MhE d dl Z d dlZd dlZd dlZd dlmZ d dlmZmZmZ G d de Z
G d de Z G d d e Z dS )
N)
BeautifulSoup)EntitySubstitutionEncodingDetector
UnicodeDammitc e Zd ZdZd Zej dg d d Zd Z d Z
d Z d Z d
Z
d Zd S )
TestUnicodeDammitz"Standalone tests of UnicodeDammit.c B d}t | }|j |k sJ d S )Nu I'm already Unicode! ☃)r unicode_markup)selfmarkupdammits U/var/www/html/test/jupyter/venv/lib/python3.11/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 p d}t |dg| j }|d | k sJ dS )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 sY 0! &4%5 +
O 2 23C D DDDDDDDr c d}t | }|j dk sJ |j dk sJ d S )Ns Sacré bleu! ☃utf-8u Sacré bleu! ☃r original_encodinglowerr
)r utf8r
s r test_detect_utf8z"TestUnicodeDammit.test_detect_utf8* sO 1 t$$'--//7::::$(DDDDDDDr c d}t |dg }|j dk sJ |j dk sJ d S )N
iso-8859-8u םולשr )r hebrewr
s r test_convert_hebrewz%TestUnicodeDammit.test_convert_hebrew0 sS $ v ~66'--//<????$(BBBBBBBr c d}t | }|j dk sJ |j d |k sJ d S )Ns ケータイ Watchr )r r r r
encode)r utf_8r
s r /test_dont_see_smart_quotes_where_there_are_nonezATestUnicodeDammit.test_dont_see_smart_quotes_where_there_are_none6 s\ I u%%'--//7::::$++G44======r c d d }t |dg }|j dk sJ d S )N
Räksmörgåsr r% r) r r r r utf8_datar
s r test_ignore_inappropriate_codecsz2TestUnicodeDammit.test_ignore_inappropriate_codecs< sL #**733 y<.99'--//7::::::r c d d }dD ]2}t ||g }|j dk sJ 3d S )Nr- r )z.utf8z...z
utF---16.!r. )r r0 bad_encodingr
s r test_ignore_invalid_codecsz,TestUnicodeDammit.test_ignore_invalid_codecsA sb #**733 : ? ?L"9|n==F+1133w>>>>> ? ?r c d d }t |dg }|j dk sJ t |ddg }|j J d S )Nr- r )exclude_encodingsr r. r/ s r test_exclude_encodingsz(TestUnicodeDammit.test_exclude_encodingsG sy #**733 yWIFFF'--//>AAAA yWn > > ; ; ;
? ? ? 0 0 0 0 0r r c D e Zd Zd Zd Zd Zd Zd Zd Zd Z d Z
d Z d
S ) TestEncodingDetectorc X t d }t |j }d|v sJ d S )Ns' u utf-�)r list encodings)r detectedrD s r Ptest_encoding_detector_replaces_junk_in_encoding_name_with_replacement_characterzeTestEncodingDetector.test_encoding_detector_replaces_junk_in_encoding_name_with_replacement_characterV s9 $$QRR+,, .);;;;;;r c L dD ] }t |d }d|j k sJ !d S )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_tag] sH
8 8D #4666Fv 777777 8 8r c d}t j j }t j t j d }|t j _ t
| }d|j u sJ d|j v sJ t |d }|j sJ t j t j
|t j _ d S # t j t j
|t j _ w xY w)NsT
بتر
ѐc d S Nr? )strs r noopzETestEncodingDetector.test_last_ditch_entity_replacement..noop~ s tr 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_replacementg s 2 *,((( 1
*.CJ &"3''F6AAAAAv 44444 m44D7 7 7 7 7 OGN + + +)0CJ & & &
OGN + + +)0CJ & 0 0 0 0s AB7 71C(c \ d}t | }d|j k sJ d|j k sJ d S )N < a > < / a > u áéutf-16le)r r
r rJ s r test_byte_order_mark_removedz1TestEncodingDetector.test_byte_order_mark_removed sB M t$$ 55555V 5555555r c L d}t | }t |dg }d|j k sJ t |dg }d|j k sJ dgd |j D k sJ d}t |dgd g
}d |j k sJ dd gd |j D k sJ d S ) Nr^ zutf-16)r r )user_encodingsr_ c g | ]
}|d S r r? .0xs r
zRTestEncodingDetector.test_known_definite_versus_user_encodings.. s E E E! E E Er r$ r% )r rb c g | ]
}|d S rd r? re s r rh zRTestEncodingDetector.test_known_definite_versus_user_encodings.. s *P*P*PA1Q4*P*P*Pr )r r tried_encodings)r rK r
beforeafterr& s r )test_known_definite_versus_user_encodingsz>TestEncodingDetector.test_known_definite_versus_user_encodings s
N t$$ txjIII633333
dG9===U 44444| E Ef.D E E EEEEE % gY ~
v 77777 &*P*P9O*P*P*PPPPPPPr c D d}t j d 5 }t |dgdgdg }d d d n # 1 swxY w Y |\ }|j }t |t
sJ |j t k sJ d|j k sJ g dd |j D k sJ d S )
Nr$ T)record shift-jisr r% )r override_encodingsrb )rp r r% c g | ]
}|d S rd r? re s r rh zKTestEncodingDetector.test_deprecated_override_encodings.. s 2 2 2aQqT 2 2 2r )
warningscatch_warningsr message
isinstanceDeprecationWarningfilename__file__r rj )r r& wr
warningru s r "test_deprecated_override_encodingsz7TestEncodingDetector.test_deprecated_override_encodings s! %
$D
1
1
1 Q"*5 $+9 ,~ F /'#566666 8++++v 77777 433 2 261 2 2 2
s
;??c T d d }d d }||z |z }t j t 5 | d d d d n # 1 swxY w Y t j | }d| d k sJ d S )Nu ☃☃☃r! u “Hi, I like Windows!” windows_1252u+ ☃☃☃“Hi, I like Windows!”☃☃☃)r) r< raisesUnicodeDecodeErrordecoder detwingle)r r! r~ rY fixeds r test_detwinglez#TestEncodingDetector.test_detwingle s !))&11
.
