3
�PfQ7 � @ s� d Z ddlmZmZmZ ddlZddlZddlZddlm Z ddl
m Z ddlm
Z
dd lm Z
dd
lmZ ddlj Z d Zejd �Ze jd
�jZG dd� de�ZG dd� de�Zdd� Zdd� ZG dd� de j�ZdS )a� Module for supporting the lxml.etree library. The idea here is to use as much
of the native library as possible, without using fragile hacks like custom element
names that break between releases. The downside of this is that we cannot represent
all possible trees; specifically the following are known to cause problems:
Text or comments as siblings of the root element
Docypes with no name
When any of these things occur, we emit a DataLossWarning
� )�absolute_import�division�unicode_literalsN� )�base� )�DataLossWarning)� constants)�etree)� _ihatexmlTz
{([^}]*)}(.*)Zasdc @ s e Zd Zdd� ZdS )� DocumentTypec C s || _ || _|| _d S )N)�name�publicId�systemId)�selfr
r r � r � /usr/lib/python3.6/etree_lxml.py�__init__# s zDocumentType.__init__N)�__name__�
__module__� __qualname__r r r r r r " s r c @ s, e Zd Zdd� Zdd� Zdd� Zee�ZdS )�Documentc C s d | _ g | _d S )N)� _elementTree� _childNodes)r r r r r * s zDocument.__init__c C s | j j� j|j� d S )N)r �getrootZaddnext�_element)r �elementr r r � appendChild. s zDocument.appendChildc C s | j S )N)r )r r r r �_getChildNodes1 s zDocument._getChildNodesN)r r r r r r �propertyZ
childNodesr r r r r ) s r c s6 g �t jdd�� d� ��fdd� ��| d� dj��S )NT)�preventDoubleDashCommentsr c
s t | d��st | d�rˆjd� | jjrz| jjp6| jjsFd| jj }nd| jj| jj| jjf }�jdd|d |f � | j� }x|j� d k r�|j� }q�W xx|d k r��||d � |j � }q�W nTt
| t �s�t
| t �r�jd d| | f � n(�jd
� x | D ]}�||d � � q�W �n�| j
tk�rn�jd d| | jf � t | d ��r| j�r�jd d| | jf � �n�tjj| j
�}|d k �r�|jd
�}|jd�}tj| }�jdd| |� j|�f � n �jdd| � j| j
�f � t | d��r�g }xr| jj� D ]d\} }
tj| �}|d k �rH|j� \}} � j| �} tj| }d|| f } n
� j| �} |j| |
f� �q�W x2t|�D ]&\} }
�jdd|d | |
f � �qpW | j�r��jd d|d | jf � |d7 }x| D ]} �| |� �q�W t | d ��r| j�r�jd d|d | jf � d S )N�tagr z #documentz
zz|%s%s� r z|%s"%s"z#document-fragmentz|%s�tailr z
|%s<%s %s>z|%s<%s>�attribz%s %sz
|%s%s="%s") �hasattr�append�docinfo� internalDTDZ public_idZ
system_url� root_namer Z getpreviousZgetnext�
isinstance�str�bytesr! � comment_type�textr# �etree_builders�
tag_regexp�match�groupr �prefixes� fromXmlNamer$ �items�groups�sorted)
r �indent�dtd_strZ next_elementZnsmatch�nsr! �prefix�
attributesr
�valueZ attr_string�child)�
infosetFilter�rv�serializeElementr r rA ; sp
"
z(testSerializer..serializeElement�
)r )r �
InfosetFilter�join)r r )r? r@ rA r �testSerializer7 s
F
rE c s$ g � � �fdd���| � dj � �S )z4Serialize an element and its child nodes to a stringc s
t | d�sH| jjr:| jjr$| jj}n d| jj }� j|� �| j� � n�| jtkrf� jd| j f � n�| j
s�� jd| jf � n.dj dd� | j
j � D ��}� jd| j|f � | j r�� j| j � x| D ] }�|� q�W � jd | jf � t | d
�o�| j
�r� j| j
� d S ) Nr! z
z z<%s>r" c S s g | ]\}}d ||f �qS )z%s="%s"r )�.0r
r= r r r �
� s z6tostring..serializeElement..z<%s %s>zr# )r% r' r( �doctyper) r& r r! r- r. r$ rD r5 r# )r r9 �attrr> )r@ rA r r rA � s*
z"tostring..serializeElement� )rD )r r )r@ rA r �tostring� s rK c sz e Zd ZeZeZdZdZeZ e
Z ddd�Z dd� Z
dd� Zd d
� Zd d � Zd
d� Zddd�Zd� fdd� Zdd� Z� ZS )� TreeBuilderNFc s� t jt|d��tjdd� �| _|| _G �fdd�dt�� G � ��fdd�d�j�}G ��fdd �d �j �}|| _
|| _ t j
j| |� d S )
N)�fullTreeT)r c s& e Zd Zd� fdd� Z� fdd�ZdS )z(TreeBuilder.__init__..AttributesNc sv |d kr i }|| _ tj| |� xR| j� D ]F\}}t|t�rVd|d � j|d �f }n
� j|�}|| j j j|< q(W d S )Nz{%s}%sr r )r �dictr r5 r* �tuple�coerceAttributer$ )r r r= �keyr
)r? r r r � s
z1TreeBuilder.__init__..Attributes.__init__c sP t j| ||� t|t�r4d|d � j|d �f }n
� j|�}|| jjj|< d S )Nz{%s}%sr r )rN � __setitem__r* rO rP r r$ )r rQ r= r
