| A"""
+ soup = self.soup(markup)
+ self.assertEqual(u"A", soup.body.decode())
diff --git a/src/bs4/tests/test_htmlparser.py b/src/bs4/tests/test_htmlparser.py
new file mode 100644
index 0000000..b45e35f
--- /dev/null
+++ b/src/bs4/tests/test_htmlparser.py
@@ -0,0 +1,32 @@
+"""Tests to ensure that the html.parser tree builder generates good
+trees."""
+
+from pdb import set_trace
+import pickle
+from bs4.testing import SoupTest, HTMLTreeBuilderSmokeTest
+from bs4.builder import HTMLParserTreeBuilder
+
+class HTMLParserTreeBuilderSmokeTest(SoupTest, HTMLTreeBuilderSmokeTest):
+
+ @property
+ def default_builder(self):
+ return HTMLParserTreeBuilder()
+
+ def test_namespaced_system_doctype(self):
+ # html.parser can't handle namespaced doctypes, so skip this one.
+ pass
+
+ def test_namespaced_public_doctype(self):
+ # html.parser can't handle namespaced doctypes, so skip this one.
+ pass
+
+ def test_builder_is_pickled(self):
+ """Unlike most tree builders, HTMLParserTreeBuilder and will
+ be restored after pickling.
+ """
+ tree = self.soup("foo")
+ dumped = pickle.dumps(tree, 2)
+ loaded = pickle.loads(dumped)
+ self.assertTrue(isinstance(loaded.builder, type(tree.builder)))
+
+
diff --git a/src/bs4/tests/test_lxml.py b/src/bs4/tests/test_lxml.py
new file mode 100644
index 0000000..a05870b
--- /dev/null
+++ b/src/bs4/tests/test_lxml.py
@@ -0,0 +1,76 @@
+"""Tests to ensure that the lxml tree builder generates good trees."""
+
+import re
+import warnings
+
+try:
+ import lxml.etree
+ LXML_PRESENT = True
+ LXML_VERSION = lxml.etree.LXML_VERSION
+except ImportError, e:
+ LXML_PRESENT = False
+ LXML_VERSION = (0,)
+
+if LXML_PRESENT:
+ from bs4.builder import LXMLTreeBuilder, LXMLTreeBuilderForXML
+
+from bs4 import (
+ BeautifulSoup,
+ BeautifulStoneSoup,
+ )
+from bs4.element import Comment, Doctype, SoupStrainer
+from bs4.testing import skipIf
+from bs4.tests import test_htmlparser
+from bs4.testing import (
+ HTMLTreeBuilderSmokeTest,
+ XMLTreeBuilderSmokeTest,
+ SoupTest,
+ skipIf,
+)
+
+@skipIf(
+ not LXML_PRESENT,
+ "lxml seems not to be present, not testing its tree builder.")
+class LXMLTreeBuilderSmokeTest(SoupTest, HTMLTreeBuilderSmokeTest):
+ """See ``HTMLTreeBuilderSmokeTest``."""
+
+ @property
+ def default_builder(self):
+ return LXMLTreeBuilder()
+
+ def test_out_of_range_entity(self):
+ self.assertSoupEquals(
+ " foo�bar ", "foobar ")
+ self.assertSoupEquals(
+ "foo�bar ", "foobar ")
+ self.assertSoupEquals(
+ "foo�bar ", "foobar ")
+
+ # In lxml < 2.3.5, an empty doctype causes a segfault. Skip this
+ # test if an old version of lxml is installed.
+
+ @skipIf(
+ not LXML_PRESENT or LXML_VERSION < (2,3,5,0),
+ "Skipping doctype test for old version of lxml to avoid segfault.")
+ def test_empty_doctype(self):
+ soup = self.soup("")
+ doctype = soup.contents[0]
+ self.assertEqual("", doctype.strip())
+
+ def test_beautifulstonesoup_is_xml_parser(self):
+ # Make sure that the deprecated BSS class uses an xml builder
+ # if one is installed.
+ with warnings.catch_warnings(record=True) as w:
+ soup = BeautifulStoneSoup("")
+ self.assertEqual(u"", unicode(soup.b))
+ self.assertTrue("BeautifulStoneSoup class is deprecated" in str(w[0].message))
+
+@skipIf(
+ not LXML_PRESENT,
+ "lxml seems not to be present, not testing its XML tree builder.")
+class LXMLXMLTreeBuilderSmokeTest(SoupTest, XMLTreeBuilderSmokeTest):
+ """See ``HTMLTreeBuilderSmokeTest``."""
+
+ @property
+ def default_builder(self):
+ return LXMLTreeBuilderForXML()
diff --git a/src/bs4/tests/test_soup.py b/src/bs4/tests/test_soup.py
new file mode 100644
index 0000000..f3e69ed
--- /dev/null
+++ b/src/bs4/tests/test_soup.py
@@ -0,0 +1,501 @@
+# -*- coding: utf-8 -*-
+"""Tests of Beautiful Soup as a whole."""
+
+from pdb import set_trace
+import logging
+import unittest
+import sys
+import tempfile
+
+from bs4 import (
+ BeautifulSoup,
+ BeautifulStoneSoup,
+)
+from bs4.element import (
+ CharsetMetaAttributeValue,
+ ContentMetaAttributeValue,
+ SoupStrainer,
+ NamespacedAttribute,
+ )
+import bs4.dammit
+from bs4.dammit import (
+ EntitySubstitution,
+ UnicodeDammit,
+ EncodingDetector,
+)
+from bs4.testing import (
+ SoupTest,
+ skipIf,
+)
+import warnings
+
+try:
+ from bs4.builder import LXMLTreeBuilder, LXMLTreeBuilderForXML
+ LXML_PRESENT = True
+except ImportError, e:
+ LXML_PRESENT = False
+
+PYTHON_3_PRE_3_2 = (sys.version_info[0] == 3 and sys.version_info < (3,2))
+
+class TestConstructor(SoupTest):
+
+ def test_short_unicode_input(self):
+ data = u"éé"
+ soup = self.soup(data)
+ self.assertEqual(u"éé", soup.h1.string)
+
+ def test_embedded_null(self):
+ data = u"foo\0bar"
+ soup = self.soup(data)
+ self.assertEqual(u"foo\0bar", soup.h1.string)
+
+ def test_exclude_encodings(self):
+ utf8_data = u"Räksmörgås".encode("utf-8")
+ soup = self.soup(utf8_data, exclude_encodings=["utf-8"])
+ self.assertEqual("windows-1252", soup.original_encoding)
+
+
+class TestWarnings(SoupTest):
+
+ def _no_parser_specified(self, s, is_there=True):
+ v = s.startswith(BeautifulSoup.NO_PARSER_SPECIFIED_WARNING[:80])
+ self.assertTrue(v)
+
+ def test_warning_if_no_parser_specified(self):
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup("")
+ msg = str(w[0].message)
+ self._assert_no_parser_specified(msg)
+
+ def test_warning_if_parser_specified_too_vague(self):
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup("", "html")
+ msg = str(w[0].message)
+ self._assert_no_parser_specified(msg)
+
+ def test_no_warning_if_explicit_parser_specified(self):
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup("", "html.parser")
+ self.assertEqual([], w)
+
+ def test_parseOnlyThese_renamed_to_parse_only(self):
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup("", parseOnlyThese=SoupStrainer("b"))
+ msg = str(w[0].message)
+ self.assertTrue("parseOnlyThese" in msg)
+ self.assertTrue("parse_only" in msg)
+ self.assertEqual(b"", soup.encode())
+
+ def test_fromEncoding_renamed_to_from_encoding(self):
+ with warnings.catch_warnings(record=True) as w:
+ utf8 = b"\xc3\xa9"
+ soup = self.soup(utf8, fromEncoding="utf8")
+ msg = str(w[0].message)
+ self.assertTrue("fromEncoding" in msg)
+ self.assertTrue("from_encoding" in msg)
+ self.assertEqual("utf8", soup.original_encoding)
+
+ def test_unrecognized_keyword_argument(self):
+ self.assertRaises(
+ TypeError, self.soup, "", no_such_argument=True)
+
+class TestWarnings(SoupTest):
+
+ def test_disk_file_warning(self):
+ filehandle = tempfile.NamedTemporaryFile()
+ filename = filehandle.name
+ try:
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup(filename)
+ msg = str(w[0].message)
+ self.assertTrue("looks like a filename" in msg)
+ finally:
+ filehandle.close()
+
+ # The file no longer exists, so Beautiful Soup will no longer issue the warning.
+ with warnings.catch_warnings(record=True) as w:
+ soup = self.soup(filename)
+ self.assertEqual(0, len(w))
+
+ def test_url_warning_with_bytes_url(self):
+ with warnings.catch_warnings(record=True) as warning_list:
+ soup = self.soup(b"http://www.crummybytes.com/")
+ # Be aware this isn't the only warning that can be raised during
+ # execution..