&
\!D( ]-
.
. JJv ',,< V@T@TTTTTTTs A..A25A2c dD ]J}| d }| d sJ t j | }||k sJ Kd S )N)u œu ₓu ðr! )r) endswithr r )r tricky_unicode_charinputoutputs r +test_detwingle_ignores_multibyte_charactersz@TestEncodingDetector.test_detwingle_ignores_multibyte_characters sj
$
# #
(..v66E>>'** * * *",U33FU????? # #r c d}| d }d}| 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 d S )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 !((11
D &&w// 3q u---555!!L$7777777!!J5555555qq~~----qq||++++ q*$%%---q)#$$,,,
Afz!4 M M MQX X X X X q4888LHHHHq !$???<OOOOq !$???GGGGGr N) r8 r9 r: rF rL r\ r` rm r| r r r r? r r rA rA U s < < <8 8 8#1 #1 #1J6 6 6 Q Q Q@
.U U U0
#
#
# 'H 'H 'H 'H 'Hr rA c e Zd ZdZd Zej dddg d Zd Z d Z
d Z d
Z d Z
d Zd
Zd Zd Zd Zd Zd Zej dg d d Zej dg d d Zej ddg d Zd S ) TestEntitySubstitutionz1Standalone tests of the EntitySubstitution class.c t | _ d S rO )r subr s r setup_methodz#TestEntitySubstitution.setup_method s
%r zoriginal,substituted)u foo∀☃õbaru foo∀☃õbar)u ‘’foo“”z ‘’foo“”c F | j | |k sJ d S rO r substitute_html)r original substituteds r test_substitute_htmlz+TestEntitySubstitution.test_substitute_html! s+ x''11[@@@@@@r c n dD ]1\ }}d}||z }||z }| j | |k sJ 2d S )N) )z⊧u ⊧)z𝔑u 𝔑)z≧̸u ≧̸)z¬ ¬)z⫬u ⫬z||)fjr )z>>)z<zfoo<bar>r r s r 'test_xml_quoting_handles_angle_bracketsz>TestEntitySubstitution.test_xml_quoting_handles_angle_bracketsm s* x&&z226FFFFFFFr c F | j d dk sJ d S )NzAT&TzAT&Tr r s r #test_xml_quoting_handles_ampersandsz:TestEntitySubstitution.test_xml_quoting_handles_ampersandsp s) x&&v..*<<<<<TestEntitySubstitution.test_unambiguous_ampersands_not_escaped sL x''//36666x,,V44>>>>>>r zmarkup,html,html5,html5raw))÷ ÷r r )z
&nonesuch;&nonesuch;r r )z÷
÷r r )z¡
¡r r c | j | |k sJ | j | |k sJ | j | |k sJ d S rO )r r substitute_html5r )r r r html5html5raws r 'test_when_entity_ampersands_are_escapedz>TestEntitySubstitution.test_when_entity_ampersands_are_escaped sm , x''//47777x((00E9999x,,V44@@@@@@r z
markup,expect)z&nosuchentity;z&nosuchentity;c | j | |k sJ | j | |k sJ d S rO r )r r expects r !test_ambiguous_ampersands_escapedz8TestEntitySubstitution.test_ambiguous_ampersands_escaped sL x''//69999x,,V44>>>>>>r N)r8 r9 r: r; r r< r= r> r r r r r r r r r r r r r r r r r? r r r r s ;;& & & [
S
C
A A
AB B B4
8
8
86 6 6L L LM M M Y Y Y 9 9 9G G G= = =O O O
6 6 6
[
? ? ? [ $
A A A" [ BC ? ? ? ? ?r r )
r< rT rs rR r
bs4.dammitr r r objectr rA r r? r r r s
D0 D0 D0 D0 D0 D0 D0 D0NCH CH CH CH CH6 CH CH CHLV? V? V? V? V?V V? V? V? V? V?r