.fromstring(TestXPathUtils.TEST_DOC)
- # check index 0
- self.assertEqual(utils.eval_xpath_getindex(doc, '//i/text()', 0), 'italic')
- # default is 'something'
- self.assertEqual(utils.eval_xpath_getindex(doc, '//i/text()', 1, default='something'), 'something')
- # default is None
- self.assertEqual(utils.eval_xpath_getindex(doc, '//i/text()', 1, default=None), None)
- # index not found
- with self.assertRaises(SearxEngineXPathException) as context:
- utils.eval_xpath_getindex(doc, '//i/text()', 1)
- self.assertEqual(context.exception.message, 'index 1 not found')
- # not a list
- with self.assertRaises(SearxEngineXPathException) as context:
- utils.eval_xpath_getindex(doc, 'count(//i)', 1)
- self.assertEqual(context.exception.message, 'the result is not a list')
|