# -*- coding: utf-8 -*-
"""Tests of Beautiful Soup as a whole."""

import unittest
from bs4 import BeautifulSoup
from bs4.element import SoupStrainer
from bs4.dammit import EntitySubstitution, UnicodeDammit
from bs4.testing import SoupTest
import warnings

class TestDeprecatedConstructorArguments(SoupTest):

    def test_parseOnlyThese_renamed_to_parse_only(self):
        with warnings.catch_warnings(record=True) as w:
            soup = self.soup("<a><b></b></a>", parseOnlyThese=SoupStrainer("b"))
        msg = str(w[0].message)
        self.assertTrue("parseOnlyThese" in msg)
        self.assertTrue("parse_only" in msg)
        self.assertEquals(b"<b></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.assertEquals("utf8", soup.original_encoding)

    def test_unrecognized_keyword_argument(self):
        self.assertRaises(
            TypeError, self.soup, "<a>", no_such_argument=True)

class TestSelectiveParsing(SoupTest):

    def test_parse_with_soupstrainer(self):
        markup = "No<b>Yes</b><a>No<b>Yes <c>Yes</c></b>"
        strainer = SoupStrainer("b")
        soup = self.soup(markup, parse_only=strainer)
        self.assertEqual(soup.encode(), b"<b>Yes</b><b>Yes <c>Yes</c></b>")


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&forall;\N{SNOWMAN}&otilde;bar")

    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),
                          "&lsquo;&rsquo;foo&ldquo;&rdquo;")

    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 &quot;Bob\'s Bar&quot;"')

    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<bar>"),
            "foo&lt;bar&gt;")

    def test_xml_quoting_handles_ampersands(self):
        self.assertEqual(self.sub.substitute_xml("AT&T"), "AT&amp;T")

    def test_xml_quoting_ignores_ampersands_when_they_are_part_of_an_entity(self):
        self.assertEqual(
            self.sub.substitute_xml("&Aacute;T&T"),
            "&Aacute;T&amp;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"<html><head></head><body><foo>Sacr\N{LATIN SMALL LETTER E WITH ACUTE} bleu!</foo></body></html>"
        self.utf8_data = self.unicode_data.encode("utf-8")
        # Just so you know what it looks like.
        self.assertEqual(
            self.utf8_data,
            b"<html><head></head><body><foo>Sacr\xc3\xa9 bleu!</foo></body></html>")

    def test_ascii_in_unicode_out(self):
        # ASCII input is converted to Unicode. The original_encoding
        # attribute is set.
        ascii = b"<foo>a</foo>"
        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, "ascii")

    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)


class TestUnicodeDammit(unittest.TestCase):
    """Standalone tests of Unicode, Dammit."""

    def test_smart_quotes_to_unicode(self):
        markup = b"<foo>\x91\x92\x93\x94</foo>"
        dammit = UnicodeDammit(markup)
        self.assertEqual(
            dammit.unicode_markup, u"<foo>\u2018\u2019\u201c\u201d</foo>")

    def test_smart_quotes_to_xml_entities(self):
        markup = b"<foo>\x91\x92\x93\x94</foo>"
        dammit = UnicodeDammit(markup, smart_quotes_to="xml")
        self.assertEqual(
            dammit.unicode_markup, "<foo>&#x2018;&#x2019;&#x201C;&#x201D;</foo>")

    def test_smart_quotes_to_html_entities(self):
        markup = b"<foo>\x91\x92\x93\x94</foo>"
        dammit = UnicodeDammit(markup, smart_quotes_to="html")
        self.assertEqual(
            dammit.unicode_markup, "<foo>&lsquo;&rsquo;&ldquo;&rdquo;</foo>")

    def test_detect_utf8(self):
        utf8 = b"\xc3\xa9"
        dammit = UnicodeDammit(utf8)
        self.assertEqual(dammit.unicode_markup, u'\xe9')
        self.assertEqual(dammit.original_encoding, 'utf-8')

    def test_convert_hebrew(self):
        hebrew = b"\xed\xe5\xec\xf9"
        dammit = UnicodeDammit(hebrew, ["iso-8859-8"])
        self.assertEqual(dammit.original_encoding, '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, '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, '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, 'utf-8')

    def test_detect_html5_style_meta_tag(self):

        for data in (
            b'<html><meta charset="euc-jp" /></html>',
            b"<html><meta charset='euc-jp' /></html>",
            b"<html><meta charset=euc-jp /></html>",
            b"<html><meta charset=euc-jp/></html>"):
            dammit = UnicodeDammit(data, is_html=True)
            self.assertEquals(
                "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 (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.

        doc = b"""\357\273\277<?xml version="1.0" encoding="UTF-8"?>
<html><b>\330\250\330\252\330\261</b>
<i>\310\322\321\220\312\321\355\344</i></html>"""
        with warnings.catch_warnings(record=True) as w:
            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)

            msg = w[0].message
            self.assertTrue(isinstance(msg, UnicodeWarning))
            self.assertTrue("Some characters could not be decoded" in str(msg))