+ self.assertTrue(any("looks like a URL" in str(w.message)
+ for w in warning_list))
+
+ def test_url_warning_with_unicode_url(self):
+ with warnings.catch_warnings(record=True) as warning_list:
+ # note - this url must differ from the bytes one otherwise
+ # python's warnings system swallows the second warning
+ soup = self.soup(u"http://www.crummyunicode.com/")
+ self.assertTrue(any("looks like a URL" in str(w.message)
+ for w in warning_list))
+
+ def test_url_warning_with_bytes_and_space(self):
+ with warnings.catch_warnings(record=True) as warning_list:
+ soup = self.soup(b"http://www.crummybytes.com/ is great")
+ self.assertFalse(any("looks like a URL" in str(w.message)
+ for w in warning_list))
+
+ def test_url_warning_with_unicode_and_space(self):
+ with warnings.catch_warnings(record=True) as warning_list:
+ soup = self.soup(u"http://www.crummyuncode.com/ is great")
+ self.assertFalse(any("looks like a URL" in str(w.message)
+ for w in warning_list))
+
+
+class TestSelectiveParsing(SoupTest):
+
+ def test_parse_with_soupstrainer(self):
+ markup = "NoYesNoYes Yes"
+ strainer = SoupStrainer("b")
+ soup = self.soup(markup, parse_only=strainer)
+ self.assertEqual(soup.encode(), b"YesYes Yes")
+
+
+class TestEntitySubstitution(unittest.TestCase):
+ """Standalone tests of the EntitySubstitution class."""
+ def setUp(self):
+ self.sub = EntitySubstitution
+
+ def test_simple_html_substitution(self):
+ # Unicode characters corresponding to named HTML entites
+ # are substituted, and no others.
+ s = u"foo\u2200\N{SNOWMAN}\u00f5bar"
+ self.assertEqual(self.sub.substitute_html(s),
+ u"foo∀\N{SNOWMAN}õbar")
+
+ def test_smart_quote_substitution(self):
+ # MS smart quotes are a common source of frustration, so we
+ # give them a special test.
+ quotes = b"\x91\x92foo\x93\x94"
+ dammit = UnicodeDammit(quotes)
+ self.assertEqual(self.sub.substitute_html(dammit.markup),
+ "‘’foo“”")
+
+ def test_xml_converstion_includes_no_quotes_if_make_quoted_attribute_is_false(self):
+ s = 'Welcome to "my bar"'
+ self.assertEqual(self.sub.substitute_xml(s, False), s)
+
+ def test_xml_attribute_quoting_normally_uses_double_quotes(self):
+ self.assertEqual(self.sub.substitute_xml("Welcome", True),
+ '"Welcome"')
+ self.assertEqual(self.sub.substitute_xml("Bob's Bar", True),
+ '"Bob\'s Bar"')
+
+ def test_xml_attribute_quoting_uses_single_quotes_when_value_contains_double_quotes(self):
+ s = 'Welcome to "my bar"'
+ self.assertEqual(self.sub.substitute_xml(s, True),
+ "'Welcome to \"my bar\"'")
+
+ def test_xml_attribute_quoting_escapes_single_quotes_when_value_contains_both_single_and_double_quotes(self):
+ s = 'Welcome to "Bob\'s Bar"'
+ self.assertEqual(
+ self.sub.substitute_xml(s, True),
+ '"Welcome to "Bob\'s Bar""')
+
+ def test_xml_quotes_arent_escaped_when_value_is_not_being_quoted(self):
+ quoted = 'Welcome to "Bob\'s Bar"'
+ self.assertEqual(self.sub.substitute_xml(quoted), quoted)
+
+ def test_xml_quoting_handles_angle_brackets(self):
+ self.assertEqual(
+ self.sub.substitute_xml("foo"),
+ "foo<bar>")
+
+ def test_xml_quoting_handles_ampersands(self):
+ self.assertEqual(self.sub.substitute_xml("AT&T"), "AT&T")
+
+ def test_xml_quoting_including_ampersands_when_they_are_part_of_an_entity(self):
+ self.assertEqual(
+ self.sub.substitute_xml("ÁT&T"),
+ "ÁT&T")
+
+ def test_xml_quoting_ignoring_ampersands_when_they_are_part_of_an_entity(self):
+ self.assertEqual(
+ self.sub.substitute_xml_containing_entities("ÁT&T"),
+ "ÁT&T")
+
+ def test_quotes_not_html_substituted(self):
+ """There's no need to do this except inside attribute values."""
+ text = 'Bob\'s "bar"'
+ self.assertEqual(self.sub.substitute_html(text), text)
+
+
+class TestEncodingConversion(SoupTest):
+ # Test Beautiful Soup's ability to decode and encode from various
+ # encodings.
+
+ def setUp(self):
+ super(TestEncodingConversion, self).setUp()
+ self.unicode_data = u'Sacr\N{LATIN SMALL LETTER E WITH ACUTE} bleu!'
+ self.utf8_data = self.unicode_data.encode("utf-8")
+ # Just so you know what it looks like.
+ self.assertEqual(
+ self.utf8_data,
+ b'Sacr\xc3\xa9 bleu!')
+
+ def test_ascii_in_unicode_out(self):
+ # ASCII input is converted to Unicode. The original_encoding
+ # attribute is set to 'utf-8', a superset of ASCII.
+ chardet = bs4.dammit.chardet_dammit
+ logging.disable(logging.WARNING)
+ try:
+ def noop(str):
+ return None
+ # Disable chardet, which will realize that the ASCII is ASCII.
+ bs4.dammit.chardet_dammit = noop
+ ascii = b"a"
+ soup_from_ascii = self.soup(ascii)
+ unicode_output = soup_from_ascii.decode()
+ self.assertTrue(isinstance(unicode_output, unicode))
+ self.assertEqual(unicode_output, self.document_for(ascii.decode()))
+ self.assertEqual(soup_from_ascii.original_encoding.lower(), "utf-8")
+ finally:
+ logging.disable(logging.NOTSET)
+ bs4.dammit.chardet_dammit = chardet
+
+ def test_unicode_in_unicode_out(self):
+ # Unicode input is left alone. The original_encoding attribute
+ # is not set.
+ soup_from_unicode = self.soup(self.unicode_data)
+ self.assertEqual(soup_from_unicode.decode(), self.unicode_data)
+ self.assertEqual(soup_from_unicode.foo.string, u'Sacr\xe9 bleu!')
+ self.assertEqual(soup_from_unicode.original_encoding, None)
+
+ def test_utf8_in_unicode_out(self):
+ # UTF-8 input is converted to Unicode. The original_encoding
+ # attribute is set.
+ soup_from_utf8 = self.soup(self.utf8_data)
+ self.assertEqual(soup_from_utf8.decode(), self.unicode_data)
+ self.assertEqual(soup_from_utf8.foo.string, u'Sacr\xe9 bleu!')
+
+ def test_utf8_out(self):
+ # The internal data structures can be encoded as UTF-8.
+ soup_from_unicode = self.soup(self.unicode_data)
+ self.assertEqual(soup_from_unicode.encode('utf-8'), self.utf8_data)
+
+ @skipIf(
+ PYTHON_3_PRE_3_2,
+ "Bad HTMLParser detected; skipping test of non-ASCII characters in attribute name.")
+ def test_attribute_name_containing_unicode_characters(self):
+ markup = u''
+ self.assertEqual(self.soup(markup).div.encode("utf8"), markup.encode("utf8"))
+
+class TestUnicodeDammit(unittest.TestCase):
+ """Standalone tests of UnicodeDammit."""
+
+ def test_unicode_input(self):
+ markup = u"I'm already Unicode! \N{SNOWMAN}"
+ dammit = UnicodeDammit(markup)
+ self.assertEqual(dammit.unicode_markup, markup)
+
+ def test_smart_quotes_to_unicode(self):
+ markup = b"\x91\x92\x93\x94"
+ dammit = UnicodeDammit(markup)
+ self.assertEqual(
+ dammit.unicode_markup, u"\u2018\u2019\u201c\u201d")
+
+ def test_smart_quotes_to_xml_entities(self):
+ markup = b"\x91\x92\x93\x94"
+ dammit = UnicodeDammit(markup, smart_quotes_to="xml")
+ self.assertEqual(
+ dammit.unicode_markup, "‘’“”")
+
+ def test_smart_quotes_to_html_entities(self):
+ markup = b"\x91\x92\x93\x94"
+ dammit = UnicodeDammit(markup, smart_quotes_to="html")
+ self.assertEqual(
+ dammit.unicode_markup, "‘’“”")
+
+ def test_smart_quotes_to_ascii(self):
+ markup = b"\x91\x92\x93\x94"
+ dammit = UnicodeDammit(markup, smart_quotes_to="ascii")
+ self.assertEqual(
+ dammit.unicode_markup, """''""""")
+
+ def test_detect_utf8(self):
+ utf8 = b"Sacr\xc3\xa9 bleu! \xe2\x98\x83"
+ dammit = UnicodeDammit(utf8)
+ self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+ self.assertEqual(dammit.unicode_markup, u'Sacr\xe9 bleu! \N{SNOWMAN}')
+
+
+ def test_convert_hebrew(self):
+ hebrew = b"\xed\xe5\xec\xf9"
+ dammit = UnicodeDammit(hebrew, ["iso-8859-8"])
+ self.assertEqual(dammit.original_encoding.lower(), 'iso-8859-8')
+ self.assertEqual(dammit.unicode_markup, u'\u05dd\u05d5\u05dc\u05e9')
+
+ def test_dont_see_smart_quotes_where_there_are_none(self):
+ utf_8 = b"\343\202\261\343\203\274\343\202\277\343\202\244 Watch"
+ dammit = UnicodeDammit(utf_8)
+ self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+ self.assertEqual(dammit.unicode_markup.encode("utf-8"), utf_8)
+
+ def test_ignore_inappropriate_codecs(self):
+ utf8_data = u"Räksmörgås".encode("utf-8")
+ dammit = UnicodeDammit(utf8_data, ["iso-8859-8"])
+ self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+
+ def test_ignore_invalid_codecs(self):
+ utf8_data = u"Räksmörgås".encode("utf-8")
+ for bad_encoding in ['.utf8', '...', 'utF---16.!']:
+ dammit = UnicodeDammit(utf8_data, [bad_encoding])
+ self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+
+ def test_exclude_encodings(self):
+ # This is UTF-8.
