Class: SiSU_XML_Munge::Trans
- Inherits:
-
Object
- Object
- SiSU_XML_Munge::Trans
- Includes:
- SiSU_Parts_XML
- Defined in:
- lib/sisu/xml_shared.rb
Instance Method Summary collapse
-
#char_enc ⇒ Object
character encode.
- #clean(str) ⇒ Object
- #default ⇒ Object
- #html(dob = '') ⇒ Object
-
#initialize(md) ⇒ Trans
constructor
A new instance of Trans.
- #markup(dob = '') ⇒ Object
- #markup_block(dob = '') ⇒ Object
-
#markup_docbook(dob = '') ⇒ Object
work on, initially a copy of fictionbook!.
- #markup_fictionbook(str = '', is = '') ⇒ Object
- #markup_group(dob = '') ⇒ Object
- #markup_light(dob = '') ⇒ Object
- #semantic_tags ⇒ Object
- #tidywords(wordlist) ⇒ Object
- #utf8(dob = '') ⇒ Object
-
#xml_sem_block_paired(matched) ⇒ Object
colon depth: many, recurs.
- #xml_semantic_tags(dob) ⇒ Object
Methods included from SiSU_Parts_XML
#_url_path_image_base, #band1, #band2, #banner_band, #black, #blue_ink, #blue_tinge, #grey, #grey_medium, #grey_pale, #home_button_only, #ico, #png_home, #png_home_button, #set_face, #set_fonts, #table1, #table2, #the_banner, #the_color, #the_font, #the_line_break, #the_png, #the_table_close, #the_url_decoration, #txt_close, #txt_open, #white, #xml_close, #xml_open, #yellow_light
Methods included from SiSU_Parts_Generic
#footer_signature, #home, #home_txt, #i_choice, #i_home_button, #i_ico, #i_new, #rl_root, #root_http, #sisu, #sisu_txt, #sisudoc, #site, #the_icon, #the_text, #the_url, #txt_home, #txt_hp, #txt_hp_alias, #txt_signature, #urify, #url_close, #url_open
Constructor Details
Instance Method Details
#char_enc ⇒ Object
character encode
111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 |
# File 'lib/sisu/xml_shared.rb', line 111 def char_enc #character encode def utf8(dob='') if @sys.locale =~/utf-?8/i # instead ucs for utf8 # String#encode Iñtërnâtiônàlizætiøn str=if defined? dob.obj then dob.obj elsif dob.is_a?(String) then dob end if str #¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûü #¢£¥§©ª«®°±²³µ¶¹º»¼½¾×÷ str=str.gsub(/</um,'<'). # '<' # < gsub(/>/um,'>'). # '>' # > gsub(/¢/um,'¢'). # '¢' # ¢ gsub(/£/um,'£'). # '£' # £ gsub(/¥/um,'¥'). # '¥' # ¥ gsub(/§/um,'§'). # '§' # § gsub(/©/um,'©'). # '©' # © gsub(/ª/um,'ª'). # 'ª' # ª gsub(/«/um,'«'). # '«' # « gsub(/®/um,'®'). # '®' # ® gsub(/°/um,'°'). # '°' # ° gsub(/±/um,'±'). # '±' # ± gsub(/²/um,'²'). # '²' # ² gsub(/³/um,'³'). # '³' # ³ gsub(/µ/um,'µ'). # 'µ' # µ gsub(/¶/um,'¶'). # '¶' # ¶ gsub(/¹/um,'¹'). # '¹' # ¹ gsub(/º/um,'º'). # 'º' # º gsub(/»/um,'»'). # '»' # » gsub(/¼/um,'¼'). # '¼' # ¼ gsub(/½/um,'½'). # '½' # ½ gsub(/¾/um,'¾'). # '¾' # ¾ gsub(/×/um,'×'). # '×' # × gsub(/÷/um,'÷'). # '÷' # ÷ gsub(/¿/um,'¿'). # '¿' # ¿ gsub(/À/um,'À'). # 'À' # À gsub(/Á/um,'Á'). # 'Á' # Á gsub(/Â/um,'Â'). # 'Â' #  gsub(/Ã/um,'Ã'). # 'Ã' # à gsub(/Ä/um,'Ä'). # 'Ä' # Ä gsub(/Å/um,'Å'). # 'Å' # Å gsub(/Æ/um,'Æ'). # 'Æ' # Æ gsub(/Ç/um,'Ç'). # 'Ç' # Ç gsub(/È/um,'È'). # 'È' # È gsub(/É/um,'É'). # 'É' # É gsub(/Ê/um,'Ê'). # 'Ê' # Ê gsub(/Ë/um,'Ë'). # 'Ë' # Ë gsub(/Ì/um,'Ì'). # 'Ì' # Ì gsub(/Í/um,'Í'). # 'Í' # Í gsub(/Î/um,'Î'). # 'Î' # Î gsub(/Ï/um,'Ï'). # 'Ï' # Ï gsub(/Ð/um,'Ð'). # 'Ð' # Ð gsub(/Ñ/um,'Ñ'). # 'Ñ' # Ñ gsub(/Ò/um,'Ò'). # 'Ò' # Ò gsub(/Ó/um,'Ó'). # 'Ó' # Ó gsub(/Ô/um,'Ô'). # 'Ô' # Ô gsub(/Õ/um,'Õ'). # 'Õ' # Õ gsub(/Ö/um,'Ö'). # 'Ö' # Ö gsub(/Ø/um,'Ø'). # 'Ø' # Ø gsub(/Ù/um,'Ù'). # 'Ù' # Ù gsub(/Ú/um,'Ú'). # 'Ú' # Ú gsub(/Û/um,'Û'). # 'Û' # Û gsub(/Ü/um,'Ü'). # 'Ü' # Ü gsub(/Ý/um,'Ý'). # 'Ý' # Ý gsub(/Þ/um,'Þ'). # 'Þ' # Þ gsub(/ß/um,'ß'). # 'ß' # ß gsub(/à/um,'à'). # 'à' # à gsub(/á/um,'á'). # 'á' # á gsub(/â/um,'â'). # 'â' # â gsub(/ã/um,'ã'). # 'ã' # ã gsub(/ä/um,'ä'). # 'ä' # ä gsub(/å/um,'å'). # 'å' # å gsub(/æ/um,'æ'). # 'æ' # æ gsub(/ç/um,'ç'). # 'ç' # ç gsub(/è/um,'è'). # 'è' # è gsub(/é/um,'é'). # '´' # é gsub(/ê/um,'ê'). # 'ˆ' # ê gsub(/ë/um,'ë'). # 'ë' # ë gsub(/ì/um,'ì'). # 'ì' # ì gsub(/í/um,'í'). # '´' # í gsub(/î/um,'î'). # 'î' # î gsub(/ï/um,'ï'). # 'ï' # ï gsub(/ð/um,'ð'). # 'ð' # ð gsub(/ñ/um,'ñ'). # 'ñ' # ñ gsub(/ò/um,'ò'). # 'ò' # ò gsub(/ó/um,'ó'). # 'ó' # ó gsub(/ô/um,'ô'). # 'ô' # ô gsub(/õ/um,'õ'). # 'õ' # õ gsub(/ö/um,'ö'). # 'ö' # ö gsub(/ø/um,'ø'). # 'ø' # ø gsub(/ù/um,'ú'). # 'ù' # ú gsub(/ú/um,'û'). # 'ú' # û gsub(/û/um,'ü'). # 'û' # ü gsub(/ü/um,'ý'). # 'ü' # ý gsub(/þ/um,'þ'). # 'þ' # þ gsub(/ÿ/um,'ÿ'). # 'ÿ' # ÿ gsub(/‘/um,'‘'). # '‘' # ‘ gsub(/’/um,'’'). # '’' # ’ gsub(/“/um,'“'). # “ # “ gsub(/”/um,'”'). # ” # ” gsub(/–/um,'–'). # – # – gsub(/—/um,'—'). # — # — gsub(/∝/um,'∝'). # ∝ # ∝ gsub(/∞/um,'∞'). # ∞ # ∞ gsub(/™/um,'™'). # ™ # ™ gsub(/✠/um,'✠'). # ✗ # ✠ gsub(/ /um,' '). # space identify gsub(/ /um,' ') # space identify end dob=if defined? dob.obj dob.obj=str dob elsif dob.is_a?(String) str end dob end end def html(dob='') if @sys.locale =~/utf-?8/i # instead ucs for utf8 # String#encode Iñtërnâtiônàlizætiøn dob.obj=dob.obj.gsub(/ /u,' '). # space identify gsub(/ /u,' ') # space identify end end self end |
#clean(str) ⇒ Object
335 336 337 338 |
# File 'lib/sisu/xml_shared.rb', line 335 def clean(str) str=str.gsub(/#{Mx[:gl_o]}(#[0-9]{3})#{Mx[:gl_c]}/u,'&\1;'). gsub(/#{Mx[:gl_o]}#([a-z]{2,4})#{Mx[:gl_c]}/u,'&\1;') end |
#default ⇒ Object
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 |
# File 'lib/sisu/xml_shared.rb', line 72 def default { pub: 'publication', conv: 'convention', vol: 'volume', pg: 'page', cty: 'city', org: 'organization', uni: 'university', dept: 'department', fac: 'faculty', inst: 'institute', co: 'company', com: 'company', conv: 'convention', dt: 'date', y: 'year', m: 'month', d: 'day', ti: 'title', au: 'author', ed: 'editor', #editor? v: 'version', #edition n: 'name', fn: 'firstname', mn: 'middlename', ln: 'lastname', in: 'initials', qt: 'quote', ct: 'cite', ref: 'reference', ab: 'abreviation', def: 'define', desc: 'description', trans: 'translate', } end |
#html(dob = '') ⇒ Object
228 229 230 231 232 233 |
# File 'lib/sisu/xml_shared.rb', line 228 def html(dob='') if @sys.locale =~/utf-?8/i # instead ucs for utf8 # String#encode Iñtërnâtiônàlizætiøn dob.obj=dob.obj.gsub(/ /u,' '). # space identify gsub(/ /u,' ') # space identify end end |
#markup(dob = '') ⇒ Object
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 |
