API tools faq
paste
Left4Cake2

Html Reader

Left4Cake2
Apr 11th, 2013
228
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Lua 9.14 KB | None | 0 0
raw download clone embed print report
  1. --[[
  2. Intializing
  3. ]]--
  4.  
  5. local tArgs = { ... }
  6. veiw = "page"
  7. shell.run("clear")
  8. file = io.open( tArgs[1] )
  9. split = function(s, pattern, maxsplit)
  10.   local pattern = pattern or ' '
  11.   local maxsplit = maxsplit or -1
  12.   local s = s
  13.   local t = {}
  14.   local patsz = #pattern
  15.   while maxsplit ~= 0 do
  16.         local curpos = 1
  17.         local found = string.find(s, pattern)
  18.         if found ~= nil then
  19.           table.insert(t, string.sub(s, curpos, found - 1))
  20.           curpos = found + patsz
  21.           s = string.sub(s, curpos)
  22.         else
  23.           table.insert(t, string.sub(s, curpos))
  24.           break
  25.         end
  26.         maxsplit = maxsplit - 1
  27.         if maxsplit == 0 then
  28.           table.insert(t, string.sub(s, curpos - patsz - 1))
  29.         end
  30.   end
  31.   return t
  32. end
  33.  
  34.  
  35. --[[
  36. Reading Html
  37. ]]--
  38.  
  39. -- where to find links
  40. herf = {"herf='(.-)'", "Herf='(.-)'", "HERF='(.-)'",
  41. "herf=\"(.-)\"", "Href=\"(.-)\"", "HREF=\"(.-)\"",
  42. "herf=(.-) ", "Href=(.-) ", "HREF=(.-) ",
  43. "herf=(.-)>", "Href=(.-)>", "HREF=(.-)>",
  44. "href='(.-)'", "Href='(.-)'", "HREF='(.-)'",
  45. "href=\"(.-)\"", "Herf=\"(.-)\"", "HERF=\"(.-)\"",
  46. "href=(.-)>", "Herf=(.-)>", "HERF=(.-)>",
  47. "href=(.-) ", "Herf=(.-) ", "HERF=(.-) ",
  48. "src='(.-)'", "Src='(.-)'", "SRC='(.-)'",
  49. "src=\"(.-)\"", "Src=\"(.-)\"", "SRC=\"(.-)\"",
  50. "src=(.-)>", "Src=(.-)>", "SRC=(.-)>",
  51. "src=(.-) ", "Src=(.-) ", "SRC=(.-) ",
  52. }
  53.  
  54. -- Tags to remove
  55. remove = {"<!DOCTYPE(.-)>", "<!doctype(.-)>",
  56. "<!--(.-)-->", "<!--(.-)-->",
  57. "<html(.-)>", "<HTML(.-)>", "</html>", "</HTML>",
  58. "<head>(.-)</head>", "<HEAD>(.-)</HEAD>",
  59. "<body>", "<body(.-)>", "</body>", "<BODY>", "<BODY(.-)>", "</BODY>",
  60. "<HEAD>(.-)</HEAD>",
  61. "<form(.-)>", "</form>", "<FORM(.-)>", "</FORM>",
  62. "<input(.-)>", "</input>", "<INPUT(.-)>", "</INPUT>",
  63. "<script>(.-)</script>","<SCRIPT>(.-)</SCRIPT>",
  64. "<script(.-)>", "</script>", "<SCRIPT(.-)>", "</SCRIPT>",
  65. "<span(.-)>", "</span>", "<SPAN(.-)>", "</SPAN>",
  66. "<div(.-)</div>", "<div(.-)>", "</div>",
  67. "<Iframe(.-)</Iframe>", "<iframe(.-)</iframe>",
  68. "<a(.-)>", "</a>", "<A(.-)>", "</A>",
  69. "<TH>", "<TH(.-)>", "</TH>", "<th>", "<th(.-)>", "</th>",
  70. "<img(.-)>", "</img>", "<IMG(.-)>", "</IMG>",
  71. "<TD>", "<TD(.-)>", "</TD>", "<td>", "<td(.-)>", "</td>",
  72. "<ol>", "<ol(.-)>", "</ol>", "<OL>", "<OL(.-)>", "</OL>",
  73. "<sub>", "</sub>", "<SUB>", "</SUB>",
  74. "</ul>", "</UL>",
  75. "</NAV>", "</Nav>", "</nav>",
  76. "<FOOTER>", "<FOOTER(.-)>", "</FOOTER>",
  77. "<Footer>", "<Footer(.-)>", "</Footer>",
  78. "<footer>", "<footer(.-)>", "</footer>",
  79. }
  80. -- Bullets
  81. Bullets = { "<li>", "<LI>", "<li(.-)>", "<LI(.-)>" }
  82.  
