ó
    F\hü$  ã                   ó�   • S r SSKrSSKrSSKrSSKrS/r\R                  " SS5      r " S S5      r	 " S S5      r
 " S	 S
5      rg)a  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://www.robotstxt.org/norobots-rfc.txt
é    NÚRobotFileParserÚRequestRatezrequests secondsc                   ód   • \ rS rSrSrSS jrS rS rS rS r	S r
S	 rS
 rS rS rS rS rSrg)r   é   zjThis class provides a set of methods to read, parse and answer
questions about a single robots.txt file.

c                 óz   • / U l         / U l        S U l        SU l        SU l        U R                  U5        SU l        g )NFr   )ÚentriesÚsitemapsÚdefault_entryÚdisallow_allÚ	allow_allÚset_urlÚlast_checked©ÚselfÚurls     Ú)/usr/lib/python3.13/urllib/robotparser.pyÚ__init__ÚRobotFileParser.__init__   s;   € ØˆŒØˆŒØ!ˆÔØ!ˆÔØˆŒØ�‰�SÔØˆÕó    c                 ó   • U R                   $ )zŸReturns 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.

)r   ©r   s    r   ÚmtimeÚRobotFileParser.mtime&   s   € ð × Ñ Ð r   c                 ó6   • SSK nUR                  5       U l        g)zISets the time the robots.txt file was last fetched to the
current time.

r   N)Útimer   )r   r   s     r   ÚmodifiedÚRobotFileParser.modified/   s   € ó
 	Ø ŸI™I›KˆÕr   c                 ón   • Xl         [        R                  R                  U5      SS u  U l        U l        g)z,Sets the URL referring to a robots.txt file.é   é   N)r   ÚurllibÚparseÚurlparseÚhostÚpathr   s     r   r   ÚRobotFileParser.set_url7   s+   € àŒÜ%Ÿ|™|×4Ñ4°SÓ9¸!¸AÐ>ÑˆŒ	�4•9r   c                 óÈ  •  [         R                  R                  U R                  5      nUR	                  5       nU R                  UR                  S5      R                  5       5        g! [         R                  R                   aY  nUR                  S;   a  SU l        O'UR                  S:¼  a  UR                  S:  a  SU l        UR                  5          SnAgSnAff = f)z4Reads the robots.txt URL and feeds it to the parser.zutf-8)i‘  i“  Ti�  iô  N)r!   ÚrequestÚurlopenr   Úreadr"   ÚdecodeÚ
splitlinesÚerrorÚ	HTTPErrorÚcoder   r   Úclose)r   ÚfÚrawÚerrs       r   r*   ÚRobotFileParser.read<   s¡   € ð
	9Ü—‘×&Ñ& t§x¡xÓ0ˆAð —&‘&“(ˆCØ�J‰J�s—z‘z 'Ó*×5Ñ5Ó7Õ8øô �|‰|×%Ñ%ó 	Ø�x‰x˜:Ó%Ø$(�Õ!Ø—‘˜S“ S§X¡X°£^Ø!%�”Ø�I‰I�K‰Kûð	ús   ‚)A* Á*C!ÂACÃC!c                 ó„   • SUR                   ;   a  U R                  c  Xl        g g U R                  R                  U5        g ©NÚ*)Ú
useragentsr
   r   Úappend)r   Úentrys     r   Ú
_add_entryÚRobotFileParser._add_entryJ   s;   € Ø�%×"Ñ"Ó"à×!Ñ!Ñ)à%*Õ"ð *ð �L‰L×Ñ Õ&r   c                 óB  • Sn[        5       nU R                  5         U GHä  nU(       d6  US:X  a  [        5       nSnO#US:X  a  U R                  U5        [        5       nSnUR                  S5      nUS:¼  a  USU nUR	                  5       nU(       d  Mv  UR                  SS5      n[        U5      S:X  d  M™  US   R	                  5       R                  5       US'   [        R                  R                  US   R	                  5       5      US'   US   S:X  aD  US:X  a  U R                  U5        [        5       nUR                  R                  US   5        SnGM=  US   S:X  a6  US:w  a-  UR                  R                  [        US   S	5      5        SnGMy  GM|  US   S
:X  a6  US:w  a-  UR                  R                  [        US   S5      5        SnGM¸  GM»  US   S:X  aG  US:w  a>  US   R	                  5       R                  5       (       a  [!        US   5      Ul        SnGM  GM  US   S:X  a¦  US:w  a�  US   R                  S5      n[        U5      S:X  au  US   R	                  5       R                  5       (       aO  US   R	                  5       R                  5       (       a)  [%        [!        US   5      [!        US   5      5      Ul        SnGM·  GMº  US   S:X  d  GMÆ  U R(                  R                  US   5        GMç     US:X  a  U R                  U5        gg)z|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.