# File 'lib/sisu/xml_shared.rb', line 246 def markup(dob='') wordlist=dob.obj.scan(/&[#0-9a-z]+;|\S+|\n/) #\n needed for tables, check though added 2005w17 dob.obj=tidywords(wordlist).join(' ').strip unless dob.is==:table dob.obj=dob.obj.gsub(/#{Mx[:br_line]}/u,'<br />'). gsub(/#{Mx[:br_paragraph]}/u,'<br />'). gsub(/#{Mx[:br_line]}|#{Mx[:br_nl]}/,'<br />') end dob.obj=dob.obj.gsub(/#{Mx[:mk_o]}:name#\S+?#{Mx[:mk_c]}/,''). gsub(/#{Mx[:mk_o]}#([a-zA-Z]+)#{Mx[:mk_c]}/,'&\1;'). gsub(/#{Mx[:mk_o]}(#[0-9]+)#{Mx[:mk_c]}/,'&\1;'). gsub(/(^|#{Mx[:gl_c]}|\s+)<\s+/,'\1< ').gsub(/\s+>(\s+|$)/,' >\1'). #gsub(/#{Mx[:fa_emphasis_o]}(.+?)#{Mx[:fa_emphasis_c]}/,'<em>\1</em>'). #reinstate gsub(/#{Mx[:fa_bold_o]}(.+?)#{Mx[:fa_bold_c]}/m,'<b>\1</b>'). gsub(/#{Mx[:fa_italics_o]}(.+?)#{Mx[:fa_italics_c]}/m,'<i>\1</i>'). gsub(/#{Mx[:fa_underscore_o]}(.+?)#{Mx[:fa_underscore_c]}/,'<u>\1</u>'). gsub(/#{Mx[:fa_superscript_o]}(.+?)#{Mx[:fa_superscript_c]}/,'<sup>\1</sup>'). gsub(/#{Mx[:fa_subscript_o]}(.+?)#{Mx[:fa_subscript_c]}/,'<sub>\1</sub>'). gsub(/#{Mx[:fa_insert_o]}(.+?)#{Mx[:fa_insert_c]}/,'<ins>\1</ins>'). gsub(/#{Mx[:fa_cite_o]}(.+?)#{Mx[:fa_cite_c]}/,'<cite>\1</cite>'). gsub(/#{Mx[:fa_strike_o]}(.+?)#{Mx[:fa_strike_c]}/,'<del>\1</del>'). gsub(/#{Mx[:fa_monospace_o]}(.+?)#{Mx[:fa_monospace_c]}/,'<tt>\1</tt>'). gsub(/<:pb>\s*/,''). #Fix gsub(/<+[-~]#>+/,'') if dob.is !=:code #embeds a red-bullet image --> dob.obj=dob.obj.gsub(/#{Mx[:fa_bold_o]}(.+?)#{Mx[:fa_bold_c]}/,'<b>\1</b>'). gsub(/#{Mx[:fa_italics_o]}(.+?)#{Mx[:fa_italics_c]}/,'<i>\1</i>'). gsub(/#{Mx[:fa_underscore_o]}(.+?)#{Mx[:fa_underscore_c]}/,'<u>\1</u>'). gsub(/#{Mx[:fa_strike_o]}(.+?)#{Mx[:fa_strike_c]}/,'<del>\1</del>') dob.obj=dob.obj.gsub(/#{Mx[:br_line]}|#{Mx[:br_nl]}/,'<br />') unless dob.is==:table dob.obj=dob.obj.gsub(/#{Mx[:br_page]}\s*/,''). gsub(/#{Mx[:br_page_new]}\s*/,''). gsub(/#{Mx[:br_page_line]}\s*/,''). gsub(/#{Mx[:pa_non_object_no_heading]}|#{Mx[:pa_non_object_dummy_heading]}/,''). gsub(/<[-~]#>/,''). gsub(/href="#{Xx[:segment]}/m,'href="'). gsub(/#{Mx[:lnk_o]}([^#{Mx[:lnk_o]}#{Mx[:lnk_c]}#{Mx[:rel_o]}#{Mx[:rel_c]}]+?)#{Mx[:lnk_c]}#{Mx[:rel_o]}(\.\.\/\S+?)#{Mx[:rel_c]}/, '<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="\2">\1</link>'). gsub(/#{Mx[:lnk_o]}([^#{Mx[:lnk_o]}#{Mx[:lnk_c]}#{Mx[:rel_o]}#{Mx[:rel_c]}]+?)#{Mx[:lnk_c]}#{Mx[:rel_o]}:(\S+?)#{Mx[:rel_c]}/, '<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="../\2">\1</link>'). gsub(/#{Mx[:lnk_o]}(.+?)#{Mx[:lnk_c]}#{Mx[:rel_o]}(\S+?)#{Mx[:rel_c]}/, '<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="#\2">\1</link>'). gsub(/(?:^|[^_\\])#{Mx[:lnk_o]}[ ]*(\S+?\.(?:jpg|png|gif))[ ]+(\d+)x(\d+)(\s+[^}]+)?#{Mx[:lnk_c]}#{Mx[:url_o]}(\S+?)#{Mx[:url_c]}/, %{<image xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:actuate="onLoad" xl:show="embed" xl:href="#{@md.file.output_path.xml.rel_image}/\\1" width="\\2" height="\\3" />[\\1] \\4}). gsub(/(?:^|[^_\\])#{Mx[:lnk_o]}[ ]*(\S+?\.(?:jpg|png|gif))([ ]+[^}]+)?#{Mx[:lnk_c]}#{Mx[:url_o]}(\S+?)#{Mx[:url_c]}/, %{<image xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:actuate="onLoad" xl:show="embed" xl:href="#{@md.file.output_path.xml.rel_image}/\\1"/>\\1}). gsub(/(?:^|[^_\\])#{Mx[:lnk_o]}[ ]*(\S+?\.(?:jpg|png|gif))[ ]+(\d+)x(\d+)(\s+[^}]+)?