  83. -- Tags to add spaces too
  84. newline = {"<br(.-)>", "<BR(.-)>", "<br />", "<BR />", "<br/>", "<BR/>", "</br>", "</BR>",
  85. "<h1>", "<h1(.-)>", "</h1>", "<H1>", "<H1(.-)>", "</H1>",
  86. "<h2>", "<h2(.-)>", "</h2>", "<H2>", "<H2(.-)>", "</H2>",
  87. "<h3>", "<h3(.-)>",  "</h3>", "<H3>", "<H3(.-)>", "</H3>",
  88. "<h4>", "<h4(.-)>",  "</h4>", "<H4>", "<H4(.-)>", "</H4>",
  89. "<h5>", "<h5(.-)>",  "</h5>", "<H5>", "<H5(.-)>", "</H5>",
  90. "<HEADER>", "<HEADER(.-)>", "</HEADER>",
  91. "<Header>", "<Header(.-)>", "</Header>",
  92. "<header>", "<header(.-)>", "</header>",
  93. "<ARTICLE>", "<ARTICLE(.-)>", "</ARTICLE>",
  94. "<Article>", "<Article(.-)>", "</Article>",
  95. "<article>", "<article(.-)>", "</article>",
  96. "<p(.-)>", "</p>", "<P(.-)>", "</P>",
  97. "<TR(.-)>", "</TR>", "<tr(.-)>", "</tr>",
  98. "<TABLE(.-)>", "<Table(.-)>", "<table(.-)>", "</TABLE>", "</Table>", "</table>",
  99. "<ul>", "<UL>", "</li>", "</LI>"
  100. }
  101.  
  102. -- Predefing Variables
  103. htmlpage = ""
  104. links = {"None"}
  105. viewpage = {}
  106. center={}
  107. hr={}
  108. centerenabled = false
  109.  
  110. -- Reading filedata
  111. while true do
  112. filedata = file:read()
  113. if filedata then
  114.   htmlpage = htmlpage .. filedata
  115. else break end
  116. end
  117. file:close()
  118.  
  119. -- Get all outgoing links
  120. i = 1
  121. ia = 1
  122.  
  123. j = 1
  124. while herf[j] do
  125.     while string.find(htmlpage, herf[j]) do
  126.         links[ia] = string.match(htmlpage, herf[j])
  127.         htmlpage = string.gsub(htmlpage, herf[j], "", 1)
  128.         ia = ia + 1
  129.     end
  130. j = j + 1
  131. end
  132.  
  133.  
  134. -- Remove useless tags
  135. i = 1
  136. while remove[i] do
  137.     while string.find(htmlpage, remove[i]) do
  138.         htmlpage = string.gsub(htmlpage, remove[i], "")
  139.     end
  140. i = i + 1
  141. end
  142.  
  143. -- HTML Entities - Spaces
  144. while string.find(htmlpage, "&#160;") do
  145.     htmlpage = string.gsub(htmlpage, "&#160;", " ")
  146. end
  147. while string.find(htmlpage, "&nbsp;") do
  148.     htmlpage = string.gsub(htmlpage, "&nbsp;", " ")
  149. end
  150.  
