o
    „¿ÑgÓ5  ã                   @  s¦   d Z ddlmZ ddlZddlZddlZej d¡Z	ej 
e	¡Ze	j e¡ eejd< e d¡e_e d¡e_eje_e dej¡e_e d	¡ZG d
d„ dejƒZdS )a  
This module imports a copy of [`html.parser.HTMLParser`][] and modifies it heavily through monkey-patches.
A copy is imported rather than the module being directly imported as this ensures that the user can import
and  use the unmodified library for their own needs.
é    )ÚannotationsNzhtml.parserÚ
htmlparserz\?>z&([a-zA-Z][-.a-zA-Z0-9]*);a”  
  <[a-zA-Z][^`\t\n\r\f />\x00]*       # tag name <= added backtick here
  (?:[\s/]*                           # optional whitespace before attribute name
    (?:(?<=['"\s/])[^`\s/>][^\s/=>]*  # attribute name <= added backtick here
      (?:\s*=+\s*                     # value indicator
        (?:'[^']*'                    # LITA-enclosed value
          |"[^"]*"                    # LIT-enclosed value
          |(?!['"])[^`>\s]*           # bare value <= added backtick here
         )
         (?:\s*,)*                    # possibly followed by a comma
       )?(?:\s|/(?!>))*
     )*
   )?
  \s*                                 # trailing whitespace
z^([ ]*\n){2}c                      s  e Zd ZU dZ‡ fdd„Z‡ fdd„Z‡ fdd„Zed:d
d„ƒZd;dd„Z	d<dd„Z
d=dd„Zd>dd„Zd?dd„Zd@dd„Zd=dd „ZdAd"d#„ZdAd$d%„Zd?d&d'„Zd?d(d)„Zd?d*d+„Zd?d,d-„ZdB‡ fd/d0„ZdB‡ fd1d2„Zd3Zd4ed5< dCd6d7„ZdBd8d9„Z‡  ZS )DÚHTMLExtractorzû
    Extract raw HTML from text.

    The raw HTML is stored in the [`htmlStash`][markdown.util.HtmlStash] of the
    [`Markdown`][markdown.Markdown] instance passed to `md` and the remaining text
    is stored in `cleandoc` as a list of strings.
    c                   s@   d|vrd|d< t dgƒ| _dg| _tƒ j|i |¤Ž || _d S )NÚconvert_charrefsFÚhrr   )ÚsetÚ
empty_tagsÚlineno_start_cacheÚsuperÚ__init__Úmd)Úselfr   ÚargsÚkwargs©Ú	__class__© úS/var/www/html/bloggers_ems/venv/lib/python3.10/site-packages/markdown/htmlparser.pyr   O   s   
zHTMLExtractor.__init__c                   s4   d| _ d| _g | _g | _g | _dg| _tƒ  ¡  dS )z1Reset this instance.  Loses all unprocessed data.Fr   N)ÚinrawÚintailÚstackÚ_cacheÚcleandocr	   r
   Úreset©r   r   r   r   r   \   s   zHTMLExtractor.resetc                   sv   t ƒ  ¡  t| jƒr | jr| js|  t | j¡¡ n|  | j¡ t| j	ƒr9| j
 | jj d | j	¡¡¡ g | _	dS dS )zHandle any buffered data.Ú N)r
   ÚcloseÚlenÚrawdatar   Ú
cdata_elemÚhandle_datar   Úunescaper   r   Úappendr   Ú	htmlStashÚstoreÚjoinr   r   r   r   r   g   s   



þzHTMLExtractor.closeÚreturnÚintc                 C  sj   t t| jƒd | jd ƒD ]}| j| }| j d|¡}|dkr$t| jƒ}| j |d ¡ q| j| jd  S )zHReturns char index in `self.rawdata` for the start of the current line. é   Ú
éÿÿÿÿ)Úranger   r	   Úlinenor   Úfindr"   )r   ÚiiÚlast_line_start_posÚlf_posr   r   r   Úline_offsetv   s   

zHTMLExtractor.line_offsetÚboolc                 C  s<   | j dkrdS | j dkrdS | j| j| j| j  …  ¡ dkS )z†
        Returns True if current position is at start of line.

        Allows for up to three blank spaces at start of line.
        r   Té   Fr   )Úoffsetr   r1   Ústripr   r   r   r   Úat_line_startƒ   s
   

 zHTMLExtractor.at_line_startÚtagÚstrc                 C  s<   | j | j }tj | j|¡}|r| j|| ¡ … S d |¡S )z™
        Returns the text of the end tag.

        If it fails to extract the actual text from the raw data, it builds a closing tag with `tag`.
        z</{}>)r1   r4   r   Ú	endendtagÚsearchr   ÚendÚformat)r   r7   ÚstartÚmr   r   r   Úget_endtag_text   s
   
zHTMLExtractor.get_endtag_textÚattrsúlist[tuple[str, str]]c                 C  sœ   || j v r|  ||¡ d S | j |¡r&| js|  ¡ r&| js&d| _| j d¡ |  	¡ }| jr;| j
 |¡ | j |¡ d S | j |¡ || jv rL|  ¡  d S d S )NTr)   )r   Úhandle_startendtagr   Úis_block_levelr   r6   r   r   r"   Úget_starttag_textr   r   ÚCDATA_CONTENT_ELEMENTSÚclear_cdata_mode)r   r7   r@   Útextr   r   r   Úhandle_starttagŸ   s   
 
þzHTMLExtractor.handle_starttagc                 C  sÖ   |   |¡}| jrc| j |¡ || jv r!| jr!| j ¡ |krn| jst| jƒdkrat | j	| j
| j t|ƒ d … ¡rA| j d¡ nd| _d| _| j | jj d | j¡¡¡ | j d¡ g | _d S d S | j |¡ d S )Nr   r)   TFr   ú

