U
|
¡ý°d ã @ sh d Z ddlZddlZddlmZ ddlmZ ddlmZm Z m
|
Z
|
ej j e ddG d d
|
d
|
e
|
e Z dS ) zDTests to ensure that the html5lib tree builder generates good trees.é N)Ú BeautifulSoup)Ú SoupStraineré )ÚHTML5LIB_PRESENTÚHTML5TreeBuilderSmokeTestÚSoupTestz?html5lib seems not to be present, not testing its tree builder.)Úreasonc @ s e Zd ZdZedd Zdd Zdd Zdd Zd
|
d Z d d Z
|
dd Z dd Z dd Z dd Zdd Zdd Zdd Zd d Zd d Zd d! Zd"S )#ÚTestHTML5LibBuilderz"See ``HTML5TreeBuilderSmokeTest``.c C s ddl m} |S )Nr )ÚHTML5TreeBuilder)Z bs4.builderr
|
)Úselfr
|
© r úNd:\z\workplace\vscode\pyvenv\venv\Lib\site-packages\bs4/tests/test_html5lib.pyÚdefault_builder s z#TestHTML5LibBuilder.default_builderc C sr t d}d}tjdd}t|d|d}W 5 Q R X | ¡ | |¡ksHt|\}|jtks\tdt |j
|
ksntd S )NÚbz <p>A <b>bold</b> statement.</p>T)ÚrecordÚhtml5lib)Z
|
parse_onlyz4the html5lib tree builder doesn't support parse_only) r ÚwarningsÚcatch_warningsr ÚdecodeZ document_forÚAssertionErrorÚfilenameÚ__file__ÚstrÚmessage)r ZstrainerÚmarkupÚwÚsoupÚwarningr r r Útest_soupstrainer s z%TestHTML5LibBuilder.test_soupstrainerc C s d}| |d¡ | d¡ dS )z8html5lib inserts <tbody> tags where other parsers don't.z[<table id="1"><tr><td>Here's another table:<table id="2"><tr><td>foo</td></tr></table></td>z<table id="1"><tbody><tr><td>Here's another table:<table id="2"><tbody><tr><td>foo</td></tr></tbody></table></td></tr></tbody></table>z{<table><thead><tr><td>Foo</td></tr></thead><tbody><tr><td>Bar</td></tr></tbody><tfoot><tr><td>Baz</td></tr></tfoot></table>N)Z assert_soup)r r r r r Ú test_correctly_nested_tables& s þÿz0TestHTML5LibBuilder.test_correctly_nested_tablesc C s$ d}| |¡}d|j ¡ ks td S )Nzy<?xml version="1.0" encoding="utf-8"?>
|
<!DOCTYPE html>
|
<html>
|
<head>
|
</head>
|
<body>
|
<p>foo</p>
|
</body>
|
</html>s
|
<p>foo</p>)r ÚpÚencoder ©r r r r r r Ú(test_xml_declaration_followed_by_doctype: s
|
z<TestHTML5LibBuilder.test_xml_declaration_followed_by_doctypec C s: d}| |¡}d|j ¡ ks tdt| d¡ks6td S )Nz%<p><em>foo</p>
|
<p>bar<a></a></em></p>zD<body><p><em>foo</em></p><em>
|
</em><p><em>bar<a></a></em></p></body>é r ©r Úbodyr r ÚlenÚfind_allr" r r r Útest_reparented_markupH s
|
z*TestHTML5LibBuilder.test_reparented_markupc C s: d}| |¡}d|j ¡ ks tdt| d¡ks6td S )Nz&<p><em>foo</p>
|
<p>bar<a></a></em></p>
|
zE<body><p><em>foo</em></p><em>
|
</em><p><em>bar<a></a></em></p>
|
</body>r$ r r% r" r r r Ú+test_reparented_markup_ends_with_whitespaceO s
|
z?TestHTML5LibBuilder.test_reparented_markup_ends_with_whitespacec C sL d}| |¡}|jdd\}}| d¡\}}|j|ks:t|j|ksHtdS )zVerify that we keep the two whitespace nodes in this
|
document distinct when reparenting the adjacent <tbody> tags.