#{Mx[:lnk_c]}image/, %{<image xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:actuate="onLoad" xl:show="embed" xl:href="#{@md.file.output_path.xml.rel_image}/\\1" width="\\2" height="\\3" />[\\1] \\4}). gsub(/(?:^|[^_\\])#{Mx[:lnk_o]}[ ]*(\S+?\.(?:jpg|png|gif))([ ]+[^}]+)?#{Mx[:lnk_c]}image/, %{<image xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:actuate="onLoad" xl:show="embed" xl:href="#{@md.file.output_path.xml.rel_image}/\\1"/>\\1}). gsub(/#{Mx[:lnk_o]}(.+?)#{Mx[:lnk_c]}#{Mx[:url_o]}(\S+?)#{Mx[:url_c]}/, '<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="\2">\1</link>'). #watch, compare html_tune gsub(/#{Mx[:url_o]}(\S+?)#{Mx[:url_c]}/, %{#{the_url_decoration.xml_open}<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="\\1">\\1</link>#{the_url_decoration.xml_close}}). gsub(/#{Mx[:url_o]}_(\S+?)#{Mx[:url_c]}/, '<link xmlns:xl="http://www.w3.org/1999/xlink" xl:type="simple" xl:href="\1">\1</link>') #escaped urls not linked, deal with later else dob.obj=dob.obj.gsub(/</m,'<').gsub(/>/m,'>') end if dob.of==:block dob.obj=dob.obj.gsub(/#{Mx[:gl_bullet]}/,'● ') end dob.obj=dob.obj.gsub(/#{Mx[:url_o]}([a-zA-Z0-9._-]+\@\S+?\.[a-zA-Z0-9._-]+)#{Mx[:url_c]}/, %{#{the_url_decoration.xml_open}\\1#{the_url_decoration.xml_close}}). gsub(/#{Dx[:url_o]}/,"#{Dx[:url_o_xml]}"). gsub(/#{Dx[:url_c]}/,"#{Dx[:url_c_xml]}"). gsub(/ |#{Mx[:nbsp]}/m,' '). gsub(/;&([^#]|(?:[^gl][^t]|[^a][^m][^p]|[^n][^b][^s][^p])[^;])/,';&\1') # pattern not to match dob end |
#markup_block(dob = '') ⇒ Object
407 408 409 410 411 412 413 414 |
# File 'lib/sisu/xml_shared.rb', line 407 def markup_block(dob='') dob.obj=dob.obj.gsub(/</,'<').gsub(/>/,'>'). gsub(/<:?br(?:\s+\/)?>/,'<br />'). gsub(/<(link xmlns:xl=".+?")>/,'<\1>'). gsub(/<(\/link)>/,'<\1>'). gsub(/<(\/?en)>/,'<\1>') dob end |
#markup_docbook(dob = '') ⇒ Object
work on, initially a copy of fictionbook!
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 |
# File 'lib/sisu/xml_shared.rb', line 363 def markup_docbook(dob='') # work on, initially a copy of fictionbook! if dob.is !=:code dob.obj=dob.obj.gsub(/#{Mx[:en_a_o]}(\d+)\s*(.+?)#{Mx[:en_a_c]}/m,'<footnote><para><!-- fn\1 -->\2</para></footnote>'). gsub(/\\\\/,'</para><para>'). gsub(/&/,'&'). #sort gsub(/#{Mx[:mk_o]}#([a-zA-Z]+)#{Mx[:mk_c]}/,'&\1;'). gsub(/(^|#{Mx[:gl_c]}|\s)&\s+/,'\1& '). #sort gsub(/#{Mx[:mk_o]}(#[0-9]+)#{Mx[:mk_c]}/,'&\1;') dob.obj=dob.obj.gsub(/#{Mx[:br_line]}|#{Mx[:br_nl]}/,'<br />') unless dob.is==:table dob.obj=dob.obj.gsub(/#{Mx[:fa_bold_o]}(.+?)#{Mx[:fa_bold_c]}/,'<b>\1</b>'). gsub(/#{Mx[:fa_italics_o]}(.+?)#{Mx[:fa_italics_c]}/,'<i>\1</i>'). gsub(/#{Mx[:fa_underscore_o]}(.+?)#{Mx[:fa_underscore_c]}/,'<u>\1</u>'). gsub(/#{Mx[:fa_superscript_o]}(.+?)#{Mx[:fa_superscript_c]}/,'<sup>\1</sup>'). gsub(/#{Mx[:fa_subscript_o]}(.+?)#{Mx[:fa_subscript_c]}/,'<sub>\1</sub>'). gsub(/#{Mx[:fa_insert_o]}(.+?)#{Mx[:fa_insert_c]}/,'<ins>\1</ins>'). gsub(/#{Mx[:fa_cite_o]}(.+?)#{Mx[:fa_cite_c]}/,'<cite>\1</cite>'). gsub(/#{Mx[:fa_strike_o]}(.+?)#{Mx[:fa_strike_c]}/,'<del>\1</del>'). gsub(/#{Mx[:fa_monospace_o]}(.+?)#{Mx[:fa_monospace_c]}/,'<tt>\1</tt>'). # tt, kbd gsub(/#{Mx[:lnk_o]}\s*(\S+?)\.(png|jpg|gif).+?