+ utf8_data = u"Räksmörgås".encode("utf-8")
+
+ # But if we exclude UTF-8 from consideration, the guess is
+ # Windows-1252.
+ dammit = UnicodeDammit(utf8_data, exclude_encodings=["utf-8"])
+ self.assertEqual(dammit.original_encoding.lower(), 'windows-1252')
+
+ # And if we exclude that, there is no valid guess at all.
+ dammit = UnicodeDammit(
+ utf8_data, exclude_encodings=["utf-8", "windows-1252"])
+ self.assertEqual(dammit.original_encoding, None)
+
+ def test_encoding_detector_replaces_junk_in_encoding_name_with_replacement_character(self):
+ detected = EncodingDetector(
+ b'')
+ encodings = list(detected.encodings)
+ assert u'utf-\N{REPLACEMENT CHARACTER}' in encodings
+
+ def test_detect_html5_style_meta_tag(self):
+
+ for data in (
+ b'',
+ b"",
+ b"",
+ b""):
+ dammit = UnicodeDammit(data, is_html=True)
+ self.assertEqual(
+ "euc-jp", dammit.original_encoding)
+
+ def test_last_ditch_entity_replacement(self):
+ # This is a UTF-8 document that contains bytestrings
+ # completely incompatible with UTF-8 (ie. encoded with some other
+ # encoding).
+ #
+ # Since there is no consistent encoding for the document,
+ # Unicode, Dammit will eventually encode the document as UTF-8
+ # and encode the incompatible characters as REPLACEMENT
+ # CHARACTER.
+ #
+ # If chardet is installed, it will detect that the document
+ # can be converted into ISO-8859-1 without errors. This happens
+ # to be the wrong encoding, but it is a consistent encoding, so the
+ # code we're testing here won't run.
+ #
+ # So we temporarily disable chardet if it's present.
+ doc = b"""\357\273\277
+\330\250\330\252\330\261
+\310\322\321\220\312\321\355\344"""
+ chardet = bs4.dammit.chardet_dammit
+ logging.disable(logging.WARNING)
+ try:
+ def noop(str):
+ return None
+ bs4.dammit.chardet_dammit = noop
+ dammit = UnicodeDammit(doc)
+ self.assertEqual(True, dammit.contains_replacement_characters)
+ self.assertTrue(u"\ufffd" in dammit.unicode_markup)
+
+ soup = BeautifulSoup(doc, "html.parser")
+ self.assertTrue(soup.contains_replacement_characters)
+ finally:
+ logging.disable(logging.NOTSET)
+ bs4.dammit.chardet_dammit = chardet
+
+ def test_byte_order_mark_removed(self):
+ # A document written in UTF-16LE will have its byte order marker stripped.
+ data = b'\xff\xfe<\x00a\x00>\x00\xe1\x00\xe9\x00<\x00/\x00a\x00>\x00'
+ dammit = UnicodeDammit(data)
+ self.assertEqual(u"áé", dammit.unicode_markup)
+ self.assertEqual("utf-16le", dammit.original_encoding)
+
+ def test_detwingle(self):
+ # Here's a UTF8 document.
+ utf8 = (u"\N{SNOWMAN}" * 3).encode("utf8")
+
+ # Here's a Windows-1252 document.
+ windows_1252 = (
+ u"\N{LEFT DOUBLE QUOTATION MARK}Hi, I like Windows!"
+ u"\N{RIGHT DOUBLE QUOTATION MARK}").encode("windows_1252")
+
+ # Through some unholy alchemy, they've been stuck together.
+ doc = utf8 + windows_1252 + utf8
+
+ # The document can't be turned into UTF-8:
+ self.assertRaises(UnicodeDecodeError, doc.decode, "utf8")
+
+ # Unicode, Dammit thinks the whole document is Windows-1252,
+ # and decodes it into "☃☃☃“Hi, I like Windows!”☃☃☃"
+
+ # But if we run it through fix_embedded_windows_1252, it's fixed:
+
+ fixed = UnicodeDammit.detwingle(doc)
+ self.assertEqual(
+ u"☃☃☃“Hi, I like Windows!”☃☃☃", fixed.decode("utf8"))
+
+ def test_detwingle_ignores_multibyte_characters(self):
+ # Each of these characters has a UTF-8 representation ending
+ # in \x93. \x93 is a smart quote if interpreted as
+ # Windows-1252. But our code knows to skip over multibyte
+ # UTF-8 characters, so they'll survive the process unscathed.
+ for tricky_unicode_char in (
+ u"\N{LATIN SMALL LIGATURE OE}", # 2-byte char '\xc5\x93'
+ u"\N{LATIN SUBSCRIPT SMALL LETTER X}", # 3-byte char '\xe2\x82\x93'
+ u"\xf0\x90\x90\x93", # This is a CJK character, not sure which one.
+ ):
+ input = tricky_unicode_char.encode("utf8")
+ self.assertTrue(input.endswith(b'\x93'))
+ output = UnicodeDammit.detwingle(input)
+ self.assertEqual(output, input)
+
+class TestNamedspacedAttribute(SoupTest):
+
+ def test_name_may_be_none(self):
+ a = NamespacedAttribute("xmlns", None)
+ self.assertEqual(a, "xmlns")
+
+ def test_attribute_is_equivalent_to_colon_separated_string(self):
+ a = NamespacedAttribute("a", "b")
+ self.assertEqual("a:b", a)
+
+ def test_attributes_are_equivalent_if_prefix_and_name_identical(self):
+ a = NamespacedAttribute("a", "b", "c")
+ b = NamespacedAttribute("a", "b", "c")
+ self.assertEqual(a, b)
+
+ # The actual namespace is not considered.
+ c = NamespacedAttribute("a", "b", None)
+ self.assertEqual(a, c)
+
+ # But name and prefix are important.
+ d = NamespacedAttribute("a", "z", "c")
+ self.assertNotEqual(a, d)
+
+ e = NamespacedAttribute("z", "b", "c")
+ self.assertNotEqual(a, e)
+
+
+class TestAttributeValueWithCharsetSubstitution(unittest.TestCase):
+
+ def test_content_meta_attribute_value(self):
+ value = CharsetMetaAttributeValue("euc-jp")
+ self.assertEqual("euc-jp", value)
+ self.assertEqual("euc-jp", value.original_value)
+ self.assertEqual("utf8", value.encode("utf8"))
+
+
+ def test_content_meta_attribute_value(self):
+ value = ContentMetaAttributeValue("text/html; charset=euc-jp")
+ self.assertEqual("text/html; charset=euc-jp", value)
+ self.assertEqual("text/html; charset=euc-jp", value.original_value)
+ self.assertEqual("text/html; charset=utf8", value.encode("utf8"))
diff --git a/src/bs4/tests/test_tree.py b/src/bs4/tests/test_tree.py
new file mode 100644
index 0000000..a4fe0b1
--- /dev/null
+++ b/src/bs4/tests/test_tree.py
@@ -0,0 +1,2044 @@
+# -*- coding: utf-8 -*-
+"""Tests for Beautiful Soup's tree traversal methods.
+
+The tree traversal methods are the main advantage of using Beautiful
+Soup over just using a parser.
+
+Different parsers will build different Beautiful Soup trees given the
+same markup, but all Beautiful Soup trees can be traversed with the
+methods tested here.
+"""
+
+from pdb import set_trace
+import copy
+import pickle
+import re
+import warnings
+from bs4 import BeautifulSoup
+from bs4.builder import (
+ builder_registry,
+ HTMLParserTreeBuilder,
+)
+from bs4.element import (
+ PY3K,
+ CData,
+ Comment,
+ Declaration,
+ Doctype,
+ NavigableString,
+ SoupStrainer,
+ Tag,
+)
+from bs4.testing import (
+ SoupTest,
+ skipIf,
+)
+
+XML_BUILDER_PRESENT = (builder_registry.lookup("xml") is not None)
+LXML_PRESENT = (builder_registry.lookup("lxml") is not None)
+
+class TreeTest(SoupTest):
+
+ def assertSelects(self, tags, should_match):
+ """Make sure that the given tags have the correct text.
+
+ This is used in tests that define a bunch of tags, each
+ containing a single string, and then select certain strings by
+ some mechanism.
+ """
+ self.assertEqual([tag.string for tag in tags], should_match)
+
+ def assertSelectsIDs(self, tags, should_match):
+ """Make sure that the given tags have the correct IDs.