  151. -- HTML Entities - inverted exclamation mark
  152. while string.find(htmlpage, "&#161;") do
  153.     htmlpage = string.gsub(htmlpage, "&#161;", "¡")
  154. end
  155. while string.find(htmlpage, "&iexcl;") do
  156.     htmlpage = string.gsub(htmlpage, "&iexcl;", " ")
  157. end
  158.  
  159. -- HTML Entities - right pointing guillemet
  160. while string.find(htmlpage, "&raquo;") do
  161.     htmlpage = string.gsub(htmlpage, "&raquo;", ">>")
  162. end
  163.  
  164. while string.find(htmlpage, "&#169;") do
  165.     htmlpage = string.gsub(htmlpage, "&#169;", "(c)")
  166. end
  167. while string.find(htmlpage, "&copy;") do
  168.     htmlpage = string.gsub(htmlpage, "&copy;", "(c)")
  169. end
  170.  
  171. -- Reduce Spaces
  172. while string.find(htmlpage, "  ") do
  173.     htmlpage = string.gsub(htmlpage, "  ", " ")
  174. end
  175. while string.find(htmlpage, "   ") do
  176.     htmlpage = string.gsub(htmlpage, "  ", " ")
  177. end
  178.  
  179.  
  180. -- Add Bullet points
  181. i = 1
  182. while Bullets[i] do
  183.     while string.find(htmlpage, Bullets[i]) do
  184.         htmlpage = string.gsub(htmlpage, Bullets[i], " * ")
  185.     end
  186. i = i + 1
  187. end
  188.  
  189. -- adding new line where needed
  190. i = 1
  191. while newline[i] do
  192.     htmlpage = string.gsub(htmlpage, newline[i], " ?!@ ")
  193.     i = i + 1
  194. end
  195.  
  196. viewpage = split(htmlpage, "?!@")
  197. i = 1
  198. while viewpage[i] do
  199.  
  200. -- Style Tags
  201.  
  202.  
  203. --center
  204. donotdeisablecenter = true
  205. if string.find(string.lower(viewpage[i]), "<center>") then
  206.   centerenabled = true
  207.   donotdeisablecenter = false
  208.   viewpage[i] = string.gsub(viewpage[i], "<center>", "")
  209.   viewpage[i] = string.gsub(viewpage[i], "<Center>", "")
  210.   viewpage[i] = string.gsub(viewpage[i], "<CENTER>", "")
  211. end
  212. if string.find(string.lower(viewpage[i]), "</center>") then
  213.   if donotdeisablecenter then centerenabled = false end
  214.   viewpage[i] = string.gsub(viewpage[i], "</center>", "")
  215.   viewpage[i] = string.gsub(viewpage[i], "</Center>", "")
  216.   viewpage[i] = string.gsub(viewpage[i], "</CENTER>", "")
  217. end
  218.  
  219. if centerenabled then
  220.   center[i] = 1
  221. else
  222.   center[i] = false
  223. end
  224.  
  225. --Horzonle Line
  226. if string.find(string.lower(viewpage[i]), "<hr>") then
  227.   viewpage[i] = string.gsub(viewpage[i], "<hr>", "")
  228.   viewpage[i] = string.gsub(viewpage[i], "<Hr>", "")
  229.   viewpage[i] = string.gsub(viewpage[i], "<HR>", "")
  230.   hr[i] = true
  231. else
  232.   hr[i] = false
  233. end
  234.  
  235.  
  236. i = i + 1
  237. end
  238.  
  239.  
  240. --[[
  241. Displaying Result
  242. ]]--
  243.  
  244. i = 0
  245. while true do
  246.  
  247. -- Displaying Page
  248.  
  249.  
  250. if veiw == "page" then
  251.     shell.run("clear")
  252.     y = 1
  253.     j = i + y
  254.     w,h = term.getSize()
  255.     while viewpage[j] and y <= h do
  256.      