#{Mx[:lnk_c]}(?:#{Mx[:url_o]}\S+?#{Mx[:url_c]}|#{Mx[:rel_o]}\S+?#{Mx[:rel_c]}|image)/m, %{#{Xx[:split]}:spaces0:<figure id="fig-\\1">\n:spaces1:<title></title>\n:spaces1:<graphic fileref="../../_sisu/image/\\1.\\2" align="center" width="50%"></graphic>\n:spaces0:</figure>#{Xx[:split]}}). # common image location, else use ./images gsub(/#{Mx[:lnk_o]}(.+?)#{Mx[:lnk_c]}#{Mx[:url_o]}(.+?)#{Mx[:url_c]}/, '<ulink url="\2">\1</ulink>'). gsub(/#{Mx[:url_o]}(.+?)#{Mx[:url_c]}/, '<ulink url="\1">\1</ulink>'). gsub(/#{Mx[:mk_o]}:name#(\S+?)#{Mx[:mk_c]}/,'<a name="\1"></a>'). gsub(/#{Mx[:gl_bullet]}/m,'● '). # not available gsub(/#{Mx[:nbsp]}/,' '). # not available gsub(/<(p|br)>/,'<\1 />') dob.obj=clean(dob.obj) elsif dob.is == :code dob.obj=dob.obj.gsub(/&/m,'&'). #sort gsub(/</,'<').gsub(/>/,'>') else # p dob.is ?? end dob end |
#markup_fictionbook(str = '', is = '') ⇒ Object
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 |
# File 'lib/sisu/xml_shared.rb', line 339 def markup_fictionbook(str='',is='') str=str.gsub(/#{Mx[:en_a_o]}([\d+*]+).+?#{Mx[:en_a_c]}/m,'<a xl:href="#footnote\1" type="note">[\1]</a>'). gsub(/&/,'&'). #sort gsub(/#{Mx[:mk_o]}#([a-zA-Z]+)#{Mx[:mk_c]}/,'&\1;'). gsub(/(^|#{Mx[:gl_c]}|\s)&\s+/,'\1& '). #sort gsub(/#{Mx[:mk_o]}(#[0-9]+)#{Mx[:mk_c]}/,'&\1;') str=str.gsub(/#{Mx[:br_line]}|#{Mx[:br_nl]}/,'<br />') unless is==:table str=str.gsub(/#{Mx[:fa_bold_o]}(.+?)#{Mx[:fa_bold_c]}/,'<b>\1</b>'). gsub(/#{Mx[:fa_italics_o]}(.+?)#{Mx[:fa_italics_c]}/,'<i>\1</i>'). gsub(/#{Mx[:fa_underscore_o]}(.+?)#{Mx[:fa_underscore_c]}/,'<u>\1</u>'). gsub(/#{Mx[:fa_superscript_o]}(.+?)#{Mx[:fa_superscript_c]}/,'<sup>\1</sup>'). gsub(/#{Mx[:fa_subscript_o]}(.+?)#{Mx[:fa_subscript_c]}/,'<sub>\1</sub>'). gsub(/#{Mx[:fa_insert_o]}(.+?)#{Mx[:fa_insert_c]}/,'<ins>\1</ins>'). gsub(/#{Mx[:fa_cite_o]}(.+?)#{Mx[:fa_cite_c]}/,'<cite>\1</cite>'). gsub(/#{Mx[:fa_strike_o]}(.+?)#{Mx[:fa_strike_c]}/,'<del>\1</del>'). gsub(/#{Mx[:fa_monospace_o]}(.+?)#{Mx[:fa_monospace_c]}/,'<tt>\1</tt>'). # tt, kbd gsub(/#{Mx[:lnk_o]}\s*(\S+?\.(?:png|jpg|gif)).+?#{Mx[:lnk_c]}(?:#{Mx[:url_o]}\S+?#{Mx[:url_c]}|#{Mx[:rel_o]}\S+?#{Mx[:rel_c]}|image)/m,'<image xl:href="#\1" />'). gsub(/#{Mx[:url_o]}(.+?)#{Mx[:url_c]}/,"#{Dx[:url_o]}\\1#{Dx[:url_c]}"). gsub(/#{Mx[:mk_o]}:name#(\S+?)#{Mx[:mk_c]}/,'<a name="\1"></a>'). gsub(/#{Mx[:gl_bullet]}/m,'● '). # not available gsub(/#{Mx[:nbsp]}/,' '). # not available gsub(/<(p|br)>/,'<\1 />') clean(str) end |
#markup_group(dob = '') ⇒ Object
399 400 401 402 403 404 405 406 |
# File 'lib/sisu/xml_shared.rb', line 399 def markup_group(dob='') dob.obj=dob.obj.gsub(/</,'<').gsub(/>/,'>'). gsub(/<:?br(?:\s+\/)?>/,'<br />'). gsub(/<(link xmlns:xl=".+?")>/,'<\1>'). gsub(/<(\/link)>/,'<\1>'). gsub(/<(\/?en)>/,'<\1>') dob end |
#markup_light(dob = '') ⇒ Object
317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 |
