math/lexer.mll

   1 {
   2     open Parser
   3     open Render_info
   4     open Tex
   5 }
   6 let space = [' ' '\t' '\n' '\r']
   7 let alpha = ['a'-'z' 'A'-'Z']
   8 let literal_id = ['a'-'z' 'A'-'Z']
   9 let literal_mn = ['0'-'9']
  10 let literal_uf_lt = [',' ':' ';' '?' '!' '\'']
  11 let delimiter_uf_lt = ['(' ')' '.']
  12 let literal_uf_op = ['+' '-' '*' '=']
  13 let delimiter_uf_op = ['/' '|']
  14 let boxchars  = ['0'-'9' 'a'-'z' 'A'-'Z' '+' '-' '*' ',' '=' '(' ')' ':' '/' ';' '?' '.' '!' ' ' '\128'-'\255']
  15 let aboxchars = ['0'-'9' 'a'-'z' 'A'-'Z' '+' '-' '*' ',' '=' '(' ')' ':' '/' ';' '?' '.' '!' ' ']
  16
  17 rule token = parse
  18     space +                     { token lexbuf }
  19   | "\\text" space * '{' boxchars + '}'
  20                                 { Texutil.tex_use_ams (); let str = Lexing.lexeme lexbuf in
  21                                   let n = String.index str '{' + 1 in
  22                                   BOX ("\\text", String.sub str n (String.length str - n - 1)) }
  23   | "\\mbox" space * '{' aboxchars + '}'
  24                                 { let str = Lexing.lexeme lexbuf in
  25                                   let n = String.index str '{' + 1 in
  26                                   BOX ("\\mbox", String.sub str n (String.length str - n - 1)) }
  27   | "\\hbox" space * '{' aboxchars + '}'
  28                                 { let str = Lexing.lexeme lexbuf in
  29                                   let n = String.index str '{' + 1 in
  30                                   BOX ("\\hbox", String.sub str n (String.length str - n - 1)) }
  31   | "\\vbox" space * '{' aboxchars + '}'
  32                                 { let str = Lexing.lexeme lexbuf in
  33                                   let n = String.index str '{' + 1 in
  34                                   BOX ("\\vbox", String.sub str n (String.length str - n - 1)) }
  35   | "\\mbox" space * '{' boxchars + '}'
  36                                 { let str = Lexing.lexeme lexbuf in
  37                                   let n = String.index str '{' + 1 in
  38                                   Texutil.tex_use_nonascii();
  39                                   BOX ("\\mbox", String.sub str n (String.length str - n - 1)) }
  40   | "\\hbox" space * '{' boxchars + '}'
  41                                 { let str = Lexing.lexeme lexbuf in
  42                                   let n = String.index str '{' + 1 in
  43                                   Texutil.tex_use_nonascii();
  44                                   BOX ("\\hbox", String.sub str n (String.length str - n - 1)) }
  45   | "\\vbox" space * '{' boxchars + '}'
  46                                 { let str = Lexing.lexeme lexbuf in
  47                                   let n = String.index str '{' + 1 in
  48                                   Texutil.tex_use_nonascii();
  49                                   BOX ("\\vbox", String.sub str n (String.length str - n - 1)) }
  50   | literal_id                  { let str = Lexing.lexeme lexbuf in LITERAL (MHTMLABLEC (FONT_IT, str,str,MI,str)) }
  51   | literal_mn                  { let str = Lexing.lexeme lexbuf in LITERAL (MHTMLABLEC (FONT_RM, str,str,MN,str)) }
  52   | literal_uf_lt               { let str = Lexing.lexeme lexbuf in LITERAL (HTMLABLEC (FONT_UFH, str,str)) }
  53   | delimiter_uf_lt             { let str = Lexing.lexeme lexbuf in DELIMITER (HTMLABLEC (FONT_UFH, str,str)) }
  54   | "-"                         { let str = Lexing.lexeme lexbuf in LITERAL (MHTMLABLEC (FONT_UFH,"-"," &minus; ",MO,str))}
  55   | literal_uf_op               { let str = Lexing.lexeme lexbuf in LITERAL (MHTMLABLEC (FONT_UFH, str," "^str^" ",MO,str)) }
