)f[< ~ d dl Z d dlZd dlZd dlmZ d dlmZmZmZ G d de Z G d de Z
G d d e Z y)
N)
BeautifulSoup)EntitySubstitutionEncodingDetector
UnicodeDammitc e Zd ZdZd Zej j dg d d Zd Z d Z
d Z d Z d
Z
d Zy )
TestUnicodeDammitz"Standalone tests of UnicodeDammit.c @ d}t | }|j |k( sJ y )Nu I'm already Unicode! ☃)r unicode_markup)selfmarkupdammits R/var/www/html/flask-app/venv/lib/python3.12/site-packages/bs4/tests/test_dammit.pytest_unicode_inputz$TestUnicodeDammit.test_unicode_input s% 3 v&$$... z smart_quotes_to,expect_converted))Nu ‘’“”)xmlz ‘’“”)htmlz ‘’“”)asciiz''""c f d}t |dg| j }|dj | k( sJ y)zbVerify the functionality of the smart_quotes_to argument
to the UnicodeDammit constructor.s windows-1252)known_definite_encodingssmart_quotes_toz
{}N)r r
format)r r expect_convertedr converteds r test_smart_quotes_toz&TestUnicodeDammit.test_smart_quotes_to sD 0! n-= +
. O 2 23C DDDDr c ~ d}t | }|j j dk( sJ |j dk( sJ y )Ns Sacré bleu! ☃utf-8u Sacré bleu! ☃r original_encodinglowerr
)r utf8r
s r test_detect_utf8z"TestUnicodeDammit.test_detect_utf8& s@ 1 t$''--/7::$$(DDDDr c d}t |dg }|j j dk( sJ |j dk( sJ y )N
iso-8859-8u םולשr )r hebrewr
s r test_convert_hebrewz%TestUnicodeDammit.test_convert_hebrew, sD $ v ~6''--/<??$$(BBBBr c d}t | }|j j dk( sJ |j j d |k( sJ y )Ns ケータイ Watchr )r r r r
encode)r utf_8r
s r /test_dont_see_smart_quotes_where_there_are_nonezATestUnicodeDammit.test_dont_see_smart_quotes_where_there_are_none2 sJ I u%''--/7::$$++G4===r c ~ dj d }t |dg }|j j dk( sJ y )N
Räksmörgåsr r% r) r r r r utf8_datar
s r test_ignore_inappropriate_codecsz2TestUnicodeDammit.test_ignore_inappropriate_codecs8 s= #**73 y<.9''--/7:::r c dj d }dD ]. }t ||g }|j j dk( r.J y )Nr- r )z.utf8z...z
utF---16.!r. )r r0 bad_encodingr
s r test_ignore_invalid_codecsz,TestUnicodeDammit.test_ignore_invalid_codecs= sI #**73 : ?L"9|n=F++113w> > ?r c dj d }t |dg }|j j dk( sJ t |ddg }|j d k( sJ y )Nr- r )exclude_encodingsr r. r/ s r test_exclude_encodingsz(TestUnicodeDammit.test_exclude_encodingsC sg #**73 yWIF''--/>AA '>)BD''4///r N)__name__
__module__ __qualname____doc__r pytestmark parametrizer r" r' r+ r1 r4 r7 r r r r sW ,/
[[ *
EEE C > ;
? 0r r c B e Zd Zd Zd Zd Zd Zd Zd Zd Z d Z
d Z y
) TestEncodingDetectorc P t d }t |j }d|v sJ y )Ns' u utf-�)r list encodings)r detectedrD s r Ptest_encoding_detector_replaces_junk_in_encoding_name_with_replacement_characterzeTestEncodingDetector.test_encoding_detector_replaces_junk_in_encoding_name_with_replacement_characterS s- # 9;++, .);;;r c N dD ] }t |d }d|j k( r J y )N)s& s& s$ s# Tis_htmlzeuc-jp)r r r datar