)r? r r rR � s
z4TreeBuilder.__init__..Attributes.__setitem__)N)r r r r rR r )r? r r �
Attributes� s rS c sx e Zd Z� ��fdd�Z�fdd�Z�fdd�Zeee�Zdd� Z� fd d
�Z eee �Z
d��fd d
� Z �fdd�Z d S )z%TreeBuilder.__init__..Elementc s* �j |�}�jj| ||d� � | �| _d S )N)� namespace)�
coerceElement�Elementr � _attributes)r r
rT )rS �builderr? r r r � s
z.TreeBuilder.__init__..Element.__init__c s$ � j |�| _| j| j| j�| j_d S )N)rU �_nameZ _getETreeTagZ
_namespacer r! )r r
)r? r r �_setName� s z.TreeBuilder.__init__..Element._setNamec s � j | j�S )N)r4 rY )r )r? r r �_getName� s z.TreeBuilder.__init__..Element._getNamec S s | j S )N)rW )r r r r �_getAttributes� s z4TreeBuilder.__init__..Element._getAttributesc s � | |�| _ d S )N)rW )r r< )rS r r �_setAttributes� s z4TreeBuilder.__init__..Element._setAttributesNc s �j |�}� jj| ||� d S )N)ZcoerceCharactersrV �
insertText)r �dataZ insertBefore)rX r? r r r^ � s
z0TreeBuilder.__init__..Element.insertTextc s � j j| |� d S )N)rV r )r r> )rX r r r � s z1TreeBuilder.__init__..Element.appendChild)N)
r r r r rZ r[ r r
r\ r] r< r^ r r )rS rX r? r r rV � s
rV c s8 e Zd Z� �fdd�Z�fdd�Zdd� Zeee�ZdS )z%TreeBuilder.__init__..Commentc s �j |�}� jj| |� d S )N)�
coerceComment�Commentr )r r_ )rX r? r r r � s
z.TreeBuilder.__init__..Comment.__init__c s � j |�}|| j_d S )N)r` r r. )r r_ )r? r r �_setData� s
z.TreeBuilder.__init__..Comment._setDatac S s | j jS )N)r r. )r r r r �_getData� s z.TreeBuilder.__init__..Comment._getDataN)r r r r rb rc r r_ r )rX r? r r ra � s ra )r/ ZgetETreeModuler
r rC r? �namespaceHTMLElementsrN rV ra � elementClass� commentClassr rL r )r rd rM rV ra r )rS rX r? r r � s zTreeBuilder.__init__c C s$ t jj| � | j| _g | _d | _d S )N)r rL �reset�insertCommentInitial�
insertComment�initial_commentsrH )r r r r rg s zTreeBuilder.resetc C s t |�S )N)rE )r r r r r rE s zTreeBuilder.testSerializerc C s t r | jjS | jjj� S d S )N)rM �documentr r )r r r r � getDocument s zTreeBuilder.getDocumentc C sF g }| j d j}|jr"|j|j� |jt|�� |jrB|j|j� |S )Nr )� openElementsr r. r& �extend�listr# )r Zfragmentr r r r � getFragment s zTreeBuilder.getFragmentc C sh |d }|d }|d }|s0t jdt� d | _n4| jj|�}||krPt jdt� | j|||�}|| _d S )Nr
r r z#lxml cannot represent empty doctypez%lxml cannot represent non-xml doctype)�warnings�warnr rH r? rU � doctypeClass)r �tokenr
r r Z coercedNamerH r r r �
insertDoctype s zTreeBuilder.insertDoctypec C s | j j|� d S )N)rj r& )r r_ �parentr r r rh , s z TreeBuilder.insertCommentInitialc sB || j kr,| j jj� d jtkr,tjdt� tt | �j
||� d S )Nr z@lxml cannot represent adjacent comments beyond the root elements�����) rk r r r! r- rq rr r �superrL ri )r r_ rv )� __class__r r �insertCommentMain1 s
z TreeBuilder.insertCommentMainc C s� d}| j r�|d| j j 7 }| j jdk s2| j jdk r�|d| jj| j jpFd� 7 }| j jr�| j j}|jd�dkr�|jd�dkr�tjdt � |j
dd �}|jd�dkr�|d
| 7 }q�|d | 7 }n|d 7 }|d
7 }| j j|d kr�tjdt � |d7 }t j |�}x*| j
D ] }| j|d �}|j|j� �qW | j� | _|j� | j_|d }|jd| j�}|dk�rb|} n d||f } | |_| j||�}
||
_| jjj|
� | jj|
� | j | _ dS )zCreate the document rootrJ z r
zGlxml cannot represent doctype with a different name to the root elementz$r_ rT z{%s}%s) rH r
r r r? Z coercePubid�findrq rr r �replacer
Z
fromstringrj rf Z addpreviousr �
documentClassrk Z getroottreer �getZdefaultNamespacer! re r r& rm rz ri ) r rt ZdocStrZsysid�root�
comment_tokenZcommentr
rT Z etree_tagZ root_elementr r r �
insertRoot7 sJ
zTreeBuilder.insertRoot)F)N)N)r r r r r� r rs re rf Z
fragmentClassr
�implementationr rg rE rl rp ru rh rz r� �
__classcell__r r )ry r rL � s
L
rL ) �__doc__Z
__future__r r r rq �re�sysrJ r r r r
r/ r Z
lxml.etreerM �compiler0 ra r! r- �objectr r rE rK rL r r r r �
s$
O)