£Á°èZ¨Ä…–K§‚«“ô4“ÒÙ´dîfUÙÃÅ WKbyÊ¦•êŽ…È®FÒ¿ÊÎóCozá¬S@6{Í:›œêZÌ:Š•_%:¢¾¾~;‘Ã~èŠ©ÊÇí`ÔÑ©úë™µ'5I¿fš×WO%ø9¾«¾DK|€ùÍD”Ýs]nHÕ¶ê×Ó¼ãžªéUWŸÈË%DÒÕ¬ï‘]/Åcx  ‰ï2ß]ä6G[]S£ÔÏ¯rs{úëóµmÒï#UQxo·õÞCe]"±/aÙ&Eã4ú9Jé_ÞåëdãöKë)AÞ                  ¯¹ægƒÛowÐø^d™ý½ßB7áyMä9ÜÖUã
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
<html>

f  c               @   sd   d  Z  d d l Z d d l Z d g Z Gd d   d  Z Gd d   d  Z Gd d   d  Z d S(	   u<   robotparser.py

    Copyright (C) 2000  Bastian Kleineidam

    You can choose between two licenses when using this package:
    1) GNU GPLv2
    2) PSF license for Python 2.2

    The robots.txt Exclusion Protocol is implemented as specified in
    http://info.webcrawler.com/mak/projects/robots/norobots-rfc.html
i    Nu   RobotFileParserc             B   s   |  Ee  Z d  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 S(   u   RobotFileParserus    This class provides a set of methods to read, parse and answer
    questions about a single robots.txt file.

    u    c             C   s>   g  |  _  d  |  _ d |  _ d |  _ |  j |  d |  _ d  S(   Ni    F(   u   entriesu   Noneu   default_entryu   Falseu   disallow_allu	   allow_allu   set_urlu   last_checked(   u   selfu   url(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __init__   s    				u   RobotFileParser.__init__c             C   s   |  j  S(   u   Returns the time the robots.txt file was last fetched.

        This is useful for long-running web spiders that need to
        check for new robots.txt files periodically.

        (   u   last_checked(   u   self(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   mtime   s    u   RobotFileParser.mtimec             C   s   d d l  } | j    |  _ d S(   uY   Sets the time the robots.txt file was last fetched to the
        current time.

        i    N(   u   timeu   last_checked(   u   selfu   time(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   modified(   s    u   RobotFileParser.modifiedc             C   s5   | |  _  t j j |  d d  \ |  _ |  _ d S(   u,   Sets the URL referring to a robots.txt file.i   i   N(   u   urlu   urllibu   parseu   urlparseu   hostu   path(   u   selfu   url(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   set_url0   s    	u   RobotFileParser.set_urlc             C   s   y t  j j |  j  } Wna t  j j k
 r| } z; | j d k rO d |  _ n | j d k rj d |  _	 n  WYd d } ~ Xn) X| j
   } |  j | j d  j    d S(   u4   Reads the robots.txt URL and feeds it to the parser.i  i  i  Nu   utf-8(   i  i  T(   u   urllibu   requestu   urlopenu   urlu   erroru	   HTTPErroru   codeu   Trueu   disallow_allu	   allow_allu   readu   parseu   decodeu
   splitlines(   u   selfu   fu   erru   raw(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   read5   s    u   RobotFileParser.readc             C   sA   d | j  k r- |  j d  k r= | |  _ q= n |  j j |  d  S(   Nu   *(   u
   useragentsu   default_entryu   Noneu   entriesu   append(   u   selfu   entry(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu
   _add_entryB   s    u   RobotFileParser._add_entryc             C   s  d } t    } x| D]} | sn | d k r@ t    } d } qn | d k rn |  j |  t    } d } qn n  | j d  } | d k r | d |  } n  | j   } | s q n  | j d d  } t |  d k r | d j   j   | d <t j j	 | d j    | d <| d d k rd| d k rG|  j |  t    } n  | j
 j | d  d } q| d d k r| d k r| j j t | d d
   d } qq| d d	 k r| d k r| j j t | d d   d } qqq q W| d k r|  j |  n  d S(   u   Parse the input lines from a robots.txt file.