  257.       --set x to center
  258.       if center[j] == 1 then
  259.                        x = math.ceil((w / 2) - (viewpage[j]:len() / 2))
  260.                        if x < 1 then
  261.                             x = 1
  262.                        end
  263.       else
  264.       -- not center
  265.             x = 1
  266.       end
  267.       term.setCursorPos(x,y)
  268.       term.clearLine()
  269.       -- Draw Horizontle line
  270.       if hr[j] then
  271.             hrX = 1
  272.             while hrX <= w do
  273.                 term.setCursorPos(hrX,y)
  274.                 write("-")
  275.                 hrX = hrX + 1
  276.             end
  277.             term.setCursorPos(x,y + 1)
  278.             term.clearLine()
  279.       end
  280.       -- Line wrap handaling
  281.       stringSize = string.len(viewpage[j])
  282.       while stringSize > w do
  283.         stringSize = stringSize - w
  284.         y = y + 1
  285.       end
  286.       --print text
  287.       write(viewpage[j])
  288.        y = y + 1
  289.        j = i + y
  290.       end
  291.     local evt, arg1, arg2 = os.pullEvent()
  292.     if arg1 == 17 or arg1 == 200 then
  293.       if viewpage[i + 21] then
  294.        i = i + 1
  295.       end
  296.     elseif arg1 == 31 or arg1 == 208 then
  297.       if viewpage[i - 1] then
  298.        i = i - 1
  299.       end
  300.     elseif arg1 == 14 or arg1 == 207 then
  301.         shell.run("clear")
  302.         break
  303.     elseif arg1 == 28 and links[1] ~= "None" then
  304.         veiw = "links"
  305.     end
  306.  
  307. -- Displaying Links
  308. elseif veiw == "links" then
  309.     shell.run("clear")
  310.     print("Use the arrows and enter to chose.")
  311.     j=0
  312.     max=-1
  313.     while true do
  314.         j=j+1
  315.         max=max+1
  316.         if links[j]==njl then break end
  317.     end
  318.     pointer = 1
  319.     term.setCursorPos(1,10)
  320.     term.clearLine(none)
  321.     print("->"..pointer..":"..links[pointer])
  322.     while true do
  323.  
  324.     list=1
  325.     while list < 10 do
  326.         term.setCursorPos(1,10-list)
  327.         term.clearLine(none)
  328.         if links[pointer - (list)] then
  329.         write(""..pointer - (list)..":"..links[pointer - (list)])
  330.     end
  331.     list=list+1
  332.     end
  333.     term.setCursorPos(1,10)
  334.     selectedItemString = "->"..pointer..":"..links[pointer] .. "<-"
  335.     if string.len(selectedItemString) > w then
  336.         diffrenct = string.len(selectedItemString) - w
  337.         selectedItemString = string.sub(selectedItemString, diffrenct)
  338.        
  339.     end
  340.     term.clearLine(none)
  341.     print(selectedItemString)
  342.     list=1
  343.     while list< 10 do
  344.         term.setCursorPos(1,10+list)
  345.         term.clearLine(none)
  346.         if links[pointer + (list)] then
  347.             write(""..pointer + (list)..":".. links[pointer + (list)])
  348.         end
  349.         list=list+1
  350.     end
  351.     local evt,arg1,arg2 = os.pullEvent()
  352.     if arg1 == 14 or arg1 == 207 then veiw = "page" break end
  353.     if arg1 == 200 and pointer > 1 then
  354.         pointer = pointer - 1
  355.     elseif arg1 == 208 and pointer < max then
  356.         pointer = pointer + 1
  357.     end
  358.     if arg1 == 28 then veiw = "none" break end
  359.  
  360.     end
  361.  
  362.  
  363. else
  364.     shell.run("clear")
  365.     shell.run("Firebox", links[pointer])
  366.     break
  367. end
  368.  
  369.  
  370. end
Advertisement
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!