+
+ This is used in tests that define a bunch of tags, each
+ containing a single string, and then select certain strings by
+ some mechanism.
+ """
+ self.assertEqual([tag['id'] for tag in tags], should_match)
+
+
+class TestFind(TreeTest):
+ """Basic tests of the find() method.
+
+ find() just calls find_all() with limit=1, so it's not tested all
+ that thouroughly here.
+ """
+
+ def test_find_tag(self):
+ soup = self.soup("1234")
+ self.assertEqual(soup.find("b").string, "2")
+
+ def test_unicode_text_find(self):
+ soup = self.soup(u'Räksmörgås')
+ self.assertEqual(soup.find(string=u'Räksmörgås'), u'Räksmörgås')
+
+ def test_unicode_attribute_find(self):
+ soup = self.soup(u'here it is')
+ str(soup)
+ self.assertEqual("here it is", soup.find(id=u'Räksmörgås').text)
+
+
+ def test_find_everything(self):
+ """Test an optimization that finds all tags."""
+ soup = self.soup("foobar")
+ self.assertEqual(2, len(soup.find_all()))
+
+ def test_find_everything_with_name(self):
+ """Test an optimization that finds all tags with a given name."""
+ soup = self.soup("foobarbaz")
+ self.assertEqual(2, len(soup.find_all('a')))
+
+class TestFindAll(TreeTest):
+ """Basic tests of the find_all() method."""
+
+ def test_find_all_text_nodes(self):
+ """You can search the tree for text nodes."""
+ soup = self.soup("Foobar\xbb")
+ # Exact match.
+ self.assertEqual(soup.find_all(string="bar"), [u"bar"])
+ self.assertEqual(soup.find_all(text="bar"), [u"bar"])
+ # Match any of a number of strings.
+ self.assertEqual(
+ soup.find_all(text=["Foo", "bar"]), [u"Foo", u"bar"])
+ # Match a regular expression.
+ self.assertEqual(soup.find_all(text=re.compile('.*')),
+ [u"Foo", u"bar", u'\xbb'])
+ # Match anything.
+ self.assertEqual(soup.find_all(text=True),
+ [u"Foo", u"bar", u'\xbb'])
+
+ def test_find_all_limit(self):
+ """You can limit the number of items returned by find_all."""
+ soup = self.soup("12345")
+ self.assertSelects(soup.find_all('a', limit=3), ["1", "2", "3"])
+ self.assertSelects(soup.find_all('a', limit=1), ["1"])
+ self.assertSelects(
+ soup.find_all('a', limit=10), ["1", "2", "3", "4", "5"])
+
+ # A limit of 0 means no limit.
+ self.assertSelects(
+ soup.find_all('a', limit=0), ["1", "2", "3", "4", "5"])
+
+ def test_calling_a_tag_is_calling_findall(self):
+ soup = self.soup("123")
+ self.assertSelects(soup('a', limit=1), ["1"])
+ self.assertSelects(soup.b(id="foo"), ["3"])
+
+ def test_find_all_with_self_referential_data_structure_does_not_cause_infinite_recursion(self):
+ soup = self.soup("")
+ # Create a self-referential list.
+ l = []
+ l.append(l)
+
+ # Without special code in _normalize_search_value, this would cause infinite
+ # recursion.
+ self.assertEqual([], soup.find_all(l))
+
+ def test_find_all_resultset(self):
+ """All find_all calls return a ResultSet"""
+ soup = self.soup("")
+ result = soup.find_all("a")
+ self.assertTrue(hasattr(result, "source"))
+
+ result = soup.find_all(True)
+ self.assertTrue(hasattr(result, "source"))
+
+ result = soup.find_all(text="foo")
+ self.assertTrue(hasattr(result, "source"))
+
+
+class TestFindAllBasicNamespaces(TreeTest):
+
+ def test_find_by_namespaced_name(self):
+ soup = self.soup('4')
+ self.assertEqual("4", soup.find("mathml:msqrt").string)
+ self.assertEqual("a", soup.find(attrs= { "svg:fill" : "red" }).name)
+
+
+class TestFindAllByName(TreeTest):
+ """Test ways of finding tags by tag name."""
+
+ def setUp(self):
+ super(TreeTest, self).setUp()
+ self.tree = self.soup("""First tag.
+ Second tag.
+ Third Nested tag. tag.""")
+
+ def test_find_all_by_tag_name(self):
+ # Find all the tags.
+ self.assertSelects(
+ self.tree.find_all('a'), ['First tag.', 'Nested tag.'])
+
+ def test_find_all_by_name_and_text(self):
+ self.assertSelects(
+ self.tree.find_all('a', text='First tag.'), ['First tag.'])
+
+ self.assertSelects(
+ self.tree.find_all('a', text=True), ['First tag.', 'Nested tag.'])
+
+ self.assertSelects(
+ self.tree.find_all('a', text=re.compile("tag")),
+ ['First tag.', 'Nested tag.'])
+
+
+ def test_find_all_on_non_root_element(self):
+ # You can call find_all on any node, not just the root.
+ self.assertSelects(self.tree.c.find_all('a'), ['Nested tag.'])
+
+ def test_calling_element_invokes_find_all(self):
+ self.assertSelects(self.tree('a'), ['First tag.', 'Nested tag.'])
+
+ def test_find_all_by_tag_strainer(self):
+ self.assertSelects(
+ self.tree.find_all(SoupStrainer('a')),
+ ['First tag.', 'Nested tag.'])
+
+ def test_find_all_by_tag_names(self):
+ self.assertSelects(
+ self.tree.find_all(['a', 'b']),
+ ['First tag.', 'Second tag.', 'Nested tag.'])
+
+ def test_find_all_by_tag_dict(self):
+ self.assertSelects(
+ self.tree.find_all({'a' : True, 'b' : True}),
+ ['First tag.', 'Second tag.', 'Nested tag.'])
+
+ def test_find_all_by_tag_re(self):
+ self.assertSelects(
+ self.tree.find_all(re.compile('^[ab]$')),
+ ['First tag.', 'Second tag.', 'Nested tag.'])
+
+ def test_find_all_with_tags_matching_method(self):
+ # You can define an oracle method that determines whether
+ # a tag matches the search.
+ def id_matches_name(tag):
+ return tag.name == tag.get('id')
+
+ tree = self.soup("""Match 1.
+ Does not match.
+ Match 2.""")
+
+ self.assertSelects(
+ tree.find_all(id_matches_name), ["Match 1.", "Match 2."])
+
+ def test_find_with_multi_valued_attribute(self):
+ soup = self.soup(
+ "1 2 3 "
+ )
+ r1 = soup.find('div', 'a d');
+ r2 = soup.find('div', re.compile(r'a d'));
+ r3, r4 = soup.find_all('div', ['a b', 'a d']);
+ self.assertEqual('3', r1.string)
+ self.assertEqual('3', r2.string)
+ self.assertEqual('1', r3.string)
+ self.assertEqual('3', r4.string)
+
+class TestFindAllByAttribute(TreeTest):
+
+ def test_find_all_by_attribute_name(self):
+ # You can pass in keyword arguments to find_all to search by
+ # attribute.
+ tree = self.soup("""
+ Matching a.
+
+ Non-matching Matching b.a.
+ """)
+ self.assertSelects(tree.find_all(id='first'),
+ ["Matching a.", "Matching b."])
+
+ def test_find_all_by_utf8_attribute_value(self):
+ peace = u"םולש".encode("utf8")
+ data = u''.encode("utf8")
+ soup = self.soup(data)
+ self.assertEqual([soup.a], soup.find_all(title=peace))
+ self.assertEqual([soup.a], soup.find_all(title=peace.decode("utf8")))
+ self.assertEqual([soup.a], soup.find_all(title=[peace, "something else"]))
+
+ def test_find_all_by_attribute_dict(self):
+ # You can pass in a dictionary as the argument 'attrs'. This
+ # lets you search for attributes like 'name' (a fixed argument
+ # to find_all) and 'class' (a reserved word in Python.)
+ tree = self.soup("""
+ Name match.
+ Class match.
+ Non-match.
+ A tag called 'name1'.
+ """)
+
+ # This doesn't do what you want.
+ self.assertSelects(tree.find_all(name='name1'),
+ ["A tag called 'name1'."])
+ # This does what you want.
+ self.assertSelects(tree.find_all(attrs={'name' : 'name1'}),
+ ["Name match."])
+
+ self.assertSelects(tree.find_all(attrs={'class' : 'class2'}),
+ ["Class match."])
+
+ def test_find_all_by_class(self):
+ tree = self.soup("""
+ Class 1.
+ Class 2.
+ Class 1.
+ Class 3 and 4.
+ """)
+
+ # Passing in the class_ keyword argument will search against
+ # the 'class' attribute.
+ self.assertSelects(tree.find_all('a', class_='1'), ['Class 1.'])
+ self.assertSelects(tree.find_all('c', class_='3'), ['Class 3 and 4.'])
+ self.assertSelects(tree.find_all('c', class_='4'), ['Class 3 and 4.'])
+
+ # Passing in a string to 'attrs' will also search the CSS class.
