From d91ac0c2a204fa79b4796cf079929fbbc6d214c8 Mon Sep 17 00:00:00 2001 From: Leonard Richardson Date: Thu, 1 Mar 2012 08:29:35 -0500 Subject: Initial port of code and tests. --- bs4/tests/test_tree.py | 228 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 228 insertions(+) (limited to 'bs4/tests/test_tree.py') diff --git a/bs4/tests/test_tree.py b/bs4/tests/test_tree.py index 4dda90e..39f78f7 100644 --- a/bs4/tests/test_tree.py +++ b/bs4/tests/test_tree.py @@ -1300,3 +1300,231 @@ class TestNavigableStringSubclasses(SoupTest): soup.insert(1, doctype) self.assertEqual(soup.encode(), b"\n") + +class TestSoupSelector(SoupTest): + + HTML = """ + + + The title + + + + +
+
+

An H1

+

Some text

+

Some more text

+

An H2

+

Another

+ Bob +

Another H2

+ me +
+

English

+

English UK

+

English US

+

French

+
+ + + + + +""" + + def setUp(self): + self.soup = BeautifulSoup(self.HTML) + + def assertSelects(self, selector, expected_ids): + el_ids = [el['id'] for el in self.soup.select(selector)] + el_ids.sort() + expected_ids.sort() + self.assertEqual(expected_ids, el_ids, + "Selector %s, expected [%s], got [%s]" % ( + selector, ', '.join(expected_ids), ', '.join(el_ids) + ) + ) + + assertSelect = assertSelects + + def assertSelectMultiple(self, *tests): + for selector, expected_ids in tests: + self.assertSelect(selector, expected_ids) + + def test_one_tag_one(self): + els = self.soup.select('title') + self.assertEqual(len(els), 1) + self.assertEqual(els[0].name, 'title') + self.assertEqual(els[0].contents, [u'The title']) + + def test_one_tag_many(self): + els = self.soup.select('div') + self.assertEqual(len(els), 3) + for div in els: + self.assertEqual(div.name, 'div') + + def test_tag_in_tag_one(self): + els = self.soup.select('div div') + self.assertSelects('div div', ['inner']) + + def test_tag_in_tag_many(self): + for selector in ('html div', 'html body div', 'body div'): + self.assertSelects(selector, ['main', 'inner', 'footer']) + + def test_tag_no_match(self): + self.assertEqual(len(self.soup.select('del')), 0) + + def test_invalid_tag(self): + self.assertEqual(len(self.soup.select('tag%t')), 0) + + def test_header_tags(self): + self.assertSelectMultiple( + ('h1', ['header1']), + ('h2', ['header2', 'header3']), + ) + + def test_class_one(self): + for selector in ('.onep', 'p.onep', 'html p.onep'): + els = self.soup.select(selector) + self.assertEqual(len(els), 1) + self.assertEqual(els[0].name, 'p') + self.assertEqual(els[0]['class'], ['onep']) + + def test_class_mismatched_tag(self): + els = self.soup.select('div.onep') + self.assertEqual(len(els), 0) + + def test_one_id(self): + for selector in ('div#inner', '#inner', 'div div#inner'): + self.assertSelects(selector, ['inner']) + + def test_bad_id(self): + els = self.soup.select('#doesnotexist') + self.assertEqual(len(els), 0) + + def test_items_in_id(self): + els = self.soup.select('div#inner p') + self.assertEqual(len(els), 3) + for el in els: + self.assertEqual(el.name, 'p') + self.assertEqual(els[1]['class'], ['onep']) + self.assert_(not els[0].has_key('class')) + + def test_a_bunch_of_emptys(self): + for selector in ('div#main del', 'div#main div.oops', 'div div#main'): + self.assertEqual(len(self.soup.select(selector)), 0) + + def test_multi_class_support(self): + for selector in ('.class1', 'p.class1', '.class2', 'p.class2', + '.class3', 'p.class3', 'html p.class2', 'div#inner .class2'): + self.assertSelects(selector, ['pmulti']) + + def test_attribute_equals(self): + self.assertSelectMultiple( + ('p[class="onep"]', ['p1']), + ('p[id="p1"]', ['p1']), + ('[class="onep"]', ['p1']), + ('[id="p1"]', ['p1']), + ('link[rel="stylesheet"]', ['l1']), + ('link[type="text/css"]', ['l1']), + ('link[href="blah.css"]', ['l1']), + ('link[href="no-blah.css"]', []), + ('[rel="stylesheet"]', ['l1']), + ('[type="text/css"]', ['l1']), + ('[href="blah.css"]', ['l1']), + ('[href="no-blah.css"]', []), + ('p[href="no-blah.css"]', []), + ('[href="no-blah.css"]', []), + ) + + def