# File 'lib/sisu/xml_shared.rb', line 317 def markup_light(dob='') dob.obj=dob.obj.gsub(/\/\{(.+?)\}\//,'<i>\1</i>'). gsub(/[*!]\{(.+?)\}[*!]/,'<b>\1</b>'). gsub(/_\{(.+?)\}_/,'<u>\1</u>'). gsub(/-\{(.+?)\}-/,'<del>\1</del>'). gsub(/<br(\s*\/)?>/,'<br />'). gsub(/<:pb>\s*/,''). gsub(/<[-~]#>/,''). gsub(/(^|#{Mx[:gl_c]}|\s)&\s+/,'\1& '). #sort gsub(/&([^;]{1,5})/,'&\1'). #sort, rough estimate, revisit #WATCH found in node not sax gsub(/(?:^|[^_\\])#{Mx[:lnk_o]}(\S+?\.(?:png|jpg|gif))[ ]+.+?#{Mx[:lnk_c]}(?:#{Mx[:url_o]}\S+?#{Mx[:url_c]}|image)/, "<image.path>#{@md.file.output_path.xml.rel_image}\/\\1</image.path>"). gsub(/ |#{Mx[:nbsp]}/,' '). gsub(/;&([^#]|(?:[^gl][^t]|[^a][^m][^p]|[^n][^b][^s][^p])[^;])/,';&\1') # pattern not to match wordlist=dob.obj.scan(/&[#0-9a-z]+;|\S+|\n/) #\n needed for tables, check though added 2005w17 dob.obj=tidywords(wordlist).join(' ').strip dob end |
#semantic_tags ⇒ Object
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 |
# File 'lib/sisu/xml_shared.rb', line 71 def def default { pub: 'publication', conv: 'convention', vol: 'volume', pg: 'page', cty: 'city', org: 'organization', uni: 'university', dept: 'department', fac: 'faculty', inst: 'institute', co: 'company', com: 'company', conv: 'convention', dt: 'date', y: 'year', m: 'month', d: 'day', ti: 'title', au: 'author', ed: 'editor', #editor? v: 'version', #edition n: 'name', fn: 'firstname', mn: 'middlename', ln: 'lastname', in: 'initials', qt: 'quote', ct: 'cite', ref: 'reference', ab: 'abreviation', def: 'define', desc: 'description', trans: 'translate', } end self end |
#tidywords(wordlist) ⇒ Object
236 237 238 239 240 241 242 243 244 245 |
# File 'lib/sisu/xml_shared.rb', line 236 def tidywords(wordlist) wordlist_new=[] wordlist.each do |x| #imperfect solution will not catch all possible cases x=x.gsub(/&/,'&') unless x =~/&\S+;/ x=x.gsub(/&([A-Z])/,'&\1') wordlist_new << x end wordlist_new end |
#utf8(dob = '') ⇒ Object
112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 |
# File 'lib/sisu/xml_shared.rb', line 112 def utf8(dob='') if @sys.locale =~/utf-?8/i # instead ucs for utf8 # String#encode Iñtërnâtiônàlizætiøn str=if defined? dob.obj then dob.obj elsif dob.is_a?(String) then dob end if str #¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûü #¢£¥§©ª«®°±²³µ¶¹º»¼½¾×÷ str=str.gsub(/</um,'<'). # '<' # < gsub(/>/um,'>'). # '>' # > gsub(/¢/um,'¢'). # '¢' # ¢ gsub(/£/um,'£'). # '£' # £ gsub(/¥/um,'¥'). # '¥' # ¥ gsub(/§/um,'§'). # '§' # § gsub(/©/um,'©'). # '©' # © gsub(/ª/um,'ª'). # 'ª' # ª gsub(/«/um,'«'). # '«' # « gsub(/®/um,'®'). # '®' # ® gsub(/°/um,'°'). # '°' # ° gsub(/±/um,'±'). # '±' # ± gsub(/²/um,'²'). # '²' # ² gsub(/³/um,'³'). # '³' # ³ gsub(/µ/um,'µ'). # 'µ' # µ gsub(/¶/um,'¶'). # '¶' # ¶ gsub(/¹/um,'¹'). # '¹' # ¹ gsub(/º/um,'º'). # 'º' # º gsub(/»/um,'»'). # '»' # » gsub(/¼/um,'¼'). # '¼' # ¼ gsub(/½/um,'½'). # '½' # ½ gsub(/¾/um,'¾'). # '¾' # ¾ gsub(/×/um,'×'). # '×' # × gsub(/÷/um,'÷'). # '÷' # ÷ gsub(/¿/um,'¿'). # '¿' # ¿ gsub(/À/um,'À'). # 'À' # À gsub(/Á/um,'Á'). # 'Á' # Á gsub(/Â/um,'Â'). # 'Â' #  gsub(/Ã/um,'Ã'). # 'Ã' # à gsub(/Ä/um,'Ä'). # 'Ä' # Ä gsub(/Å/um,'Å'). # 'Å' # Å gsub(/Æ/um,'Æ'). # 'Æ' # Æ gsub(/Ç/um,'Ç'). # 'Ç' # Ç gsub(/È/um,'È'). # 'È' # È gsub(/É/um,'É'). # 'É' # É gsub(/Ê/um,'Ê'). # 'Ê' # Ê gsub(/Ë/um,'Ë'). # 'Ë' # Ë gsub(/Ì/um,'Ì'). # 'Ì' # Ì gsub(/Í/um,'Í'). # 'Í' # Í gsub(/Î/um,'Î'). # 'Î' # Î gsub(/Ï/um,'Ï'). # 'Ï' # Ï gsub(/Ð/um,'Ð'). # 'Ð' # Ð gsub(/Ñ/um,'Ñ'). # 