r   r   é   Ú#NÚ:z
user-agentÚdisallowFÚallowTzcrawl-delayzrequest-rateÚ/Úsitemap)ÚEntryr   r;   ÚfindÚstripÚsplitÚlenÚlowerr!   r"   Úunquoter8   r9   Ú	rulelinesÚRuleLineÚisdigitÚintÚdelayr   Úreq_rater	   )r   ÚlinesÚstater:   ÚlineÚiÚnumberss          r   r"   ÚRobotFileParser.parseS   s¸  € ð ˆÜ“ˆà�‰ŒÜˆDÞØ˜A“:Ü!›G�EØ‘EØ˜a“ZØ—O‘O EÔ*Ü!›G�EØ�Eà—	‘	˜#“ˆAØ�A‹vØ˜B˜Q�x�Ø—:‘:“<ˆDÞÙØ—:‘:˜c 1Ó%ˆDÜ�4‹y˜A�~Ø˜q™'Ÿ-™-›/×/Ñ/Ó1��Q‘Ü Ÿ,™,×.Ñ.¨t°A©w¯}©}«Ó?��Q‘Ø˜‘7˜lÓ*Ø “zØŸ™¨Ô.Ü %£˜Ø×$Ñ$×+Ñ+¨D°©GÔ4Ø“EØ˜!‘W 
Ó*Ø “zØŸ™×.Ñ.¬x¸¸Q¹ÀÓ/GÔHØ !›ò "ð ˜!‘W Ó'Ø “zØŸ™×.Ñ.¬x¸¸Q¹ÀÓ/FÔGØ !›ò "ð ˜!‘W Ó-Ø “zð   ™7Ÿ=™=›?×2Ñ2×4Ñ4Ü*-¨d°1©g«,˜EœKØ !›ò "ð ˜!‘W Ó.Ø “zØ"& q¡'§-¡-°Ó"4˜ä ›L¨AÓ-°'¸!±*×2BÑ2BÓ2D×2LÑ2L×2NÑ2NØ '¨¡
× 0Ñ 0Ó 2× :Ñ :× <Ñ <Ü-8¼¸WÀQ¹Z»Ì#ÈgÐVWÉjË/Ó-Z˜EœNØ !›ò "ð ˜!‘W 	Ö)ð
 —M‘M×(Ñ(¨¨a©×1ño ðp �A‹:Ø�O‰O˜EÕ"ð r   c                 ó¢  • U R                   (       a  gU R                  (       a  gU R                  (       d  g[        R                  R                  [        R                  R                  U5      5      n[        R                  R                  SSUR                  UR                  UR                  UR                  45      n[        R                  R                  U5      nU(       d  SnU R                   H,  nUR                  U5      (       d  M  UR                  U5      s  $    U R                   (       a  U R                   R                  U5      $ g)z=using the parsed robots.txt decide if useragent can fetch urlFTÚ rC   )r   r   r   r!   r"   r#   rK   Ú
urlunparser%   ÚparamsÚqueryÚfragmentÚquoter   Ú
applies_toÚ	allowancer
   )r   Ú	useragentr   Ú
parsed_urlr:   s        r   Ú	can_fetchÚRobotFileParser.can_fetchœ   sø   € à××ØØ�>�>Øð
 × × Øô —\‘\×*Ñ*¬6¯<©<×+?Ñ+?ÀÓ+DÓEˆ
Ü�l‰l×%Ñ% r¨"¨Z¯_©_Ø×Ñ˜j×.Ñ.°
×0CÑ0Cð'Eó Fˆä�l‰l× Ñ  Ó%ˆÞØˆCØ—\”\ˆEØ×Ñ 	×*Ó*Ø—‘ sÓ+Ò+ñ "ð ××Ø×%Ñ%×/Ñ/°Ó4Ð4àr   c                 óì   • U R                  5       (       d  g U R                   H'  nUR                  U5      (       d  M  UR                  s  $    U R                  (       a  U R                  R                  $ g ©N)r   r   r_   rP   r
   ©r   ra   r:   s      r   Úcrawl_delayÚRobotFileParser.crawl_delay¹   sY   € Ø�z‰z�|‰|ØØ—\”\ˆEØ×Ñ 	×*Ó*Ø—{‘{Ò"ñ "ð ××Ø×%Ñ%×+Ñ+Ð+Ør   c                 óì   • U R                  5       (       d  g U R                   H'  nUR                  U5      (       d  M  UR                  s  $    U R                  (       a  U R                  R                  $ g rf   )r   r   r_   rQ   r
   rg   s      r   Úrequest_rateÚRobotFileParser.request_rateÃ   sY   € Ø�z‰z�|‰|ØØ—\”\ˆEØ×Ñ 	×*Ó*Ø—~‘~Ò%ñ "ð ××Ø×%Ñ%×.Ñ.Ð.Ør   c                 ó>   • U R                   (       d  g U R                   $ rf   )r	   r   s    r   Ú	site_mapsÚRobotFileParser.site_mapsÍ   s   € Ø�}�}ØØ�}‰}Ðr   c                 ó�   • U R                   nU R                  b  XR                  /-   nSR                  [        [        U5      5      $ )Nz

)r   r
   ÚjoinÚmapÚstr)r   r   s     r   Ú__str__ÚRobotFileParser.__str__Ò   s>   € Ø—,‘,ˆØ×ÑÑ)Ø×!3Ñ!3Ð 4Ñ4ˆGØ�{‰{œ3œs GÓ,Ó-Ð-r   )	r   r
   r   r   r$   r   r%   r	   r   N)rY   )Ú__name__Ú