  56   | delimiter_uf_op             { let str = Lexing.lexeme lexbuf in DELIMITER (MHTMLABLEC (FONT_UFH, str," "^str^" ",MO,str)) }
  57   | "\\" alpha +                { Texutil.find (Lexing.lexeme lexbuf) }
  58   | "\\sqrt" space * "["        { FUN_AR1opt "\\sqrt" }
  59   | "\\xleftarrow" space * "["  { Texutil.tex_use_ams(); FUN_AR1opt "\\xleftarrow" }
  60   | "\\xrightarrow" space * "[" { Texutil.tex_use_ams(); FUN_AR1opt "\\xrightarrow" }
  61   | "\\,"                       { LITERAL (HTMLABLE (FONT_UF, "\\,","&nbsp;")) }
  62   | "\\ "                       { LITERAL (HTMLABLE (FONT_UF, "\\ ","&nbsp;")) }
  63   | "\\;"                       { LITERAL (HTMLABLE (FONT_UF, "\\;","&nbsp;")) }
  64   | "\\!"                       { LITERAL (TEX_ONLY "\\!") }
  65   | "\\{"                       { DELIMITER (HTMLABLEC(FONT_UFH,"\\{","{")) }
  66   | "\\}"                       { DELIMITER (HTMLABLEC(FONT_UFH,"\\}","}")) }
  67   | "\\|"                       { DELIMITER (HTMLABLE (FONT_UFH,"\\|","||")) }
  68   | "\\_"                       { LITERAL (HTMLABLEC(FONT_UFH,"\\_","_")) }
  69   | "\\#"                       { LITERAL (HTMLABLE (FONT_UFH,"\\#","#")) }
  70   | "\\%"                       { LITERAL (HTMLABLE (FONT_UFH,"\\%","%")) }
  71   | "\\$"                       { LITERAL (HTMLABLE (FONT_UFH,"\\$","$")) }
  72   | "\\&"                       { LITERAL (HTMLABLEC (FONT_RM,"\\&","&amp;")) }
  73   | "&"                         { NEXT_CELL }
  74   | "\\\\"                      { NEXT_ROW }
  75   | "\\begin{matrix}"           { Texutil.tex_use_ams(); BEGIN__MATRIX }
  76   | "\\end{matrix}"             { END__MATRIX }
  77   | "\\begin{pmatrix}"          { Texutil.tex_use_ams(); BEGIN_PMATRIX }
  78   | "\\end{pmatrix}"            { END_PMATRIX }
  79   | "\\begin{bmatrix}"          { Texutil.tex_use_ams(); BEGIN_BMATRIX }
  80   | "\\end{bmatrix}"            { END_BMATRIX }
  81   | "\\begin{Bmatrix}"          { Texutil.tex_use_ams(); BEGIN_BBMATRIX }
  82   | "\\end{Bmatrix}"            { END_BBMATRIX }
  83   | "\\begin{vmatrix}"          { Texutil.tex_use_ams(); BEGIN_VMATRIX }
  84   | "\\end{vmatrix}"            { END_VMATRIX }
  85   | "\\begin{Vmatrix}"          { Texutil.tex_use_ams(); BEGIN_VVMATRIX }
  86   | "\\end{Vmatrix}"            { END_VVMATRIX }
  87   | "\\begin{array}"            { Texutil.tex_use_ams(); BEGIN_ARRAY }
  88   | "\\end{array}"              { END_ARRAY }
  89   | "\\begin{align}"            { Texutil.tex_use_ams(); BEGIN_ALIGN }
  90   | "\\end{align}"              { END_ALIGN }
  91   | "\\begin{alignat}"          { Texutil.tex_use_ams(); BEGIN_ALIGNAT }
  92   | "\\end{alignat}"            { END_ALIGNAT }
  93   | "\\begin{smallmatrix}"      { Texutil.tex_use_ams(); BEGIN_SMALLMATRIX }
  94   | "\\end{smallmatrix}"        { END_SMALLMATRIX }
  95   | "\\begin{cases}"            { Texutil.tex_use_ams(); BEGIN_CASES }
  96   | "\\end{cases}"              { END_CASES }
  97   | '>'                         { LITERAL (HTMLABLEC(FONT_UFH,">"," &gt; ")) }
  98   | '<'                         { LITERAL (HTMLABLEC(FONT_UFH,"<"," &lt; ")) }
  99   | '%'                         { LITERAL (HTMLABLEC(FONT_UFH,"\\%","%")) }
 100   | '$'                         { LITERAL (HTMLABLEC(FONT_UFH,"\\$","$")) }
 101   | '~'                         { LITERAL (HTMLABLE (FONT_UF, "~","&nbsp;")) }
 102   | '['                         { DELIMITER (HTMLABLEC(FONT_UFH,"[","[")) }
 103   | ']'                         { SQ_CLOSE }
 104   | '{'                         { CURLY_OPEN }
 105   | '}'                         { CURLY_CLOSE }
 106   | '^'                         { SUP }
 107   | '_'                         { SUB }
 108   | eof                         { EOF }