+ self.assertSelects(tree.find_all('a', '1'), ['Class 1.'])
+ self.assertSelects(tree.find_all(attrs='1'), ['Class 1.', 'Class 1.'])
+ self.assertSelects(tree.find_all('c', '3'), ['Class 3 and 4.'])
+ self.assertSelects(tree.find_all('c', '4'), ['Class 3 and 4.'])
+
+ def test_find_by_class_when_multiple_classes_present(self):
+ tree = self.soup("Found it")
+
+ f = tree.find_all("gar", class_=re.compile("o"))
+ self.assertSelects(f, ["Found it"])
+
+ f = tree.find_all("gar", class_=re.compile("a"))
+ self.assertSelects(f, ["Found it"])
+
+ # If the search fails to match the individual strings "foo" and "bar",
+ # it will be tried against the combined string "foo bar".
+ f = tree.find_all("gar", class_=re.compile("o b"))
+ self.assertSelects(f, ["Found it"])
+
+ def test_find_all_with_non_dictionary_for_attrs_finds_by_class(self):
+ soup = self.soup("Found it")
+
+ self.assertSelects(soup.find_all("a", re.compile("ba")), ["Found it"])
+
+ def big_attribute_value(value):
+ return len(value) > 3
+
+ self.assertSelects(soup.find_all("a", big_attribute_value), [])
+
+ def small_attribute_value(value):
+ return len(value) <= 3
+
+ self.assertSelects(
+ soup.find_all("a", small_attribute_value), ["Found it"])
+
+ def test_find_all_with_string_for_attrs_finds_multiple_classes(self):
+ soup = self.soup('')
+ a, a2 = soup.find_all("a")
+ self.assertEqual([a, a2], soup.find_all("a", "foo"))
+ self.assertEqual([a], soup.find_all("a", "bar"))
+
+ # If you specify the class as a string that contains a
+ # space, only that specific value will be found.
+ self.assertEqual([a], soup.find_all("a", class_="foo bar"))
+ self.assertEqual([a], soup.find_all("a", "foo bar"))
+ self.assertEqual([], soup.find_all("a", "bar foo"))
+
+ def test_find_all_by_attribute_soupstrainer(self):
+ tree = self.soup("""
+ Match.
+ Non-match.""")
+
+ strainer = SoupStrainer(attrs={'id' : 'first'})
+ self.assertSelects(tree.find_all(strainer), ['Match.'])
+
+ def test_find_all_with_missing_attribute(self):
+ # You can pass in None as the value of an attribute to find_all.
+ # This will match tags that do not have that attribute set.
+ tree = self.soup("""ID present.
+ No ID present.
+ ID is empty.""")
+ self.assertSelects(tree.find_all('a', id=None), ["No ID present."])
+
+ def test_find_all_with_defined_attribute(self):
+ # You can pass in None as the value of an attribute to find_all.
+ # This will match tags that have that attribute set to any value.
+ tree = self.soup("""ID present.
+ No ID present.
+ ID is empty.""")
+ self.assertSelects(
+ tree.find_all(id=True), ["ID present.", "ID is empty."])
+
+ def test_find_all_with_numeric_attribute(self):
+ # If you search for a number, it's treated as a string.
+ tree = self.soup("""Unquoted attribute.
+ Quoted attribute.""")
+
+ expected = ["Unquoted attribute.", "Quoted attribute."]
+ self.assertSelects(tree.find_all(id=1), expected)
+ self.assertSelects(tree.find_all(id="1"), expected)
+
+ def test_find_all_with_list_attribute_values(self):
+ # You can pass a list of attribute values instead of just one,
+ # and you'll get tags that match any of the values.
+ tree = self.soup("""1
+ 2
+ 3
+ No ID.""")
+ self.assertSelects(tree.find_all(id=["1", "3", "4"]),
+ ["1", "3"])
+
+ def test_find_all_with_regular_expression_attribute_value(self):
+ # You can pass a regular expression as an attribute value, and
+ # you'll get tags whose values for that attribute match the
+ # regular expression.
+ tree = self.soup("""One a.
+ Two as.
+ Mixed as and bs.
+ One b.
+ No ID.""")
+
+ self.assertSelects(tree.find_all(id=re.compile("^a+$")),
+ ["One a.", "Two as."])
+
+ def test_find_by_name_and_containing_string(self):
+ soup = self.soup("foobarfoo")
+ a = soup.a
+
+ self.assertEqual([a], soup.find_all("a", text="foo"))
+ self.assertEqual([], soup.find_all("a", text="bar"))
+ self.assertEqual([], soup.find_all("a", text="bar"))
+
+ def test_find_by_name_and_containing_string_when_string_is_buried(self):
+ soup = self.soup("foofoo")
+ self.assertEqual(soup.find_all("a"), soup.find_all("a", text="foo"))
+
+ def test_find_by_attribute_and_containing_string(self):
+ soup = self.soup('foofoo')
+ a = soup.a
+
+ self.assertEqual([a], soup.find_all(id=2, text="foo"))
+ self.assertEqual([], soup.find_all(id=1, text="bar"))
+
+
+
+
+class TestIndex(TreeTest):
+ """Test Tag.index"""
+ def test_index(self):
+ tree = self.soup("""
+ Identical
+ Not identical
+ Identical
+
+ Identical with child
+ Also not identical
+ Identical with child
+
+ ''')
+ self.start = self.tree.b
+
+
+ def test_parent(self):
+ self.assertEqual(self.start.parent['id'], 'bottom')
+ self.assertEqual(self.start.parent.parent['id'], 'middle')
+ self.assertEqual(self.start.parent.parent.parent['id'], 'top')
+
+ def test_parent_of_top_tag_is_soup_object(self):
+ top_tag = self.tree.contents[0]
+ self.assertEqual(top_tag.parent, self.tree)
+
+ def test_soup_object_has_no_parent(self):
+ self.assertEqual(None, self.tree.parent)
+
+ def test_find_parents(self):
+ self.assertSelectsIDs(
+ self.start.find_parents('ul'), ['bottom', 'middle', 'top'])
+ self.assertSelectsIDs(
+ self.start.find_parents('ul', id="middle"), ['middle'])
+
+ def test_find_parent(self):
+ self.assertEqual(self.start.find_parent('ul')['id'], 'bottom')
+ self.assertEqual(self.start.find_parent('ul', id='top')['id'], 'top')
+
+ def test_parent_of_text_element(self):
+ text = self.tree.find(text="Start here")
+ self.assertEqual(text.parent.name, 'b')
+
+ def test_text_element_find_parent(self):
+ text = self.tree.find(text="Start here")
+ self.assertEqual(text.find_parent('ul')['id'], 'bottom')
+
+ def test_parent_generator(self):
+ parents = [parent['id'] for parent in self.start.parents
+ if parent is not None and 'id' in parent.attrs]
+ self.assertEqual(parents, ['bottom', 'middle', 'top'])
+
+
+class ProximityTest(TreeTest):
+
+ def setUp(self):
+ super(TreeTest, self).setUp()
+ self.tree = self.soup(
+ 'OneTwoThree')
+
+
+class TestNextOperations(ProximityTest):
+
+ def setUp(self):
+ super(TestNextOperations, self).setUp()
+ self.start = self.tree.b
+
+ def test_next(self):
+ self.assertEqual(self.start.next_element, "One")
+ self.assertEqual(self.start.next_element.next_element['id'], "2")
+
+ def test_next_of_last_item_is_none(self):
+ last = self.tree.find(text="Three")
+ self.assertEqual(last.next_element, None)
+
+ def test_next_of_root_is_none(self):
+ # The document root is outside the next/previous chain.
+ self.assertEqual(self.tree.next_element, None)
+
+ def test_find_all_next(self):
+ self.assertSelects(self.start.find_all_next('b'), ["Two", "Three"])
+ self.start.find_all_next(id=3)
+ self.assertSelects(self.start.find_all_next(id=3), ["Three"])
+
+ def test_find_next(self):
+ self.assertEqual(self.start.find_next('b')['id'], '2')
+ self.assertEqual(self.start.find_next(text="Three"), "Three")
+
+ def test_find_next_for_text_element(self):
+ text = self.tree.find(text="One")
+ self.assertEqual(text.find_next("b").string, "Two")
+ self.assertSelects(text.find_all_next("b"), ["Two", "Three"])
+
+ def test_next_generator(self):
+ start = self.tree.find(text="Two")
+ successors = [node for node in start.next_elements]
+ # There are two successors: the final tag and its text contents.
+ tag, contents = successors
+ self.assertEqual(tag['id'], '3')
+ self.assertEqual(contents, "Three")
+
+class TestPreviousOperations(ProximityTest):
+
+ def setUp(self):
+ super(TestPreviousOperations, self).setUp()
+ self.end = self.tree.find(text="Three")
+
+ def test_previous(self):
+ self.assertEqual(self.end.previous_element['id'], "3")
+ self.assertEqual(self.end.previous_element.previous_element, "Two")
+
+ def test_previous_of_first_item_is_none(self):
+ first = self.tree.find('html')
+ self.assertEqual(first.previous_element, None)
+
+ def test_previous_of_root_is_none(self):
+ # The document root is outside the next/previous chain.
+ # XXX This is broken!
