File: //usr/local/CyberCP/lib/python3.10/site-packages/bs4/builder/__pycache__/_html5lib.cpython-310.pyc
o
�h�J �
@ s d Z dgZddlZddlZddlmZmZmZmZm Z ddl
m Z m Z ddl
Z
ddlmZmZ ddl
mZmZmZmZ z
ddlmZ d ZW n ey_ Z zdd
lmZ d ZW Y dZ[ndZ[ww G d d� de �Z G d
d� dej �Z G dd� de �Z G dd� dej!�Z"G dd� de"�Z#dS )�MIT�HTML5TreeBuilder� N)�DetectsXMLParsedAsHTML�
PERMISSIVE�HTML�HTML_5�HTMLTreeBuilder)�NamespacedAttribute�nonwhitespace_re)�
namespaces�prefixes)�Comment�Doctype�NavigableString�Tag)�_baseF)�baseTc @ sH e Zd ZdZdZeeeegZdZ d
dd�Z
dd� Z d d
� Z d d � Z
dS )r a Use html5lib to build a tree.
Note that this TreeBuilder does not support some features common
to HTML TreeBuilders. Some of these features could theoretically
be implemented, but at the very least it's quite difficult,
because html5lib moves the parse tree around as it's being built.
* This TreeBuilder doesn't use different subclasses of NavigableString
based on the name of the tag in which the string was found.
* You can't use a SoupStrainer to parse only part of a document.
�html5libTNc c s: � || _ |r
tjddd� tj|dd� |d d dfV d S )NzjYou provided a value for exclude_encoding, but the html5lib tree builder doesn't support exclude_encoding.� ��
stacklevelF)�user_specified_encoding�warnings�warnr � warn_if_markup_looks_like_xml)�self�markupr �document_declared_encoding�exclude_encodings� r �H/usr/local/CyberCP/lib/python3.10/site-packages/bs4/builder/_html5lib.py�prepare_markup@ s ���z HTML5TreeBuilder.prepare_markupc C s� | j jd ur
tjddd� tj| jd�}|| j_t � }t
|t �s-t r(| j
|d< n| j
|d< |j|fi |��}t
|t �r?d |_n|jjjd }t
|t �sN|j}||_d | j_d S )Nz�You provided a value for parse_only, but the html5lib tree builder doesn't support parse_only. The entire document will be parsed.� r )�tree�override_encoding�encodingr )�soup�
parse_onlyr r r �
HTMLParser�create_treebuilder�underlying_builder�parser�dict�
isinstance�str� new_html5libr �parse�original_encoding� tokenizer�stream� charEncoding�name)r r r+ � extra_kwargs�docr1 r r r �feedW s( �
zHTML5TreeBuilder.feedc C s t || j| jd�| _| jS )N)�store_line_numbers)�TreeBuilderForHtml5libr& r9 r* )r �namespaceHTMLElementsr r r r) v s
�z#HTML5TreeBuilder.create_treebuilderc C s d| S )zSee `TreeBuilder`.z)<html><head></head><body>%s</body></html>r )r �fragmentr r r �test_fragment_to_document} s z*HTML5TreeBuilder.test_fragment_to_document)NN)�__name__�
__module__� __qualname__�__doc__�NAMEr r r �features�TRACKS_LINE_NUMBERSr! r8 r) r= r r r r r * s
� c sj e Zd Z d� fdd� Zdd� Zdd� Zd d
� Zd d � Zd
d� Zdd� Z dd� Z
dd� Z dd� Z � Z
S )r: NTc sN |r|| _ nddlm} | dd|i|��| _ tt| ��|� d | _|| _d S )Nr ��
BeautifulSoup� � html.parserr9 )rG rH )r& �bs4rF �superr: �__init__r+ r9 )r r; r& r9 �kwargsrF �� __class__r r rK � s ���
z TreeBuilderForHtml5lib.__init__c C s | j �� t| j | j d �S �N)r& �reset�Element�r r r r �
documentClass� s
z$TreeBuilderForHtml5lib.documentClassc C s6 |d }|d }|d }t �|||�}| j�|� d S )Nr5 �publicId�systemId)r �for_name_and_idsr& �object_was_parsed)r �tokenr5 rT rU �doctyper r r �
insertDoctype� s
z$TreeBuilderForHtml5lib.insertDoctypec C sZ i }| j r| jr| j jj�� \}}||d< |d |d< | jj||fi |��}t|| j|�S )N�
sourceline� � sourcepos)r+ r9 r2 r3 �positionr& �new_tagrQ )r r5 � namespacerL r[ r] �tagr r r � elementClass� s z#TreeBuilderForHtml5lib.elementClassc C s t t|�| j�S rO )�TextNoder
r& )r �datar r r � commentClass� s z#TreeBuilderForHtml5lib.commentClassc C s0 ddl m} |dd�| _d| j_t| j| jd �S )Nr rE rG rH z[document_fragment])rI rF r& r5 rQ )r rF r r r �
fragmentClass� s z$TreeBuilderForHtml5lib.fragmentClassc C s | j �|j� d S rO )r& �append�element�r �noder r r � appendChild� s z"TreeBuilderForHtml5lib.appendChildc C s | j S rO )r& rR r r r � getDocument� s z"TreeBuilderForHtml5lib.getDocumentc C s t j�| �jS rO )�treebuilder_base� TreeBuilder� getFragmentrh rR r r r ro � � z"TreeBuilderForHtml5lib.getFragmentc sB ddl m� g �t�d��d� ���fdd� ��|d� d���S )Nr rE z8^(.*?)(?: PUBLIC "(.*?)"(?: "(.*?)")?| SYSTEM "(.*?)")?$c s� t | � �r t | t�rW��| �}|rK|�d�}|jdkr>|�d�p"d}|�d�p.|�d�p.d}��dd| |||f � d S ��dd| |f � d S ��d d| f � d S t | t�ri��d
d| | f � d S t | t�r{��d d| | f � d S | jr�d t | j | j
f }n| j
}��d
d| |f � | j r�g }t | j �
� �D ]$\}}t |t�r�d t |j |j
f }t |t �r�d�|�}|�||f� q�t|�D ]\}}��dd|d ||f � q�|d7 }| jD ]}�||� q�d S )Nr\ � rG r r" z|%s<!DOCTYPE %s "%s" "%s">� z|%s<!DOCTYPE %s>z|%s<!DOCTYPE >z|%s<!-- %s -->z|%s"%s"z%s %sz|%s<%s>z
|%s%s="%s")r- r �match�group� lastindexrg r
r r` r r5 �attrs�list�itemsr �join�sorted�children) rh �indent�mr5 rT rU �
attributes�value�child�rF �
doctype_re�rv�serializeElementr r r� � sL
�
�
�z?TreeBuilderForHtml5lib.testSerializer.<locals>.serializeElement�
)r )rI rF �re�compilery �r rh r r� r �testSerializer� s
)
z%TreeBuilderForHtml5lib.testSerializer)NT)r> r? r@ rK rS rZ rb re rf rk rl ro r� �
__classcell__r r rM r r: � s �
r: c @ sL e Zd Zdd� Zdd� Zdd� Zdd� Zd d
� Zd d � Zd
d� Z dd� Z
dS )�AttrListc C s || _ t| j j�| _d S rO )rh r, rv r� r r r rK � s zAttrList.__init__c C s t | j�� ��� S rO )rw rv rx �__iter__rR r r r r� � � zAttrList.__iter__c C s^ | j jpi }||�dg �v s | j j|v r(||�| j jg �v r(t|t�s(t�|�}|| j |< d S )N�*)rh �cdata_list_attributes�getr5 r- rw r
�findall)r r5 r � list_attrr r r � __setitem__ s
zAttrList.__setitem__c C � t | j�� �S rO )rw rv rx rR r r r rx rp zAttrList.itemsc C r� rO �rw rv �keysrR r r r r� rp z
AttrList.keysc C s
t | j�S rO )�lenrv rR r r r �__len__ �
zAttrList.__len__c C s
| j | S rO )rv �r r5 r r r � __getitem__ r� zAttrList.__getitem__c C s |t | j�� �v S rO r� r� r r r � __contains__ r� zAttrList.__contains__N) r> r? r@ rK r� r� rx r� r� r� r� r r r r r� � s r� c @ sx e Zd Zdd� Zdd� Zdd� Zdd� Zeee�Zdd