s r test_detect_html5_style_meta_tagz5TestEncodingDetector.test_detect_html5_style_meta_tagY s2 4 8D
#46Fv 7 77 7
8r c , d}t j j }t j t j
d }|t j _ t
| }d|j k( sJ d|j v sJ t |d }|j sJ t j t j |t j _ y # t j t j |t j _ w xY w)NsT
بتر
ѐc y Nr? )strs r noopzETestEncodingDetector.test_last_ditch_entity_replacement..noopy s r Tu �z html.parser) bs4r
chardet_dammitloggingdisableWARNINGr contains_replacement_charactersr
r NOTSET)r docchardetrQ r
soups r "test_last_ditch_entity_replacementz7TestEncodingDetector.test_last_ditch_entity_replacementc s 2 **++( 0
(,CJJ %"3'F6AAA Av 4 44 4 m4D77 77 OOGNN +(/CJJ %
OOGNN +(/CJJ %s
A C :Dc b d}t | }d|j k( sJ d|j k( sJ y )N < a > < / a > u áéutf-16le)r r
r rJ s r test_byte_order_mark_removedz1TestEncodingDetector.test_byte_order_mark_removed s6 M t$ 5 555V 5 5555r c d}t | }t |dg }d|j k( sJ t |dg }d|j k( sJ dg|j D cg c] }|d c}k( sJ d}t |dgd g
}d |j k( sJ dd g|j D cg c] }|d c}k( sJ y c c}w c c}w ) Nr^ zutf-16)r r )user_encodingsr_ r r$ r% )r rb r r tried_encodings)r rK r
beforeafterxr& s r )test_known_definite_versus_user_encodingsz>TestEncodingDetector.test_known_definite_versus_user_encodings s
N t$ txjI63333
dG9=U 4 444|f.D.D E! EEE % v /;n> v 7 777 &9O9O*PA1Q4*PPPP F +Qs B7# B<c d}t |dgdgdg }d|j k( sJ g d|j D cg c] }|d c}k( sJ y c c}w )Nr$ shift-jisr r% )r override_encodingsrb )rj r r% r rc )r r& r
rg s r "test_deprecated_override_encodingsz7TestEncodingDetector.test_deprecated_override_encodings sg % &1] 'y (>
v 7 777 4!11 2aQqT 2
2s A c , dj d }dj d }||z |z }t j t 5 |j d d d d t j | }d|j d k( sJ y # 1 sw Y 5xY w)Nu ☃☃☃r! u “Hi, I like Windows!” windows_1252u+ ☃☃☃“Hi, I like Windows!”☃☃☃)r) r< raisesUnicodeDecodeErrordecoder detwingle)r r! rn rY fixeds r test_detwinglez#TestEncodingDetector.test_detwingle s !))&1
./5vn/E
\!D( ]]-
. JJv '',< V@TTTT s B
Bc dD ]B }|j d }|j d sJ t j | }||k( rBJ y )N)u œu ₓu ðr! )r) endswithr rr )r tricky_unicode_charinputoutputs r +test_detwingle_ignores_multibyte_charactersz@TestEncodingDetector.test_detwingle_ignores_multibyte_characters sO
$ #
(..v6E>>'* *",,U3FU? " #r c d}|j d }d}|j d }t j } ||d J d ||d k( sJ d ||d k( sJ d || k( sJ d || k( sJ d d
z } |||z J |||z J |||z dd dk( sJ ||d dk( sJ |d |z d dk( sJ |d
|z d J y )Nz0r z,FrH r Tz
iso-8859-1 i )rI search_entire_document)r~ a)r) r find_declared_encoding)r html_unicode
html_bytes xml_unicode xml_bytesmspacers r test_find_declared_encodingz0TestEncodingDetector.test_find_declared_encoding s J !((1
C &&w/ 3 3 u-55!L$777!J555q~--q|++ *$%--)#$,,
fz!4 M
48LHH !$?<OO !$?GGGr N) r8 r9 r: rF rL r\ r` rh rl rt r{ r r? r r rA rA Q s3 < 8!0F6 Q>
$U.