+ #self.assertEqual(self.tree.previous_element, None)
+ pass
+
+ def test_find_all_previous(self):
+ # The tag containing the "Three" node is the predecessor
+ # of the "Three" node itself, which is why "Three" shows up
+ # here.
+ self.assertSelects(
+ self.end.find_all_previous('b'), ["Three", "Two", "One"])
+ self.assertSelects(self.end.find_all_previous(id=1), ["One"])
+
+ def test_find_previous(self):
+ self.assertEqual(self.end.find_previous('b')['id'], '3')
+ self.assertEqual(self.end.find_previous(text="One"), "One")
+
+ def test_find_previous_for_text_element(self):
+ text = self.tree.find(text="Three")
+ self.assertEqual(text.find_previous("b").string, "Three")
+ self.assertSelects(
+ text.find_all_previous("b"), ["Three", "Two", "One"])
+
+ def test_previous_generator(self):
+ start = self.tree.find(text="One")
+ predecessors = [node for node in start.previous_elements]
+
+ # There are four predecessors: the tag containing "One"
+ # the tag, the tag, and the tag.
+ b, body, head, html = predecessors
+ self.assertEqual(b['id'], '1')
+ self.assertEqual(body.name, "body")
+ self.assertEqual(head.name, "head")
+ self.assertEqual(html.name, "html")
+
+
+class SiblingTest(TreeTest):
+
+ def setUp(self):
+ super(SiblingTest, self).setUp()
+ markup = '''
+
+
+
+
+
+
+
+
+
+
+ '''
+ # All that whitespace looks good but makes the tests more
+ # difficult. Get rid of it.
+ markup = re.compile("\n\s*").sub("", markup)
+ self.tree = self.soup(markup)
+
+
+class TestNextSibling(SiblingTest):
+
+ def setUp(self):
+ super(TestNextSibling, self).setUp()
+ self.start = self.tree.find(id="1")
+
+ def test_next_sibling_of_root_is_none(self):
+ self.assertEqual(self.tree.next_sibling, None)
+
+ def test_next_sibling(self):
+ self.assertEqual(self.start.next_sibling['id'], '2')
+ self.assertEqual(self.start.next_sibling.next_sibling['id'], '3')
+
+ # Note the difference between next_sibling and next_element.
+ self.assertEqual(self.start.next_element['id'], '1.1')
+
+ def test_next_sibling_may_not_exist(self):
+ self.assertEqual(self.tree.html.next_sibling, None)
+
+ nested_span = self.tree.find(id="1.1")
+ self.assertEqual(nested_span.next_sibling, None)
+
+ last_span = self.tree.find(id="4")
+ self.assertEqual(last_span.next_sibling, None)
+
+ def test_find_next_sibling(self):
+ self.assertEqual(self.start.find_next_sibling('span')['id'], '2')
+
+ def test_next_siblings(self):
+ self.assertSelectsIDs(self.start.find_next_siblings("span"),
+ ['2', '3', '4'])
+
+ self.assertSelectsIDs(self.start.find_next_siblings(id='3'), ['3'])
+
+ def test_next_sibling_for_text_element(self):
+ soup = self.soup("Foobarbaz")
+ start = soup.find(text="Foo")
+ self.assertEqual(start.next_sibling.name, 'b')
+ self.assertEqual(start.next_sibling.next_sibling, 'baz')
+
+ self.assertSelects(start.find_next_siblings('b'), ['bar'])
+ self.assertEqual(start.find_next_sibling(text="baz"), "baz")
+ self.assertEqual(start.find_next_sibling(text="nonesuch"), None)
+
+
+class TestPreviousSibling(SiblingTest):
+
+ def setUp(self):
+ super(TestPreviousSibling, self).setUp()
+ self.end = self.tree.find(id="4")
+
+ def test_previous_sibling_of_root_is_none(self):
+ self.assertEqual(self.tree.previous_sibling, None)
+
+ def test_previous_sibling(self):
+ self.assertEqual(self.end.previous_sibling['id'], '3')
+ self.assertEqual(self.end.previous_sibling.previous_sibling['id'], '2')
+
+ # Note the difference between previous_sibling and previous_element.
+ self.assertEqual(self.end.previous_element['id'], '3.1')
+
+ def test_previous_sibling_may_not_exist(self):
+ self.assertEqual(self.tree.html.previous_sibling, None)
+
+ nested_span = self.tree.find(id="1.1")
+ self.assertEqual(nested_span.previous_sibling, None)
+
+ first_span = self.tree.find(id="1")
+ self.assertEqual(first_span.previous_sibling, None)
+
+ def test_find_previous_sibling(self):
+ self.assertEqual(self.end.find_previous_sibling('span')['id'], '3')
+
+ def test_previous_siblings(self):
+ self.assertSelectsIDs(self.end.find_previous_siblings("span"),
+ ['3', '2', '1'])
+
+ self.assertSelectsIDs(self.end.find_previous_siblings(id='1'), ['1'])
+
+ def test_previous_sibling_for_text_element(self):
+ soup = self.soup("Foobarbaz")
+ start = soup.find(text="baz")
+ self.assertEqual(start.previous_sibling.name, 'b')
+ self.assertEqual(start.previous_sibling.previous_sibling, 'Foo')
+
+ self.assertSelects(start.find_previous_siblings('b'), ['bar'])
+ self.assertEqual(start.find_previous_sibling(text="Foo"), "Foo")
+ self.assertEqual(start.find_previous_sibling(text="nonesuch"), None)
+
+
+class TestTagCreation(SoupTest):
+ """Test the ability to create new tags."""
+ def test_new_tag(self):
+ soup = self.soup("")
+ new_tag = soup.new_tag("foo", bar="baz")
+ self.assertTrue(isinstance(new_tag, Tag))
+ self.assertEqual("foo", new_tag.name)
+ self.assertEqual(dict(bar="baz"), new_tag.attrs)
+ self.assertEqual(None, new_tag.parent)
+
+ def test_tag_inherits_self_closing_rules_from_builder(self):
+ if XML_BUILDER_PRESENT:
+ xml_soup = BeautifulSoup("", "lxml-xml")
+ xml_br = xml_soup.new_tag("br")
+ xml_p = xml_soup.new_tag("p")
+
+ # Both the
and tag are empty-element, just because
+ # they have no contents.
+ self.assertEqual(b"
", xml_br.encode())
+ self.assertEqual(b" ", xml_p.encode())
+
+ html_soup = BeautifulSoup("", "html.parser")
+ html_br = html_soup.new_tag("br")
+ html_p = html_soup.new_tag("p")
+
+ # The HTML builder users HTML's rules about which tags are
+ # empty-element tags, and the new tags reflect these rules.
+ self.assertEqual(b"
", html_br.encode())
+ self.assertEqual(b"", html_p.encode())
+
+ def test_new_string_creates_navigablestring(self):
+ soup = self.soup("")
+ s = soup.new_string("foo")
+ self.assertEqual("foo", s)
+ self.assertTrue(isinstance(s, NavigableString))
+
+ def test_new_string_can_create_navigablestring_subclass(self):
+ soup = self.soup("")
+ s = soup.new_string("foo", Comment)
+ self.assertEqual("foo", s)
+ self.assertTrue(isinstance(s, Comment))
+
+class TestTreeModification(SoupTest):
+
+ def test_attribute_modification(self):
+ soup = self.soup('')
+ soup.a['id'] = 2
+ self.assertEqual(soup.decode(), self.document_for(''))
+ del(soup.a['id'])
+ self.assertEqual(soup.decode(), self.document_for(''))
+ soup.a['id2'] = 'foo'
+ self.assertEqual(soup.decode(), self.document_for(''))
+
+ def test_new_tag_creation(self):
+ builder = builder_registry.lookup('html')()
+ soup = self.soup("", builder=builder)
+ a = Tag(soup, builder, 'a')
+ ol = Tag(soup, builder, 'ol')
+ a['href'] = 'http://foo.com/'
+ soup.body.insert(0, a)
+ soup.body.insert(1, ol)
+ self.assertEqual(
+ soup.body.encode(),
+ b'
')
+
+ def test_append_to_contents_moves_tag(self):
+ doc = """Don't leave me here.
+ Don\'t leave! """
+ soup = self.soup(doc)
+ second_para = soup.find(id='2')
+ bold = soup.b
+
+ # Move the tag to the end of the second paragraph.
+ soup.find(id='2').append(soup.b)
+
+ # The tag is now a child of the second paragraph.