d �Z d d
� Z
dd� Z dd� Z dd� Z
dd� Zdd� Zee�Zd S )rQ c C s& t j�| |j� || _|| _|| _d S rO )rm �NoderK r5 rh r& r` )r rh r& r` r r r rK s
zElement.__init__c C s* d }}t |t�r| }}n t |t�r|}n|jjtkr%|j }}| |_n|j}| |_t |t�s:|jd ur:|j�� |d urd| jjrd| jjd jtkrd| jjd }| j �
|| �}|� |� || j _ d S t |t�ro| j �
|�}| jjrz| j�
d�}n| jjd ur�| j �
� }n| j}| j j|| j|d� d S )N���F)�parent�most_recent_element)r- r. r rh rN r r� �extract�contentsr& �
new_string� replace_with�_most_recent_element�_last_descendant� next_elementrW )r rj � string_childr� � old_element� new_elementr� r r r rk s:
�zElement.appendChildc C s t | jt�ri S t| j�S rO )r- rh r
r� rR r r r �
getAttributesU s
zElement.getAttributesc C s� |d urJt |�dkrLg }t|�� �D ]\}}t|t�r&t|� }||= |||< q| jj�| j |� t|�� �D ] \}}|| j
|< q6| jj� | j
� d S d S d S )Nr ) r� rw rx r- �tupler r& �builder�$_replace_cdata_list_attribute_valuesr5 rh �set_up_substitutions)r r~ �converted_attributesr5 r �new_namer r r �
setAttributesZ s
�� �zElement.setAttributesNc C s6 t | j�|�| j�}|r| �||� d S | �|� d S rO )rc r& r� � insertBeforerk )r rd r� �textr r r �
insertTextp s zElement.insertTextc C s� | j �|j �}|j jtkr4| j jr4| j j|d jtkr4| j j|d }| j�||j �}|�|� d S | j �||j � | |_ d S )Nr\ )
rh �indexrN r r� r& r� r� �insertr� )r rj �refNoder� �old_node�new_strr r r r� w s
zElement.insertBeforec C s |j �� d S rO )rh r� ri r r r � removeChild� rp zElement.removeChildc C s� | j }|j }|j}|�dd�}t|j�dkr |jd }|j}nd}|j}|j}t|�dkrf|d } |dur9|| _n|| _|| _|durG| |_n| |_|durQ| |_|d �dd�}
||
_|durc|
|_d|
_|D ] } || _|j� | � qhg |_||_dS )z1Move all of this tag's children into another tag.Fr r� NT)
rh � next_siblingr� r� r� r� �previous_element�previous_siblingr� rg ) r �
new_parentrh �new_parent_element�final_next_element�new_parents_last_descendant�new_parents_last_child�(new_parents_last_descendant_next_element� to_append� first_child�last_childs_last_descendantr� r r r �reparentChildren� s>
zElement.reparentChildrenc C sB | j �| jj| j�}t|| j | j�}| jD ] \}}||j|< q|S rO )r& r_ rh r5 r` rQ r~ )r ra rj �keyr r r r � cloneNode� s
zElement.cloneNodec C s | j jS rO )rh r� rR r r r �
hasContent� s zElement.hasContentc C s$ | j d kr td | jfS | j | jfS )N�html)r` r r5 rR r r r � getNameTuple� s
zElement.getNameTuplerO )r> r? r@ rK rk r� r� �propertyr~ r� r� r� r� r� r� r� � nameTupler r r r rQ s 6
B rQ c @ s e Zd Zdd� Zdd� ZdS )rc c C s t j�| d � || _|| _d S rO )rm r� rK rh r& )r rh r& r r r rK � s
zTextNode.__init__c C s t �rO )�NotImplementedErrorrR r r r r� � s zTextNode.cloneNodeN)r> r? r@ rK r� r r r r rc � s rc )$� __license__�__all__r r� � bs4.builderr r r r r � bs4.elementr r
r �html5lib.constantsr r r
r r r �html5lib.treebuildersr rm r/ � ImportError�er r rn r: �objectr� r� rQ rc r r r r �<module> s0 � ��Xx C