'Ñ' # Ñ gsub(/Ò/um,'Ò'). # 'Ò' # Ò gsub(/Ó/um,'Ó'). # 'Ó' # Ó gsub(/Ô/um,'Ô'). # 'Ô' # Ô gsub(/Õ/um,'Õ'). # 'Õ' # Õ gsub(/Ö/um,'Ö'). # 'Ö' # Ö gsub(/Ø/um,'Ø'). # 'Ø' # Ø gsub(/Ù/um,'Ù'). # 'Ù' # Ù gsub(/Ú/um,'Ú'). # 'Ú' # Ú gsub(/Û/um,'Û'). # 'Û' # Û gsub(/Ü/um,'Ü'). # 'Ü' # Ü gsub(/Ý/um,'Ý'). # 'Ý' # Ý gsub(/Þ/um,'Þ'). # 'Þ' # Þ gsub(/ß/um,'ß'). # 'ß' # ß gsub(/à/um,'à'). # 'à' # à gsub(/á/um,'á'). # 'á' # á gsub(/â/um,'â'). # 'â' # â gsub(/ã/um,'ã'). # 'ã' # ã gsub(/ä/um,'ä'). # 'ä' # ä gsub(/å/um,'å'). # 'å' # å gsub(/æ/um,'æ'). # 'æ' # æ gsub(/ç/um,'ç'). # 'ç' # ç gsub(/è/um,'è'). # 'è' # è gsub(/é/um,'é'). # '´' # é gsub(/ê/um,'ê'). # 'ˆ' # ê gsub(/ë/um,'ë'). # 'ë' # ë gsub(/ì/um,'ì'). # 'ì' # ì gsub(/í/um,'í'). # '´' # í gsub(/î/um,'î'). # 'î' # î gsub(/ï/um,'ï'). # 'ï' # ï gsub(/ð/um,'ð'). # 'ð' # ð gsub(/ñ/um,'ñ'). # 'ñ' # ñ gsub(/ò/um,'ò'). # 'ò' # ò gsub(/ó/um,'ó'). # 'ó' # ó gsub(/ô/um,'ô'). # 'ô' # ô gsub(/õ/um,'õ'). # 'õ' # õ gsub(/ö/um,'ö'). # 'ö' # ö gsub(/ø/um,'ø'). # 'ø' # ø gsub(/ù/um,'ú'). # 'ù' # ú gsub(/ú/um,'û'). # 'ú' # û gsub(/û/um,'ü'). # 'û' # ü gsub(/ü/um,'ý'). # 'ü' # ý gsub(/þ/um,'þ'). # 'þ' # þ gsub(/ÿ/um,'ÿ'). # 'ÿ' # ÿ gsub(/‘/um,'‘'). # '‘' # ‘ gsub(/’/um,'’'). # '’' # ’ gsub(/“/um,'“'). # “ # “ gsub(/”/um,'”'). # ” # ” gsub(/–/um,'–'). # – # – gsub(/—/um,'—'). # — # — gsub(/∝/um,'∝'). # ∝ # ∝ gsub(/∞/um,'∞'). # ∞ # ∞ gsub(/™/um,'™'). # ™ # ™ gsub(/✠/um,'✠'). # ✗ # ✠ gsub(/ /um,' '). # space identify gsub(/ /um,' ') # space identify end dob=if defined? dob.obj dob.obj=str dob elsif dob.is_a?(String) str end dob end end |
#xml_sem_block_paired(matched) ⇒ Object
colon depth: many, recurs
415 416 417 418 419 420 421 422 423 424 425 426 427 428 |
# File 'lib/sisu/xml_shared.rb', line 415 def xml_sem_block_paired(matched) # colon depth: many, recurs matched=matched.gsub(/\b(au):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:au]} depth="many">\\2</sem:#{@ab[:au]}>}). gsub(/\b(vol):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:vol]} depth="many">\\2</sem:#{@ab[:vol]}>}). gsub(/\b(pub):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:pub]} depth="many">\\2</sem:#{@ab[:pub]}>}). gsub(/\b(ref):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:ref]} depth="many">\\2</sem:#{@ab[:ref]}>}). gsub(/\b(desc):\{(.+?)\}:\1\b/m,%{<sem:#{@ab[:desc]} depth="many">\\2</sem:#{@ab[:desc]}>}). gsub(/\b(conv):\{(.+?)\}:\1\b/m,%{<sem:#{@ab[:conv]} depth="many">\\2</sem:#{@ab[:conv]}>}). gsub(/\b(ct):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:ct]} depth="many">\\2</sem:#{@ab[:ct]}>}). gsub(/\b(cty):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:cty]} depth="many">\\2</sem:#{@ab[:cty]}>}). gsub(/\b(org):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:org]} depth="many">\\2</sem:#{@ab[:org]}>}). gsub(/\b(dt):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:dt]} depth="many">\\2</sem:#{@ab[:dt]}>}). gsub(/\b(n):\{(.+?)\}:\1\b/m, %{<sem:#{@ab[:n]} depth="many">\\2</sem:#{@ab[:n]}>}). gsub(/([a-z]+(?:[_:.][a-z]+)*)(?::\{(.+?)\}:\1)/m,'<sem:\1 depth="many">\2</sem:\1>') end |
#xml_semantic_tags(dob) ⇒ Object
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 |