+ self.assertEqual(bold.parent, second_para)
+
+ self.assertEqual(
+ soup.decode(), self.document_for(
+ 'Don\'t leave me . \n'
+ 'Don\'t leave!here '))
+
+ def test_replace_with_returns_thing_that_was_replaced(self):
+ text = ""
+ soup = self.soup(text)
+ a = soup.a
+ new_a = a.replace_with(soup.c)
+ self.assertEqual(a, new_a)
+
+ def test_unwrap_returns_thing_that_was_replaced(self):
+ text = ""
+ soup = self.soup(text)
+ a = soup.a
+ new_a = a.unwrap()
+ self.assertEqual(a, new_a)
+
+ def test_replace_with_and_unwrap_give_useful_exception_when_tag_has_no_parent(self):
+ soup = self.soup("FooBar")
+ a = soup.a
+ a.extract()
+ self.assertEqual(None, a.parent)
+ self.assertRaises(ValueError, a.unwrap)
+ self.assertRaises(ValueError, a.replace_with, soup.c)
+
+ def test_replace_tag_with_itself(self):
+ text = "Foo"
+ soup = self.soup(text)
+ c = soup.c
+ soup.c.replace_with(c)
+ self.assertEqual(soup.decode(), self.document_for(text))
+
+ def test_replace_tag_with_its_parent_raises_exception(self):
+ text = ""
+ soup = self.soup(text)
+ self.assertRaises(ValueError, soup.b.replace_with, soup.a)
+
+ def test_insert_tag_into_itself_raises_exception(self):
+ text = ""
+ soup = self.soup(text)
+ self.assertRaises(ValueError, soup.a.insert, 0, soup.a)
+
+ def test_replace_with_maintains_next_element_throughout(self):
+ soup = self.soup('onethree ')
+ a = soup.a
+ b = a.contents[0]
+ # Make it so the tag has two text children.
+ a.insert(1, "two")
+
+ # Now replace each one with the empty string.
+ left, right = a.contents
+ left.replaceWith('')
+ right.replaceWith('')
+
+ # The tag is still connected to the tree.
+ self.assertEqual("three", soup.b.string)
+
+ def test_replace_final_node(self):
+ soup = self.soup("Argh!")
+ soup.find(text="Argh!").replace_with("Hooray!")
+ new_text = soup.find(text="Hooray!")
+ b = soup.b
+ self.assertEqual(new_text.previous_element, b)
+ self.assertEqual(new_text.parent, b)
+ self.assertEqual(new_text.previous_element.next_element, new_text)
+ self.assertEqual(new_text.next_element, None)
+
+ def test_consecutive_text_nodes(self):
+ # A builder should never create two consecutive text nodes,
+ # but if you insert one next to another, Beautiful Soup will
+ # handle it correctly.
+ soup = self.soup("Argh!")
+ soup.b.insert(1, "Hooray!")
+
+ self.assertEqual(
+ soup.decode(), self.document_for(
+ "Argh!Hooray!"))
+
+ new_text = soup.find(text="Hooray!")
+ self.assertEqual(new_text.previous_element, "Argh!")
+ self.assertEqual(new_text.previous_element.next_element, new_text)
+
+ self.assertEqual(new_text.previous_sibling, "Argh!")
+ self.assertEqual(new_text.previous_sibling.next_sibling, new_text)
+
+ self.assertEqual(new_text.next_sibling, None)
+ self.assertEqual(new_text.next_element, soup.c)
+
+ def test_insert_string(self):
+ soup = self.soup("")
+ soup.a.insert(0, "bar")
+ soup.a.insert(0, "foo")
+ # The string were added to the tag.
+ self.assertEqual(["foo", "bar"], soup.a.contents)
+ # And they were converted to NavigableStrings.
+ self.assertEqual(soup.a.contents[0].next_element, "bar")
+
+ def test_insert_tag(self):
+ builder = self.default_builder
+ soup = self.soup(
+ "Findlady!", builder=builder)
+ magic_tag = Tag(soup, builder, 'magictag')
+ magic_tag.insert(0, "the")
+ soup.a.insert(1, magic_tag)
+
+ self.assertEqual(
+ soup.decode(), self.document_for(
+ "Findthelady!"))
+
+ # Make sure all the relationships are hooked up correctly.
+ b_tag = soup.b
+ self.assertEqual(b_tag.next_sibling, magic_tag)
+ self.assertEqual(magic_tag.previous_sibling, b_tag)
+
+ find = b_tag.find(text="Find")
+ self.assertEqual(find.next_element, magic_tag)
+ self.assertEqual(magic_tag.previous_element, find)
+
+ c_tag = soup.c
+ self.assertEqual(magic_tag.next_sibling, c_tag)
+ self.assertEqual(c_tag.previous_sibling, magic_tag)
+
+ the = magic_tag.find(text="the")
+ self.assertEqual(the.parent, magic_tag)
+ self.assertEqual(the.next_element, c_tag)
+ self.assertEqual(c_tag.previous_element, the)
+
+ def test_append_child_thats_already_at_the_end(self):
+ data = ""
+ soup = self.soup(data)
+ soup.a.append(soup.b)
+ self.assertEqual(data, soup.decode())
+
+ def test_move_tag_to_beginning_of_parent(self):
+ data = ""
+ soup = self.soup(data)
+ soup.a.insert(0, soup.d)
+ self.assertEqual("", soup.decode())
+
+ def test_insert_works_on_empty_element_tag(self):
+ # This is a little strange, since most HTML parsers don't allow
+ # markup like this to come through. But in general, we don't
+ # know what the parser would or wouldn't have allowed, so
+ # I'm letting this succeed for now.
+ soup = self.soup("
")
+ soup.br.insert(1, "Contents")
+ self.assertEqual(str(soup.br), "
Contents")
+
+ def test_insert_before(self):
+ soup = self.soup("foobar")
+ soup.b.insert_before("BAZ")
+ soup.a.insert_before("QUUX")
+ self.assertEqual(
+ soup.decode(), self.document_for("QUUXfooBAZbar"))
+
+ soup.a.insert_before(soup.b)
+ self.assertEqual(
+ soup.decode(), self.document_for("QUUXbarfooBAZ"))
+
+ def test_insert_after(self):
+ soup = self.soup("foobar")
+ soup.b.insert_after("BAZ")
+ soup.a.insert_after("QUUX")
+ self.assertEqual(
+ soup.decode(), self.document_for("fooQUUXbarBAZ"))
+ soup.b.insert_after(soup.a)
+ self.assertEqual(
+ soup.decode(), self.document_for("QUUXbarfooBAZ"))
+
+ def test_insert_after_raises_exception_if_after_has_no_meaning(self):
+ soup = self.soup("")
+ tag = soup.new_tag("a")
+ string = soup.new_string("")
+ self.assertRaises(ValueError, string.insert_after, tag)
+ self.assertRaises(NotImplementedError, soup.insert_after, tag)
+ self.assertRaises(ValueError, tag.insert_after, tag)
+
+ def test_insert_before_raises_notimplementederror_if_before_has_no_meaning(self):
+ soup = self.soup("")
+ tag = soup.new_tag("a")
+ string = soup.new_string("")
+ self.assertRaises(ValueError, string.insert_before, tag)
+ self.assertRaises(NotImplementedError, soup.insert_before, tag)
+ self.assertRaises(ValueError, tag.insert_before, tag)
+
+ def test_replace_with(self):
+ soup = self.soup(
+ "There's no business like show business ")
+ no, show = soup.find_all('b')
+ show.replace_with(no)
+ self.assertEqual(
+ soup.decode(),
+ self.document_for(
+ "There's business like no business "))
+
+ self.assertEqual(show.parent, None)
+ self.assertEqual(no.parent, soup.p)
+ self.assertEqual(no.next_element, "no")
+ self.assertEqual(no.next_sibling, " business")
+
+ def test_replace_first_child(self):
+ data = ""
+ soup = self.soup(data)
+ soup.b.replace_with(soup.c)
+ self.assertEqual("", soup.decode())
+
+ def test_replace_last_child(self):
+ data = ""
+ soup = self.soup(data)
+ soup.c.replace_with(soup.b)
+ self.assertEqual("", soup.decode())
+
+ def test_nested_tag_replace_with(self):
+ soup = self.soup(
+ """Wereservetherighttorefuseservice""")
+
+ # Replace the entire tag and its contents ("reserve the
+ # right") with the tag ("refuse").
+ remove_tag = soup.b
+ move_tag = soup.f
+ remove_tag.replace_with(move_tag)
+
+ self.assertEqual(
+ soup.decode(), self.document_for(
+ "Werefusetoservice"))
+
+ # The tag is now an orphan.
+ self.assertEqual(remove_tag.parent, None)
+ self.assertEqual(remove_tag.find(text="right").next_element, None)
+ self.assertEqual(remove_tag.previous_element, None)
+ self.assertEqual(remove_tag.next_sibling, None)
+ self.assertEqual(remove_tag.previous_sibling, None)
+
+ # The tag is now connected to the tag.
+ self.assertEqual(move_tag.parent, soup.a)
+ self.assertEqual(move_tag.previous_element, "We")
+ self.assertEqual(move_tag.next_element.next_element, soup.e)
+ self.assertEqual(move_tag.next_sibling, None)
+
+ # The gap where the tag used to be has been mended, and
+ # the word "to" is now connected to the tag.
+ to_text = soup.find(text="to")
+ g_tag = soup.g
+ self.assertEqual(to_text.next_element, g_tag)
+ self.assertEqual(to_text.next_sibling, g_tag)
+ self.assertEqual(g_tag.previous_element, to_text)
+ self.assertEqual(g_tag.previous_sibling, to_text)
+
+ def test_unwrap(self):
+ tree = self.soup("""
+ Unneeded formatting is unneeded
+ """)
+ tree.em.unwrap()
+ self.assertEqual(tree.em, None)
+ self.assertEqual(tree.p.text, "Unneeded formatting is unneeded")
+
+ def test_wrap(self):
+ soup = self.soup("I wish I was bold.")