__module__Ú__qualname__Ú__firstlineno__Ú__doc__r   r   r   r   r*   r;   r"   rc   rh   rk   rn   rt   Ú__static_attributes__© r   r   r   r      sE   † ñô
ò!ò(ò?ò
9ò'òG#òRò:òòõ
.r   c                   ó*   • \ rS rSrSrS rS rS rSrg)rM   éÙ   zhA rule line is a single "Allow:" (allowance==True) or "Disallow:"
(allowance==False) followed by a path.c                 óî   • US:X  a	  U(       d  Sn[         R                  R                  [         R                  R                  U5      5      n[         R                  R	                  U5      U l        X l        g )NrY   T)r!   r"   rZ   r#   r^   r%   r`   )r   r%   r`   s      r   r   ÚRuleLine.__init__Ü   sN   € Ø�2‹:žiàˆIÜ�|‰|×&Ñ&¤v§|¡|×'<Ñ'<¸TÓ'BÓCˆÜ—L‘L×&Ñ& tÓ,ˆŒ	Ø"�r   c                 ód   • U R                   S:H  =(       d    UR                  U R                   5      $ r6   )r%   Ú
startswith)r   Úfilenames     r   r_   ÚRuleLine.applies_toä   s%   € Ø�y‰y˜CÑ×A 8×#6Ñ#6°t·y±yÓ#AÐAr   c                 óL   • U R                   (       a  SOSS-   U R                  -   $ )NÚAllowÚDisallowz: ©r`   r%   r   s    r   rt   ÚRuleLine.__str__ç   s   € ØŸ>Ÿ>‘¨z¸TÑAÀDÇIÁIÑMÐMr   rˆ   N)	rv   rw   rx   ry   rz   r   r_   rt   r{   r|   r   r   rM   rM   Ù   s   † ñ1ò#òBõNr   rM   c                   ó0   • \ rS rSrSrS rS rS rS rSr	g)	rE   éë   z?An entry has one or more user-agents and zero or more rulelinesc                 ó<   • / U l         / U l        S U l        S U l        g rf   )r8   rL   rP   rQ   r   s    r   r   ÚEntry.__init__í   s   € ØˆŒØˆŒØˆŒ
Øˆ�r   c                 ó¦  • / nU R                    H  nUR                  SU 35        M     U R                  b  UR                  SU R                   35        U R                  b7  U R                  nUR                  SUR                   SUR
                   35        UR                  [        [        U R                  5      5        SR                  U5      $ )NzUser-agent: zCrawl-delay: zRequest-rate: rC   Ú
)r8   r9   rP   rQ   ÚrequestsÚsecondsÚextendrr   rs   rL   rq   )r   ÚretÚagentÚrates       r   rt   ÚEntry.__str__ó   s¢   € ØˆØ—_”_ˆEØ�J‰J˜ e WÐ-Ö.ñ %à�:‰:Ñ!Ø�J‰J˜ t§z¡z lÐ3Ô4Ø�=‰=Ñ$Ø—=‘=ˆDØ�J‰J˜¨¯© °a¸¿¹°~ÐFÔGØ�
‰
”3”s˜DŸN™NÓ+Ô,Ø�y‰y˜‹~Ðr   c                 ó¬   • UR                  S5      S   R                  5       nU R                   H"  nUS:X  a    gUR                  5       nX!;   d  M"    g   g)z2check if this entry applies to the specified agentrC   r   r7   TF)rH   rJ   r8   )r   ra   r”   s      r   r_   ÚEntry.applies_toÿ   sQ   € ð —O‘O CÓ(¨Ñ+×1Ñ1Ó3ˆ	Ø—_”_ˆEØ˜‹|áØ—K‘K“MˆEØÕ!Ùñ %ð r   c                 ór   • U R                    H'  nUR                  U5      (       d  M  UR                  s  $    g)zJPreconditions:
- our agent applies to this entry
- filename is URL decodedT)rL   r_   r`   )r   rƒ   rT   s      r   r`   ÚEntry.allowance  s0   € ð —N”NˆDØ�‰˜x×(Ó(Ø—~‘~Ò%ñ #ð r   )rP   rQ   rL   r8   N)
rv   rw   rx   ry   rz   r   rt   r_   r`   r{   r|   r   r   rE   rE   ë   s   † ÙIòò
òõr   rE   )rz   ÚcollectionsÚurllib.errorr!   Úurllib.parseÚurllib.requestÚ__all__Ú
namedtupler   r   rM   rE   r|   r   r   Ú<module>r¡      sV   ðñ
ó Û Û Û àÐ
€à×$Ò$ ]Ð4FÓG€÷.ñ .÷DNñ N÷$(ò (r   