test_attribute_tilde(self): + self.assertSelectMultiple( + ('p[class~="class1"]', ['pmulti']), + ('p[class~="class2"]', ['pmulti']), + ('p[class~="class3"]', ['pmulti']), + ('[class~="class1"]', ['pmulti']), + ('[class~="class2"]', ['pmulti']), + ('[class~="class3"]', ['pmulti']), + ('a[rel~="friend"]', ['bob']), + ('a[rel~="met"]', ['bob']), + ('[rel~="friend"]', ['bob']), + ('[rel~="met"]', ['bob']), + ) + + def test_attribute_startswith(self): + self.assertSelectMultiple( + ('[rel^="style"]', ['l1']), + ('link[rel^="style"]', ['l1']), + ('notlink[rel^="notstyle"]', []), + ('[rel^="notstyle"]', []), + ('link[rel^="notstyle"]', []), + ('link[href^="bla"]', ['l1']), + ('a[href^="http://"]', ['bob', 'me']), + ('[href^="http://"]', ['bob', 'me']), + ('[id^="p"]', ['pmulti', 'p1']), + ('[id^="m"]', ['me', 'main']), + ('div[id^="m"]', ['main']), + ('a[id^="m"]', ['me']), + ) + + def test_attribute_endswith(self): + self.assertSelectMultiple( + ('[href$=".css"]', ['l1']), + ('link[href$=".css"]', ['l1']), + ('link[id$="1"]', ['l1']), + ('[id$="1"]', ['l1', 'p1', 'header1']), + ('div[id$="1"]', []), + ('[id$="noending"]', []), + ) + + def test_attribute_contains(self): + self.assertSelectMultiple( + # From test_attribute_startswith + ('[rel*="style"]', ['l1']), + ('link[rel*="style"]', ['l1']), + ('notlink[rel*="notstyle"]', []), + ('[rel*="notstyle"]', []), + ('link[rel*="notstyle"]', []), + ('link[href*="bla"]', ['l1']), + ('a[href*="http://"]', ['bob', 'me']), + ('[href*="http://"]', ['bob', 'me']), + ('[id*="p"]', ['pmulti', 'p1']), + ('div[id*="m"]', ['main']), + ('a[id*="m"]', ['me']), + # From test_attribute_endswith + ('[href*=".css"]', ['l1']), + ('link[href*=".css"]', ['l1']), + ('link[id*="1"]', ['l1']), + ('[id*="1"]', ['l1', 'p1', 'header1']), + ('div[id*="1"]', []), + ('[id*="noending"]', []), + # New for this test + ('[href*="."]', ['bob', 'me', 'l1']), + ('a[href*="."]', ['bob', 'me']), + ('link[href*="."]', ['l1']), + ('div[id*="n"]', ['main', 'inner']), + ('div[id*="nn"]', ['inner']), + ) + + def test_attribute_exact_or_hypen(self): + self.assertSelectMultiple( + ('p[lang|="en"]', ['lang-en', 'lang-en-gb', 'lang-en-us']), + ('[lang|="en"]', ['lang-en', 'lang-en-gb', 'lang-en-us']), + ('p[lang|="fr"]', ['lang-fr']), + ('p[lang|="gb"]', []), + ) + + def test_attribute_exists(self): + self.assertSelectMultiple( + ('[rel]', ['l1', 'bob', 'me']), + ('link[rel]', ['l1']), + ('a[rel]', ['bob', 'me']), + ('[lang]', ['lang-en', 'lang-en-gb', 'lang-en-us', 'lang-fr']), + ('p[class]', ['p1', 'pmulti']), + ('[blah]', []), + ('p[blah]', []), + ) -- cgit v1.2.3 From 8bbc84dfc6324a32066b87cf3a78ce0eb719e289 Mon Sep 17 00:00:00 2001 From: Leonard Richardson Date: Thu, 1 Mar 2012 08:49:27 -0500 Subject: Got tests to pass on Python 2 and Python 3. --- bs4/tests/test_tree.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'bs4/tests/test_tree.py') diff --git a/bs4/tests/test_tree.py b/bs4/tests/test_tree.py index 39f78f7..bfc4218 100644 --- a/bs4/tests/test_tree.py +++ b/bs4/tests/test_tree.py @@ -1412,7 +1412,7 @@ class TestSoupSelector(SoupTest): for el in els: self.assertEqual(el.name, 'p') self.assertEqual(els[1]['class'], ['onep']) - self.assert_(not els[0].has_key('class')) + self.assertFalse(els[0].has_key('class')) def test_a_bunch_of_emptys(self): for selector in ('div#main del', 'div#main div.oops', 'div div#main'): -- cgit v1.2.3 From 4a5136d31bf07a7b28b58343f0c32e41d895e110 Mon Sep 17 00:00:00 2001 From: Leonard Richardson Date: Thu, 1 Mar 2012 09:06:25 -0500 Subject: Test that CSS selectors work within the tree as well as at the top level. --- bs4/tests/test_tree.py | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'bs4/tests/test_tree.py') diff --git a/bs4/tests/test_tree.py b/bs4/tests/test_tree.py index bfc4218..3f32736 100644 --- a/bs4/tests/test_tree.py +++ b/bs4/tests/test_tree.py @@ -1301,7 +1301,7 @@ class TestNavigableStringSubclasses(SoupTest): self.assertEqual(soup.encode(), b"\n") -class TestSoupSelector(SoupTest): +class TestSoupSelector(TreeTest): HTML = """ @@ -1528,3 +1528,12 @@ class TestSoupSelector(SoupTest): ('[blah]', []), ('p[blah]', []), ) + + def test_select_on_element(self): + # Other tests operate on the tree; this operates on an element + # within the tree. + inner = self.soup.find("div", id="main") + selected = inner.select("div") + # The
tag was selected. The