+ value = soup.string.wrap(soup.new_tag("b"))
+ self.assertEqual(value.decode(), "I wish I was bold.")
+ self.assertEqual(
+ soup.decode(), self.document_for("I wish I was bold."))
+
+ def test_wrap_extracts_tag_from_elsewhere(self):
+ soup = self.soup("I wish I was bold.")
+ soup.b.next_sibling.wrap(soup.b)
+ self.assertEqual(
+ soup.decode(), self.document_for("I wish I was bold."))
+
+ def test_wrap_puts_new_contents_at_the_end(self):
+ soup = self.soup("I like being bold.I wish I was bold.")
+ soup.b.next_sibling.wrap(soup.b)
+ self.assertEqual(2, len(soup.b.contents))
+ self.assertEqual(
+ soup.decode(), self.document_for(
+ "I like being bold.I wish I was bold."))
+
+ def test_extract(self):
+ soup = self.soup(
+ 'Some content. Nav crap More content.')
+
+ self.assertEqual(len(soup.body.contents), 3)
+ extracted = soup.find(id="nav").extract()
+
+ self.assertEqual(
+ soup.decode(), "Some content. More content.")
+ self.assertEqual(extracted.decode(), 'Nav crap ')
+
+ # The extracted tag is now an orphan.
+ self.assertEqual(len(soup.body.contents), 2)
+ self.assertEqual(extracted.parent, None)
+ self.assertEqual(extracted.previous_element, None)
+ self.assertEqual(extracted.next_element.next_element, None)
+
+ # The gap where the extracted tag used to be has been mended.
+ content_1 = soup.find(text="Some content. ")
+ content_2 = soup.find(text=" More content.")
+ self.assertEqual(content_1.next_element, content_2)
+ self.assertEqual(content_1.next_sibling, content_2)
+ self.assertEqual(content_2.previous_element, content_1)
+ self.assertEqual(content_2.previous_sibling, content_1)
+
+ def test_extract_distinguishes_between_identical_strings(self):
+ soup = self.soup("foobar")
+ foo_1 = soup.a.string
+ bar_1 = soup.b.string
+ foo_2 = soup.new_string("foo")
+ bar_2 = soup.new_string("bar")
+ soup.a.append(foo_2)
+ soup.b.append(bar_2)
+
+ # Now there are two identical strings in the tag, and two
+ # in the tag. Let's remove the first "foo" and the second
+ # "bar".
+ foo_1.extract()
+ bar_2.extract()
+ self.assertEqual(foo_2, soup.a.string)
+ self.assertEqual(bar_2, soup.b.string)
+
+ def test_extract_multiples_of_same_tag(self):
+ soup = self.soup("""
+
+
+
+
+
+
+
+
+
+""")
+ [soup.script.extract() for i in soup.find_all("script")]
+ self.assertEqual("\n\n\n", unicode(soup.body))
+
+
+ def test_extract_works_when_element_is_surrounded_by_identical_strings(self):
+ soup = self.soup(
+ '\n'
+ 'hi\n'
+ '')
+ soup.find('body').extract()
+ self.assertEqual(None, soup.find('body'))
+
+
+ def test_clear(self):
+ """Tag.clear()"""
+ soup = self.soup("String Italicized and another ")
+ # clear using extract()
+ a = soup.a
+ soup.p.clear()
+ self.assertEqual(len(soup.p.contents), 0)
+ self.assertTrue(hasattr(a, "contents"))
+
+ # clear using decompose()
+ em = a.em
+ a.clear(decompose=True)
+ self.assertEqual(0, len(em.contents))
+
+ def test_string_set(self):
+ """Tag.string = 'string'"""
+ soup = self.soup(" ")
+ soup.a.string = "foo"
+ self.assertEqual(soup.a.contents, ["foo"])
+ soup.b.string = "bar"
+ self.assertEqual(soup.b.contents, ["bar"])
+
+ def test_string_set_does_not_affect_original_string(self):
+ soup = self.soup("foobar")
+ soup.b.string = soup.c.string
+ self.assertEqual(soup.a.encode(), b"barbar")
+
+ def test_set_string_preserves_class_of_string(self):
+ soup = self.soup("")
+ cdata = CData("foo")
+ soup.a.string = cdata
+ self.assertTrue(isinstance(soup.a.string, CData))
+
+class TestElementObjects(SoupTest):
+ """Test various features of element objects."""
+
+ def test_len(self):
+ """The length of an element is its number of children."""
+ soup = self.soup("123")
+
+ # The BeautifulSoup object itself contains one element: the
+ # tag.
+ self.assertEqual(len(soup.contents), 1)
+ self.assertEqual(len(soup), 1)
+
+ # The tag contains three elements: the text node "1", the
+ # tag, and the text node "3".
+ self.assertEqual(len(soup.top), 3)
+ self.assertEqual(len(soup.top.contents), 3)
+
+ def test_member_access_invokes_find(self):
+ """Accessing a Python member .foo invokes find('foo')"""
+ soup = self.soup('')
+ self.assertEqual(soup.b, soup.find('b'))
+ self.assertEqual(soup.b.i, soup.find('b').find('i'))
+ self.assertEqual(soup.a, None)
+
+ def test_deprecated_member_access(self):
+ soup = self.soup('')
+ with warnings.catch_warnings(record=True) as w:
+ tag = soup.bTag
+ self.assertEqual(soup.b, tag)
+ self.assertEqual(
+ '.bTag is deprecated, use .find("b") instead.',
+ str(w[0].message))
+
+ def test_has_attr(self):
+ """has_attr() checks for the presence of an attribute.
+
+ Please note note: has_attr() is different from
+ __in__. has_attr() checks the tag's attributes and __in__
+ checks the tag's chidlren.
+ """
+ soup = self.soup("")
+ self.assertTrue(soup.foo.has_attr('attr'))
+ self.assertFalse(soup.foo.has_attr('attr2'))
+
+
+ def test_attributes_come_out_in_alphabetical_order(self):
+ markup = ''
+ self.assertSoupEquals(markup, '')
+
+ def test_string(self):
+ # A tag that contains only a text node makes that node
+ # available as .string.
+ soup = self.soup("foo")
+ self.assertEqual(soup.b.string, 'foo')
+
+ def test_empty_tag_has_no_string(self):
+ # A tag with no children has no .stirng.
+ soup = self.soup("")
+ self.assertEqual(soup.b.string, None)
+
+ def test_tag_with_multiple_children_has_no_string(self):
+ # A tag with no children has no .string.
+ soup = self.soup("foo")
+ self.assertEqual(soup.b.string, None)
+
+ soup = self.soup("foobar")
+ self.assertEqual(soup.b.string, None)
+
+ # Even if all the children are strings, due to trickery,
+ # it won't work--but this would be a good optimization.
+ soup = self.soup("foo")
+ soup.a.insert(1, "bar")
+ self.assertEqual(soup.a.string, None)
+
+ def test_tag_with_recursive_string_has_string(self):
+ # A tag with a single child which has a .string inherits that
+ # .string.
+ soup = self.soup("foo")
+ self.assertEqual(soup.a.string, "foo")
+ self.assertEqual(soup.string, "foo")
+
+ def test_lack_of_string(self):
+ """Only a tag containing a single text node has a .string."""
+ soup = self.soup("feo")
+ self.assertFalse(soup.b.string)
+
+ soup = self.soup("")
+ self.assertFalse(soup.b.string)
+
+ def test_all_text(self):
+ """Tag.text and Tag.get_text(sep=u"") -> all child text, concatenated"""
+ soup = self.soup("ar t ")
+ self.assertEqual(soup.a.text, "ar t ")
+ self.assertEqual(soup.a.get_text(strip=True), "art")
+ self.assertEqual(soup.a.get_text(","), "a,r, , t ")
+ self.assertEqual(soup.a.get_text(",", strip=True), "a,r,t")
+
+ def test_get_text_ignores_comments(self):
+ soup = self.soup("foobar")
+ self.assertEqual(soup.get_text(), "foobar")
+
+ self.assertEqual(
+ soup.get_text(types=(NavigableString, Comment)), "fooIGNOREbar")
+ self.assertEqual(
+ soup.get_text(types=None), "fooIGNOREbar")
+
+ def test_all_strings_ignores_comments(self):
+ soup = self.soup("foobar")
+ self.assertEqual(['foo', 'bar'], list(soup.strings))
+
+class TestCDAtaListAttributes(SoupTest):
+
+ """Testing cdata-list attributes like 'class'.
+ """
+ def test_single_value_becomes_list(self):
+ soup = self.soup("")
+ self.assertEqual(["foo"],soup.a['class'])
+
+ def test_multiple_values_becomes_list(self):
+ soup = self.soup("")
+ self.assertEqual(["foo", "bar"], soup.a['class'])
+
+ def test_multiple_values_separated_by_weird_whitespace(self):
+ soup = self.soup("")
+ self.assertEqual(["foo", "bar", "baz"],soup.a['class'])
+
+ def test_attributes_joined_into_string_on_output(self):
+ soup = self.soup("")
+ self.assertEqual(b'', soup.a.encode())
+
+ def test_accept_charset(self):
+ soup = self.soup(' |