# File 'lib/sisu/xml_shared.rb', line 429 def (dob) if @md.sem_tag dob.obj.gsub!(/([a-z]+(?:[_:.][a-z]+)*)(?::\{(.+?)\}:\1)/m) {|c| xml_sem_block_paired(c) } dob.obj.gsub!(/([a-z]+(?:[_:.][a-z]+)*)(?::\{(.+?)\}:\1)/m) {|c| xml_sem_block_paired(c) } dob.obj.gsub!(/([a-z]+(?:[_:.][a-z]+)*)(?::\{(.+?)\}:\1)/m) {|c| xml_sem_block_paired(c) } dob.obj=dob.obj.gsub(/:\{(.+?)\}:au\b/m, %{<sem:#{@ab[:au]} depth="one">\\1</sem:#{@ab[:au]}>}). gsub(/:\{(.+?)\}:n\b/m, %{<sem:#{@ab[:n]} depth="one">\\1</sem:#{@ab[:n]}>}). gsub(/:\{(.+?)\}:ti\b/m, %{<sem:#{@ab[:ti]} depth="one">\\1</sem:#{@ab[:ti]}>}). gsub(/:\{(.+?)\}:ref\b/m, %{<sem:#{@ab[:ref]} depth="one">\\1</sem:#{@ab[:ref]}>}). gsub(/:\{(.+?)\}:desc\b/m, %{<sem:#{@ab[:desc]} depth="one">\\1</sem:#{@ab[:desc]}>}). gsub(/:\{(.+?)\}:cty\b/m, %{<sem:#{@ab[:cty]} depth="one">\\1</sem:#{@ab[:cty]}>}). gsub(/:\{(.+?)\}:org\b/m, %{<sem:#{@ab[:org]} depth="one">\\1</sem:#{@ab[:org]}>}). gsub(/:\{(.+?)\}:([a-z]+(?:[_:.][a-z]+)*)/m,'<sem:\2 depth="one">\1</sem:\2>'). gsub(/;\{([^}]+(?![;]))\};ti\b/m, %{<sem:#{@ab[:ti]} depth="zero">\\1</sem:#{@ab[:ti]}>}). gsub(/;\{([^}]+(?![;]))\};qt\b/m, %{<sem:#{@ab[:qt]} depth="zero">\\1</sem:#{@ab[:qt]}>}). gsub(/;\{([^}]+(?![;]))\};ref\b/m, %{<sem:#{@ab[:ref]} depth="zero">\\1</sem:#{@ab[:ref]}>}). gsub(/;\{([^}]+(?![;]))\};ed\b/m, %{<sem:#{@ab[:ed]} depth="zero">\\1</sem:#{@ab[:ed]}>}). gsub(/;\{([^}]+(?![;]))\};v\b/m, %{<sem:#{@ab[:v]} depth="zero">\\1</sem:#{@ab[:v]}>}). gsub(/;\{([^}]+(?![;]))\};desc\b/m, %{<sem:#{@ab[:desc]} depth="zero">\\1</sem:#{@ab[:desc]}>}). gsub(/;\{([^}]+(?![;]))\};def\b/m, %{<sem:#{@ab[:def]} depth="zero">\\1</sem:#{@ab[:def]}>}). gsub(/;\{([^}]+(?![;]))\};trans\b/m, %{<sem:#{@ab[:trans]} depth="zero">\\1</sem:#{@ab[:trans]}>}). gsub(/;\{([^}]+(?![;]))\};y\b/m, %{<sem:#{@ab[:y]} depth="zero">\\1</sem:#{@ab[:y]}>}). gsub(/;\{([^}]+(?![;]))\};ab\b/m, %{<sem:#{@ab[:ab]} depth="zero">\\1</sem:#{@ab[:ab]}>}). gsub(/;\{([^}]+(?![;]))\};pg\b/m, %{<sem:#{@ab[:pg]} depth="zero">\\1</sem:#{@ab[:pg]}>}). gsub(/;\{([^}]+(?![;]))\};fn?\b/m, %{<sem:#{@ab[:fn]} depth="zero">\\1</sem:#{@ab[:fn]}>}). gsub(/;\{([^}]+(?![;]))\};mn?\b/m, %{<sem:#{@ab[:mn]} depth="zero">\\1</sem:#{@ab[:mn]}>}). gsub(/;\{([^}]+(?![;]))\};ln?\b/m, %{<sem:#{@ab[:ln]} depth="zero">\\1</sem:#{@ab[:ln]}>}). gsub(/;\{([^}]+(?![;]))\};in\b/m, %{<sem:#{@ab[:in]} depth="zero">\\1</sem:#{@ab[:in]}>}). gsub(/;\{([^}]+(?![;]))\};uni\b/m, %{<sem:#{@ab[:uni]} depth="zero">\\1</sem:#{@ab[:uni]}>}). gsub(/;\{([^}]+(?![;]))\};fac\b/m, %{<sem:#{@ab[:fac]} depth="zero">\\1</sem:#{@ab[:fac]}>}). gsub(/;\{([^}]+(?![;]))\};inst\b/m, %{<sem:#{@ab[:inst]} depth="zero">\\1</sem:#{@ab[:inst]}>}). gsub(/;\{([^}]+(?![;]))\};dept\b/m, %{<sem:#{@ab[:dpt]} depth="zero">\\1</sem:#{@ab[:dept]}>}). gsub(/;\{([^}]+(?![;]))\};org\b/m, %{<sem:#{@ab[:org]} depth="zero">\\1</sem:#{@ab[:org]}>}). gsub(/;\{([^}]+(?![;]))\};com?\b/m, %{<sem:#{@ab[:com]} depth="zero">\\1</sem:#{@ab[:com]}>}). gsub(/;\{([^}]+(?![;]))\};cty\b/m, %{<sem:#{@ab[:cty]} depth="zero">\\1</sem:#{@ab[:cty]}>}). gsub(/;\{([^}]+(?![;]))\};([a-z]+(?:[_:.][a-z]+)*)/m,'<sem:\2 depth="zero">\1</sem:\2>') end dob end |