        We allow that a user-agent: line is not preceded by
        one or more blank lines.
        i    i   i   u   #Nu   :u
   user-agentu   disallowu   allowFT(   u   Entryu
   _add_entryu   findu   stripu   splitu   lenu   loweru   urllibu   parseu   unquoteu
   useragentsu   appendu	   rulelinesu   RuleLineu   Falseu   True(   u   selfu   linesu   stateu   entryu   lineu   i(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   parseK   sJ    
				 	u   RobotFileParser.parsec             C   s   |  j  r d S|  j r d St j j t j j |   } t j j d d | j	 | j
 | j | j f  } t j j |  } | s d } n  x- |  j D]" } | j |  r | j |  Sq W|  j r |  j j |  Sd S(   u=   using the parsed robots.txt decide if useragent can fetch urlu    u   /FT(   u   disallow_allu   Falseu	   allow_allu   Trueu   urllibu   parseu   urlparseu   unquoteu
   urlunparseu   pathu   paramsu   queryu   fragmentu   quoteu   entriesu
   applies_tou	   allowanceu   default_entry(   u   selfu	   useragentu   urlu
   parsed_urlu   entry(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu	   can_fetch~   s     				u   RobotFileParser.can_fetchc             C   s   d j  d d   |  j D  S(   Nu    c             S   s    g  |  ] } t  |  d   q S(   u   
(   u   str(   u   .0u   entry(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu
   <listcomp>   s   	 u+   RobotFileParser.__str__.<locals>.<listcomp>(   u   joinu   entries(   u   self(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __str__   s    u   RobotFileParser.__str__N(   u   __name__u
   __module__u   __qualname__u   __doc__u   __init__u   mtimeu   modifiedu   set_urlu   readu
   _add_entryu   parseu	   can_fetchu   __str__(   u
   __locals__(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   RobotFileParser   s   		3c             B   s>   |  Ee  Z d  Z d Z d d   Z d d   Z d d   Z d S(	   u   RuleLineuo   A rule line is a single "Allow:" (allowance==True) or "Disallow:"
       (allowance==False) followed by a path.c             C   s\   | d k r | r d } n  t j j t j j |   } t j j |  |  _ | |  _ d  S(   Nu    T(   u   Trueu   urllibu   parseu
   urlunparseu   urlparseu   quoteu   pathu	   allowance(   u   selfu   pathu	   allowance(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __init__   s
    	u   RuleLine.__init__c             C   s   |  j  d k p | j |  j   S(   Nu   *(   u   pathu
   startswith(   u   selfu   filename(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu
   applies_to   s    u   RuleLine.applies_toc             C   s   |  j  r d p d d |  j S(   Nu   Allowu   Disallowu   : (   u	   allowanceu   path(   u   self(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __str__   s    u   RuleLine.__str__N(   u   __name__u
   __module__u   __qualname__u   __doc__u   __init__u
   applies_tou   __str__(   u
   __locals__(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   RuleLine   s   u   RuleLinec             B   sJ   |  Ee  Z d  Z d Z d d   Z d d   Z d d   Z d d	   Z d
 S(   u   Entryu?   An entry has one or more user-agents and zero or more rulelinesc             C   s   g  |  _  g  |  _ d  S(   N(   u
   useragentsu	   rulelines(   u   self(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __init__   s    	u   Entry.__init__c             C   sj   g  } x' |  j  D] } | j d | d g  q Wx* |  j D] } | j t |  d g  q: Wd j |  S(   Nu   User-agent: u   
u    (   u
   useragentsu   extendu	   rulelinesu   stru   join(   u   selfu   retu   agentu   line(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   __str__   s    u   Entry.__str__c             C   s]   | j  d  d j   } x= |  j D]2 } | d k r9 d S| j   } | | k r# d Sq# Wd S(   u2   check if this entry applies to the specified agentu   /i    u   *TF(   u   splitu   loweru
   useragentsu   Trueu   False(   u   selfu	   useragentu   agent(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu
   applies_to   s    u   Entry.applies_toc             C   s.   x' |  j  D] } | j |  r
 | j Sq
 Wd S(   uZ   Preconditions:
        - our agent applies to this entry
        - filename is URL decodedT(   u	   rulelinesu
   applies_tou	   allowanceu   True(   u   selfu   filenameu   line(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu	   allowance   s    u   Entry.allowanceN(   u   __name__u
   __module__u   __qualname__u   __doc__u   __init__u   __str__u
   applies_tou	   allowance(   u
   __locals__(    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   Entry   s
   u   Entry(   u   __doc__u   urllib.parseu   urllibu   urllib.requestu   __all__u   RobotFileParseru   RuleLineu   Entry(    (    (    u7   /opt/alt/python33/lib64/python3.3/urllib/robotparser.pyu   <module>   s
   	