# (Hr rA c e Zd ZdZd Zej j dddg d Zd Z d Z
d Z d
Z d Z
d Zd
Zd Zd Zd Zd Zd Zy)TestEntitySubstitutionz1Standalone tests of the EntitySubstitution class.c t | _ y rO )r subr s r setup_methodz#TestEntitySubstitution.setup_method s %r zoriginal,substituted)u foo∀☃õbaru foo∀☃õbar)u ‘’foo“”z ‘’foo“”c D | j j | |k( sJ y rO r substitute_html)r original substituteds r test_substitute_htmlz+TestEntitySubstitution.test_substitute_html s! xx''1[@@@r c p dD ]1 \ }}d}||z }||z }| j j | |k( r1J y )N)
)z⊧u ⊧)z𝔑u 𝔑)z≧̸u ≧̸)z¬ ¬)z⫬u ⫬z||)fjr )z>>)z<<)z&&z3 %s 4r )r entityutemplateraw
with_entitiess r test_html5_entityz(TestEntitySubstitution.test_html5_entity% sM
BIFA0 HQ,C $v -M88++C0MA A7 Br c d}d}| j j | |k( sJ d}d}| j j | |k( sJ y )Nu fjords ⊔ penguinszfjords ⊔ penguinsu fjords ⊔︀ penguinszfjords ⊔︀ penguinsr )r rK r s r )test_html5_entity_with_variation_selectorz@TestEntitySubstitution.test_html5_entity_with_variation_selectorC sO (*xx''-77-+xx''-777r c J d}| j j |d |k( sJ y )NWelcome to "my bar"Fr substitute_xmlr ss r Itest_xml_converstion_includes_no_quotes_if_make_quoted_attribute_is_falsez`TestEntitySubstitution.test_xml_converstion_includes_no_quotes_if_make_quoted_attribute_is_falseO s& !xx&&q%0A555r c | j j dd dk( sJ | j j dd dk( sJ y )NWelcomeTz "Welcome"z Bob's Barz "Bob's Bar"r r s r 6test_xml_attribute_quoting_normally_uses_double_quoteszMTestEntitySubstitution.test_xml_attribute_quoting_normally_uses_double_quotesS s= xx&&y$7;FFxx&&{D9^KKKr c J d}| j j |d dk( sJ y )Nr Tz'Welcome to "my bar"'r r s r Otest_xml_attribute_quoting_uses_single_quotes_when_value_contains_double_quoteszfTestEntitySubstitution.test_xml_attribute_quoting_uses_single_quotes_when_value_contains_double_quotesW s' !xx&&q$/3LLLLr c J d}| j j |d dk( sJ y )NWelcome to "Bob's Bar"Tz""Welcome to "Bob's Bar""r r s r btest_xml_attribute_quoting_escapes_single_quotes_when_value_contains_both_single_and_double_quoteszyTestEntitySubstitution.test_xml_attribute_quoting_escapes_single_quotes_when_value_contains_both_single_and_double_quotes[ s' %xx&&q$/3XXXXr c H d}| j j | |k( sJ y )Nr r )r quoteds r zfoo<bar>r r s r 'test_xml_quoting_handles_angle_bracketsz>TestEntitySubstitution.test_xml_quoting_handles_angle_bracketsc s xx&&z26FFFFr c D | j j d dk( sJ y )NzAT&TzAT&Tr r s r #test_xml_quoting_handles_ampersandsz:TestEntitySubstitution.test_xml_quoting_handles_ampersandsf s xx&&v.*<< r r r r r r r r r r r r r r? r r r r s ;& [[
2
C
A A B<
86LMY9G=O_6r r ) r< rT rR r
bs4.dammitr r r objectr rA r r? r r r sI
C0 C0J{H6 {H|c6V c6r