)r?   r   r   r"   r   Úpopr   Úblank_line_reÚmatchr   r1   r4   r   r   r   r#   r$   r%   )r   r7   rG   r   r   r   Úhandle_endtag´   s$   

þ$
ózHTMLExtractor.handle_endtagÚdatac                 C  s:   | j r
d|v r
d| _ | jr| j |¡ d S | j |¡ d S )Nr)   F)r   r   r   r"   r   ©r   rN   r   r   r   r    Ï   s
   zHTMLExtractor.handle_dataÚis_blockc                 C  sÆ   | j s| jr| j |¡ dS |  ¡ r[|r[t | j| j| j	 t
|ƒ d… ¡r+|d7 }nd| _| jr6| jd nd}| d¡sH| d¡rH| j d¡ | j | jj |¡¡ | j d¡ dS | j |¡ dS )z Handle empty tags (`<data>`). Nr)   Tr*   r   rI   )r   r   r   r"   r6   rK   rL   r   r1   r4   r   r   Úendswithr   r#   r$   )r   rN   rP   Úitemr   r   r   Úhandle_empty_tag×   s   $
zHTMLExtractor.handle_empty_tagc                 C  s   | j |  ¡ | j |¡d d S )N©rP   )rS   rD   r   rC   )r   r7   r@   r   r   r   rB   î   s   z HTMLExtractor.handle_startendtagÚnamec                 C  ó   | j d |¡dd d S )Nz&#{};FrT   ©rS   r<   ©r   rU   r   r   r   Úhandle_charrefñ   ó   zHTMLExtractor.handle_charrefc                 C  rV   )Nz&{};FrT   rW   rX   r   r   r   Úhandle_entityrefô   rZ   zHTMLExtractor.handle_entityrefc                 C  rV   )Nz	<!--{}-->TrT   rW   rO   r   r   r   Úhandle_comment÷   rZ   zHTMLExtractor.handle_commentc                 C  rV   )Nz<!{}>TrT   rW   rO   r   r   r   Úhandle_declú   rZ   zHTMLExtractor.handle_declc                 C  rV   )Nz<?{}?>TrT   rW   rO   r   r   r   Ú	handle_piý   rZ   zHTMLExtractor.handle_pic                 C  s,   |  d¡rdnd}| jd ||¡dd d S )NzCDATA[z]]>z]>z<![{}{}TrT   )Ú
startswithrS   r<   )r   rN   r;   r   r   r   Úunknown_decl   s   zHTMLExtractor.unknown_declÚic                   ó,   |   ¡ s| jrtƒ  |¡S |  d¡ |d S )Nz<?é   )r6   r   r
   Úparse_pir    ©r   ra   r   r   r   rd     ó   
zHTMLExtractor.parse_pic                   rb   )Nz<!rc   )r6   r   r
   Úparse_html_declarationr    re   r   r   r   rg     rf   z$HTMLExtractor.parse_html_declarationNz
str | NoneÚ_HTMLExtractor__starttag_textc                 C  s   | j S )z)Return full source of start tag: `<...>`.)rh   r   r   r   r   rD     s   zHTMLExtractor.get_starttag_textc                 C  sê  d | _ |  |¡}|dk r|S | j}|||… | _ g }tj ||d ¡}|s)J dƒ‚| ¡ }| d¡ ¡  | _	}||k r™tj
 ||¡}|sEnT| ddd¡\}	}
}|
sTd }n-|d d… d  krf|dd … ks{n |d d… d  kry|dd … krn n|dd… }|rˆt |¡}| |	 ¡ |f¡ | ¡ }||k s;|||…  ¡ }|d	vrÖ|  ¡ \}}d
| j v rÄ|| j  d
¡ }t| j ƒ| j  d
¡ }n|t| j ƒ }|  |||… ¡ |S | d¡rã|  ||¡ |S || jv rí|  |¡ |  ||¡ |S )Nr   r(   z#unexpected call to parse_starttag()rc   r3   ú'r*   ú")ú>ú/>r)   rl   )rh   Úcheck_for_whole_start_tagr   r   Útagfind_tolerantrL   r;   ÚgroupÚlowerÚlasttagÚattrfind_tolerantr!   r"   r5   ÚgetposÚcountr   Úrfindr    rQ   rB   rE   Úset_cdata_moderH   )r   ra   Úendposr   r@   rL   Úkr7   r>   ÚattrnameÚrestÚ	attrvaluer;   r,   r4   r   r   r   Úparse_starttag  sX   
&(
ó

ÿ

ý
zHTMLExtractor.parse_starttag)r&   r'   )r&   r2   )r7   r8   r&   r8   )r7   r8   r@   rA   )r7   r8   )rN   r8   )rN   r8   rP   r2   )rU   r8   )ra   r'   r&   r'   )r&   r8   )Ú__name__Ú
__module__Ú__qualname__Ú__doc__r   r   r   Úpropertyr1   r6   r?   rH   rM   r    rS   rB   rY   r[   r\   r]   r^   r`   rd   rg   rh   Ú__annotations__rD   r|   Ú__classcell__r   r   r   r   r   F   s2   
 













r   )r€   Ú
__future__r   ÚreÚimportlib.utilÚ	importlibÚsysÚutilÚ	find_specÚspecÚmodule_from_specr   ÚloaderÚexec_moduleÚmodulesÚcompileÚpicloseÚ	entityrefÚ
incompleteÚVERBOSEÚlocatestarttagend_tolerantrK   Ú
HTMLParserr   r   r   r   r   Ú<module>   s"   
ò
