g&0ddlZddlZddlmZmZmZddZy)N) HTML2Text __version__configc d}Gdd}tj}|jddtjd|jdd d tj d |jd ddtj d |jddd tjd |jddd tjd |jddd tjd |jdddtjd |jddd tjd |jd d d!tjd"#|jd$d%d tjd& |jd'd(d tjd) |jd*d+d tjd, |jd-d.d tj d/ |jd0d1d tj"d2 |jd3d4d d5d6d7#|jd8d9d d:d6d;#|jdd6d?#|jd@dAdBt$tj&dCD|jdEdFdGt$tj(dHD|jdIdJd dKd6dL#|jdMd dNd6dO#|jdPd dQtj*dR#|jdSd dTtj,dU#|jdVd dWtj.dX#|jdYd dZtj0d[#|jd\dd]tj2d^#|jd_dd`tj4da#|jdbd dctj6dd#|jded dftj8dg#|jdhditj:dj|jdkdltj<dm|jdndotj>dp|jdqdrdsjAtCtDtFt|jdudvw|jdxdvdyz|jd{d|d tjHd} |jK}|jLr?|jLd~k7r0tO|jLd5}|jQ}dddn(tRjTjVjQ} jY|jZ|j\}ti|} |jjrd~| _6|jnrd| _8d| _9|jt| _:|jv| _<|jz| _=|j|| _>|j~| _?|j| _@|j| _A|j| _B|j| _C|j| _D|j| _E|j| _F|j| _G|j| _H|j| _I|j| _J|j| _K|j| _L|j| _M|j| _N|j| _O|j| _P|j| _Q|j| _R|j| _S|j| _T|j| _U|j| _V|j| _W|j| _XtRjj| j|y#1swYxYw#t^$rU}|j`dz|jbz}|d|jdzz }|d|jbzdzz }tg||d}~wwxYw)Nc,eZdZdZdZdZdZdZdZdZ dZ y ) main..bcolorszzzzzzzzN) __name__ __module__ __qualname__HEADEROKBLUEOKGREENWARNINGFAILENDCBOLD UNDERLINEK/var/www/openai/venv/lib/python3.12/site-packages/crawl4ai/html2text/cli.pybcolorsr s* rrz--default-image-altdefault_image_altz3The default alt string for images with missing ones)destdefaulthelpz --pad-tables pad_tables store_truez-pad the cells to equal column width in tables)ractionrrz--no-wrap-links wrap_links store_falsez"don't wrap links during conversionz--wrap-list-itemswrap_list_itemsz!wrap list items during conversionz --wrap-tables wrap_tablesz wrap tablesz--ignore-emphasisignore_emphasisz)don't include any formatting for emphasisz--reference-links inline_linksz1use reference style links instead of inline linksz--ignore-links ignore_linksz&don't include any formatting for linksz--ignore-mailto-linksignore_mailto_linkszdon't include mailto: links)r rrrz--protect-links protect_linkszCprotect links from line breaks surrounding them with angle bracketsz--ignore-images ignore_imagesz'don't include any formatting for imagesz--images-as-htmlimages_as_htmlzWAlways write image tags as raw html; preserves `height`, `width` and `alt` if possible.z--images-to-alt images_to_altz&Discard image data, only keep alt textz--images-with-sizeimages_with_sizezMWrite image tags with height and width attrs as raw html to retain dimensionsz-gz --google-doc google_docFz(convert an html-exported Google Documentz-dz--dash-unordered-list ul_style_dashz6use a dash rather than a star for unordered list itemsz-ez--asterisk-emphasisem_style_asteriskz=use an asterisk rather than an underscore for emphasized textz-bz --body-width body_widthz3number of characters per output line, 0 for no wrap)rtyperrz-iz--google-list-indent list_indentz,number of pixels Google indents nested listsz-sz--hide-strikethroughhide_strikethroughzDhide strike-through text. only relevant when -g is specified as wellz --escape-all escape_snobzbEscape all special characters. Output is less readable, but avoids corner case formatting issues.z--bypass-tables bypass_tablesz2Format tables in HTML rather than Markdown syntax.z--ignore-tables ignore_tableszAIgnore table-related tags (table, th, td, tr) while keeping rows.z--single-line-breaksingle_line_breakzhUse a single line break after a block element rather than two line breaks. NOTE: Requires --body-width=0z--unicode-snob unicode_snobzUse unicode throughout documentz--no-automatic-linksuse_automatic_linksz.Do not use automatic links wherever applicablez--no-skip-internal-linksskip_internal_linkszDo not skip internal linksz--links-after-paralinks_each_paragraphz2Put links after each paragraph instead of documentz --mark-code mark_codez.Mark program code blocks with [code]...[/code]z--decode-errors decode_errorszZWhat to do in case of decode errors.'ignore', 'strict' and 'replace' are acceptable valuesz --open-quote open_quotez!The character used to open quotesz --close-quote close_quotez"The character used to close quotesz --versionversion.)r rAfilename?)nargsencodingzutf-8)rErz--include-sup-subinclude_sup_subzInclude the sup and sub tags-rbzWarning:z Use the z--decode-errors=ignorez flag.)baseurl*__)\argparseArgumentParser add_argumentrDEFAULT_IMAGE_ALT PAD_TABLES WRAP_LINKSWRAP_LIST_ITEMS WRAP_TABLESIGNORE_EMPHASIS INLINE_LINKSIGNORE_ANCHORSIGNORE_MAILTO_LINKS PROTECT_LINKS IGNORE_IMAGESIMAGES_AS_HTML IMAGES_TO_ALTIMAGES_WITH_SIZEint BODY_WIDTHGOOGLE_LIST_INDENT BYPASS_TABLES IGNORE_TABLESSINGLE_LINE_BREAK UNICODE_SNOBUSE_AUTOMATIC_LINKSSKIP_INTERNAL_LINKSLINKS_EACH_PARAGRAPH MARK_CODE DECODE_ERRORS OPEN_QUOTE CLOSE_QUOTEjoinmapstrrINCLUDE_SUP_SUB parse_argsrCopenreadsysstdinbufferdecoderFr>UnicodeDecodeErrorrrrprintrr/ ul_item_markr0 emphasis_mark strong_markr1r3google_list_indentr%r'r(r)r*r+r,r-r.r4r5r6r7r8r&r9r:r;r<r=r!r#r$rrr?r@rGstdoutwritehandle) rJrpargsfpdatahtmlerrwarninghs rmainrs>G !ANN (( B  NN !! < NN !! 1 NN && 0 NN ""  NN && 8 NN ## @ NN %% 5 NN "** * NN $$ R NN $$ 6 NN %% !  NN $$ 5 NN ''   NN   7 NN   E NN   L NN   !! B NN   )) ; NN  ! V NN  -  NN $$ A NN $$ S NN (( 4  NN ## . NN "** = NN" "** ) NN #++ A NN   = NN $$ NN !! 0  NN "" 1  NNIsxxC8M/NNN:SN)NN:S'N:NN && +  <<>D }}#- $-- &"779D' &yy$$&{{4==$*<*<= '"A   ??AL++A,,A&&AN 44A((AO((AO**A((AO..A??AL22A$$AM((AO((AO00A&&AN&&AN 44A 44A!66A..AK??AL,,A$$AM??AL00A??AL$$AM,,AJJQXXd^$k' & //J.=;00+gll:XEE g s%&_5)&`5_? a  Aaa )returnN)rMrsrrrrrrrrrs ,,C%r