
Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­Â�Â­
<!DOCTYPE html>
<html>
3

  \€"  ã               @   sX   d dl Z d dlZd dlZdgZe jddƒZG dd„ dƒZG dd„ dƒZG dd	„ d	ƒZ	dS )
é    NÚRobotFileParserÚRequestRatezrequests secondsc               @   sf   e Zd Zddd„Zdd„ Zdd„ Zdd	„ Zd
d„ Zdd„ Zdd„ Z	dd„ Z
dd„ Zdd„ Zdd„ ZdS )r   Ú c             C   s,   g | _ d | _d| _d| _| j|ƒ d| _d S )NFr   )ÚentriesÚdefault_entryÚdisallow_allÚ	allow_allÚset_urlÚlast_checked)ÚselfÚurl© r   ú*/usr/lib64/python3.6/urllib/robotparser.pyÚ__init__   s    
zRobotFileParser.__init__c             C   s   | j S )N)r
   )r   r   r   r   Úmtime$   s    zRobotFileParser.mtimec             C   s   dd l }|j ƒ | _d S )Nr   )Útimer
   )r   r   r   r   r   Úmodified-   s    zRobotFileParser.modifiedc             C   s&   || _ tjj|ƒdd… \| _| _d S )Né   é   )r   ÚurllibÚparseÚurlparseÚhostÚpath)r   r   r   r   r   r	   5   s    zRobotFileParser.set_urlc             C   s†   yt jj| jƒ}W nR t jjk
rd } z2|jdkr:d| _n|jdkrT|jdk rTd| _W Y d d }~X nX |j	ƒ }| j
|jdƒjƒ ƒ d S )Né‘  é“  Ti�  iô  zutf-8)r   r   )r   ZrequestZurlopenr   ÚerrorZ	HTTPErrorÚcoder   r   Úreadr   ÚdecodeÚ
splitlines)r   ÚfÚerrÚrawr   r   r   r   :   s    
zRobotFileParser.readc             C   s,   d|j kr| jd kr(|| _n| jj|ƒ d S )NÚ*)Ú
useragentsr   r   Úappend)r   Úentryr   r   r   Ú
_add_entryG   s    

zRobotFileParser._add_entryc             C   s6  d}t ƒ }| jƒ  �x|D �]þ}|sT|dkr8t ƒ }d}n|dkrT| j|ƒ t ƒ }d}|jdƒ}|dkrr|d |… }|jƒ }|s€q|jddƒ}t|ƒdkr|d jƒ jƒ |d< tj	j
|d jƒ ƒ|d< |d dk�r |dkrê| j|ƒ t ƒ }|jj|d ƒ d}q|d dk�r4|dk�r|jjt|d dƒƒ d}q|d d	k�rh|dk�r|jjt|d d
ƒƒ d}q|d dk�r¦|dk�r|d jƒ jƒ �r t|d ƒ|_d}q|d dkr|dkr|d jdƒ}t|ƒdk�r|d jƒ jƒ �r|d jƒ jƒ �rtt|d ƒt|d ƒƒ|_d}qW |dk�r2| j|ƒ d S )Nr   r   é   ú#ú:z
user-agentZdisallowFZallowTzcrawl-delayzrequest-rateú/)ÚEntryr   r(   ÚfindÚstripÚsplitÚlenÚlowerr   r   Úunquoter%   r&   Ú	rulelinesÚRuleLineÚisdigitÚintÚdelayr   Úreq_rate)r   ÚlinesÚstater'   ÚlineÚiZnumbersr   r   r   r   P   sd    






 
zRobotFileParser.parsec             C   s    | j r
dS | jrdS | jsdS tjjtjj|ƒƒ}tjjdd|j|j	|j
|jfƒ}tjj|ƒ}|sfd}x"| jD ]}|j|ƒrn|j|ƒS qnW | jrœ| jj|ƒS dS )NFTr   r,   )r   r   r
   r   r   r   r3   Ú
urlunparser   ZparamsZqueryZfragmentÚquoter   Ú
applies_toÚ	allowancer   )r   Ú	useragentr   Z
parsed_urlr'   r   r   r   Ú	can_fetch“   s$    
zRobotFileParser.can_fetchc             C   s4   | j ƒ sd S x| jD ]}|j|ƒr|jS qW | jjS )N)r   r   r@   r8   r   )r   rB   r'   r   r   r   Úcrawl_delay°   s    

zRobotFileParser.crawl_delayc             C   s4   | j ƒ sd S x| jD ]}|j|ƒr|jS qW | jjS )N)r   r   r@   r9   r   )r   rB   r'   r   r   r   Úrequest_rate¸   s    

zRobotFileParser.request_ratec             C   s0   | j }| jd k	r|| jg }djtt|ƒƒd S )NÚ
)r   r   ÚjoinÚmapÚstr)r   r   r   r   r   Ú__str__À   s    
zRobotFileParser.__str__N)r   )Ú__name__Ú
__module__Ú__qualname__r   r   r   r	   r   r(   r   rC   rD   rE   rJ   r   r   r   r   r      s   
		Cc               @   s$   e Zd Zdd„ Zdd„ Zdd„ ZdS )r5   c             C   s>   |dkr| rd}t jjt jj|ƒƒ}t jj|ƒ| _|| _d S )Nr   T)r   r   r>   r   r?   r   rA   )r   r   rA   r   r   r   r   Ê   s
    zRuleLine.__init__c             C   s   | j dkp|j| j ƒS )Nr$   )r   Ú
startswith)r   Úfilenamer   r   r   r@   Ò   s    zRuleLine.applies_toc             C   s   | j r
dndd | j S )NZAllowZDisallowz: )rA   r   )r   r   r   r   rJ   Õ   s    zRuleLine.__str__N)rK   rL   rM   r   r@   rJ   r   r   r   r   r5   Ç   s   r5   c               @   s,   e Zd Zdd„ Zdd„ Zdd„ Zdd„ Zd	S )
r-   c             C   s   g | _ g | _d | _d | _d S )N)r%   r4   r8   r9   )r   r   r   r   r   Û   s    zEntry.__init__c             C   s�   g }x| j D ]}|jd|› �ƒ qW | jd k	r@|jd| j› �ƒ | jd k	rj| j}|jd|j› d|j› �ƒ |jtt| j	ƒƒ |jdƒ dj
|ƒS )NzUser-agent: zCrawl-delay: zRequest-rate: r,   r   rF   )r%   r&   r8   r9   ZrequestsZsecondsÚextendrH   rI   r4   rG   )r   ZretÚagentZrater   r   r   rJ   á   s    


zEntry.__str__c             C   sF   |j dƒd jƒ }x.| jD ]$}|dkr*dS |jƒ }||krdS qW dS )Nr,   r   r$   TF)r0   r2   r%   )r   rB   rQ   r   r   r   r@   î   s    zEntry.applies_toc             C   s$   x| j D ]}|j|ƒr|jS qW dS )NT)r4   r@   rA   )r   rO   r<   r   r   r   rA   û   s    

zEntry.allowanceN)rK   rL   rM   r   rJ   r@   rA   r   r   r   r   r-   Ù   s   r-   )
ÚcollectionsZurllib.parser   Zurllib.requestÚ__all__Ú
namedtupler   r   r5   r-   r   r   r   r   Ú<module>   s    2