|
z,<table> <tbody><tbody><ims></tbody> </table>ú ©ÚstringÚtbodyN)r r( Ú next_elementr )r r r Zspace1Zspace2Ztbody1Ztbody2r r r Ú<test_reparented_markup_containing_identical_whitespace_nodesU s
|
zPTestHTML5LibBuilder.test_reparented_markup_containing_identical_whitespace_nodesc C s^ d}| |¡}|j}d|jks"t|jdd}|jddd }||jksLt||jksZtd S )NzF<div><a>aftermath<p><noscript>target</noscript>aftermath</a></p></div>Útargetr, Z aftermathéÿÿÿÿ)r Únoscriptr/ r Úfindr( Zprevious_element)r r r r3 r1 Zfinal_aftermathr r r Ú*test_reparented_markup_containing_children` s
|
z>TestHTML5LibBuilder.test_reparented_markup_containing_childrenc C s$ d}| |¡}t| d¡s tdS )z(Processing instructions become comments.s <?PITarget PIContent?>z<!--?PITarget PIContent?-->N)r r Ú
|
startswithr r" r r r Útest_processing_instructionp s
|
z/TestHTML5LibBuilder.test_processing_instructionc C s8 d}| |¡}| d¡\}}||ks(t||k s4td S )Ns <a class="my_class"><p></a>Úa)r r( r )r r r Za1Za2r r r Útest_cloned_multivalue_nodev s
|
|
z/TestHTML5LibBuilder.test_cloned_multivalue_nodec C s$ d}| |¡}d|j ¡ ks td S )Ns <table><td></tbody>Az><body>A<table><tbody><tr><td></td></tr></tbody></table></body>)r r& r r r" r r r Útest_foster_parenting} s
|
z)TestHTML5LibBuilder.test_foster_parentingc C sL d}| |¡}dd |dD dd |dD t| d¡dksHtd S )
|
z
|
Test that extraction does not destroy the tree.
|
|
https://bugs.launchpad.net/beautifulsoup/+bug/1782928
|
zW
|
<html><head></head>
|
<style>
|
</style><script></script><body><p>hello</p></body></html>
|
c S s g | ] }| ¡ qS r ©Úextract©Ú.0Úsr r r Ú
|
<listcomp> s z7TestHTML5LibBuilder.test_extraction.<locals>.<listcomp>Úscriptc S s g | ] }| ¡ qS r r; r= r r r r@ s Ústyler r N)r r' r( r r" r r r Útest_extraction s
|
|
z#TestHTML5LibBuilder.test_extractionc C sF d}| |¡}g }| d¡D ]}| | d¡¡ q t|dksBtdS )z
|
Test that empty comment does not break structure.
|
|
https://bugs.launchpad.net/beautifulsoup/+bug/1806598
|
zI
|
<html>
|
<body>
|
<form>
|
<!----><input type="text">
|
</form>
|
</body>
|
</html>
|
ÚformÚinputr N)r r( Úextendr' r )r r r ÚinputsrD r r r Útest_empty_comment s
|
z&TestHTML5LibBuilder.test_empty_commentc C sz d}| |¡}d|jjks td|jjks.td|j d¡jksDt| j |dd}d|jjjksdtd|jjjksvtd S )Nz=
|
<p>
|
|
<sourceline>
|
<b>text</b></sourceline><sourcepos></p>r$ é Ú
|
sourcelineF)Zstore_line_numbersÚ sourcepos)r r rJ r rK r4 Únamer" r r r Útest_tracking_line_numbersª s
|
z.TestHTML5LibBuilder.test_tracking_line_numbersc C s d S )Nr )r r r r Ú test_special_string_containers¸ s z2TestHTML5LibBuilder.test_special_string_containersc C sj dD ]`\}}}d| }| |¡j}| ¡ }d| d¡ }||ksDt|jdd}d| }||kstqd S )N))z⇄u âs ⇄)z⊧u â§s ⊧)z𝔑u ðs 𝔑)z≧̸u â§Ì¸s ≧̸)z¬õ ¬s ¬)z⫬u ⫬s ⫬)z"ú"ó ")z∴õ â´ó ∴)z ∴rR rS )z ∴rR rS )zfjÚfjs fj)z⊔u âs ⊔)z⊔︀u âï¸s ⊔︀)z'ú'ó ')z|ú|ó |z <div>%s</div>s <div>%s</div>Úutf8Úhtml)Ú formatter)r Údivr! r ) r Z input_elementZoutput_unicodeZoutput_elementr r\ Zwithout_elementÚexpectZ with_elementr r r Útest_html5_attributes¾ s z)TestHTML5LibBuilder.test_html5_attributesN)Ú__name__Ú
|
__module__Ú __qualname__Ú__doc__Úpropertyr r r r# r) r* r0 r5 r7 r9 r: rC rH rM rN r^ r r r r r s$
|
r )rb Zpytestr Zbs4r Z bs4.elementr Ú r r r ÚmarkZskipifr r r r r Ú<module> s þ
|
