import os
import sys

# The runtime environment assumes ISO8859/1 (European including Spain)
# CH, LL, and RR are not supported glyphs so we handle them here using 2 letters
# Note that when compiled as a stand-alone .exe under DOS/Windows,
# the default DOS character set is *not* ISO8859/1 and therefore the
# accented characters will not be displayed correctly.

# Customise here:
TITLE	= "Juez Lingüístico FISE"
Version	= "2.0"
ENTER_WORD	= "Teclee la palabra por la que desea consultar"
SUBMIT	= "Consultar"

GOOD_WORD	= "PALABRA ACEPTADA"
BAD_WORD	= "PALABRA NO ACEPTADA"

INVALID_ENTRY	= "Caracteres no válidos"
WRONG_SIZE	= "La palabra consultada debe tener entre dos y quince fichas"
NO_ARG_GIVEN	= "(vacio)"
WORD	= "palabra"
PROGNAME	= "juez2.cgi"

static showroot = (0==0)

#include "FISE/FISE-ROOT.h" # directory is protected.  File is not available for download

LATIN_CAPITAL_LETTER_N_WITH_TILDE	= 0xD1
LATIN_SMALL_LETTER_N_WITH_TILDE	= 0xF1
LATIN_SMALL_LETTER_A_WITH_ACUTE	= 0xE1
LATIN_SMALL_LETTER_E_WITH_ACUTE	= 0xE9
LATIN_SMALL_LETTER_I_WITH_ACUTE	= 0xED
LATIN_SMALL_LETTER_O_WITH_ACUTE	= 0xF3
LATIN_SMALL_LETTER_U_WITH_ACUTE	= 0xFA
LATIN_SMALL_LETTER_U_WITH_DIAERESIS	= 0xFC
LATIN_SMALL_LETTER_E_WITH_GRAVE	= 0xE8
LATIN_SMALL_LETTER_I_WITH_CIRCUMFLEX	= 0xEE
LATIN_CAPITAL_LETTER_A_WITH_ACUTE	= 0xC1
LATIN_CAPITAL_LETTER_E_WITH_ACUTE	= 0xC9
LATIN_CAPITAL_LETTER_I_WITH_ACUTE	= 0xCD
LATIN_CAPITAL_LETTER_O_WITH_ACUTE	= 0xD3
LATIN_CAPITAL_LETTER_U_WITH_ACUTE	= 0xDA
LATIN_CAPITAL_LETTER_U_WITH_DIAERESIS	= 0xDC
LATIN_CAPITAL_LETTER_E_WITH_GRAVE	= 0xC8
LATIN_CAPITAL_LETTER_I_WITH_CIRCUMFLEX	= 0xCE

#
# Web param stuff derived directly from code by Harry H. Cheng, 11/24/1995 *
# 

unescapechar(url) {
  from = url; to = from
  hex[3]; hexp=hex # contains xx of %xx 

  while True:
      c = *from += 1
      if c == '+': c = ' '
      if c == '%':
        *hexp++ = *from += 1; *hexp++ = *from += 1; *hexp = '\0'
        c = strtol(hex, NULL, 16); hexp = hex
      *to++ = c
      if c == '\0': return

def getpairs(name, value):
{                    # content length 
  num=1

  if (not strcmp("POST", os.getenv("REQUEST_METHOD"))) {
    # POST 
    if ((cl = int(os.getenv("CONTENT_LENGTH"))) == 0) return 0
    if (clientinput = () malloc(sizeof(char) * (cl+1))) == NULL: return -2
    fgets(clientinput, cl+1, stdin) # get the client input  
  else:
    # GET 
    if ((qs = os.getenv("QUERY_STRING")) == NULL) return 0; else:
      cl = len(qs) 
      if (clientinput = () malloc(sizeof(char) * (cl+1))) == NULL: return -2
      clientinput = qs

  token = strchr(clientinput, '&') # & is the name/value pair separator 
  while token != NULL:
    token = strchr(++token, '&')
    num += 1 # obtain the total number of pairs 

  nam = () malloc(sizeof()*num)
  val = () malloc(sizeof()*num)
  if nam == NULL or val == NULL: return -2

  for (i = 0, token = strtok(clientinput, "&"); token != NULL; i++) {
    l1 = len(token); l2 = strcspn(token,"=")
    nam[i] = () malloc(sizeof(char) * (l2+1))
    if nam[i] == NULL: return -2
    strncpy(nam[i], token, l2) 
    nam[i][l2] = '\0'
    unescapechar(nam[i])
    if l1 != l2+1: # name=value& 
      val[i] = ()malloc(sizeof(char) * (l1-l2))
      if val[i] == NULL: return -2
      strcpy(val[i], token+l2+1)
      unescapechar(val[i])
    else: val[i] = NULL # special case of name=& 
    token = strtok(NULL, "&")
  free(clientinput)
  *name = nam; *value = val
  return num # return the number of entries 

#
# Dictionary lookup code.  See also 'token.c'                       *
# 

typedef NODE

ROOT_NODE	= 1		# Root of dag 

V_END_OF_WORD	= 23	# Bit number of W 
M_END_OF_WORD	= (1L << V_END_OF_WORD)
V_END_OF_NODE	= 22	# Bit number of N 
M_END_OF_NODE	= (1L << V_END_OF_NODE)

V_LETTER	= 24
M_LETTER	= 0xFF

M_NODE_POINTER	= 0x1FFFFFL	# Bit mask for node pointer 

typedef INDEX
MAX_WORD_LEN	= 128

def dawg_pr(NODE *dawg, INDEX node, len):
  static word[MAX_WORD_LEN]
  NODE *edge

  for (edge = (NODE *)&dawg[node]; ; edge++) {
    c = *edge           # Don't rewrite this - its avoiding a MSC bug 
    c = c >> V_LETTER
    c = c & M_LETTER
    word[len] = (char)c
    if (*edge & M_END_OF_WORD) != 0:
      word[len+1] = '\0'
      s = word
      while True:
        if (s == NULL) or (*s == '\0'): break
        ws = s
        s = strchr(ws, ';')
        if s != NULL: *s++ = '\0'

        ws2 = ws
        s2 = strchr(ws2, ' ')
        if s2 != NULL: *s2 = '\0'

	fprintf(stdout, "<FONT color=blue><A HREF=\"http://buscon.rae.es/draeI/SrvltGUIBusUsual?LEMA=%s&origen=RAE&TIPO_BUS=3\">%s</A></FONT>&nbsp;\n" % (ws2, ws))
      fprintf(stdout, "<BR>")
    c = *edge & M_NODE_POINTER
    if (*edge & M_NODE_POINTER) != 0:
      dawg_pr (dawg, c, len + 1)
    if (*edge & M_END_OF_NODE) != 0: break # End of node 

def dawg_print(NODE *dawg, INDEX node):
  if not showroot: return
  fprintf(stdout, "La palabra proviene de:<BR>")
  dawg_pr(dawg, node, 0)
  fprintf(stdout, "<BR>")

def dawg_checkeq(NODE *dict, unsigned word):
  NODE *edge = dict+ROOT_NODE

  if edge == dict) return(0!=0:
  while True:
    ch = (int)(((*edge >> V_LETTER) & M_LETTER))
    ch = ch&255
    if ch == '=':
      if *word == '\0':
	dawg_print(dict, (*edge & M_NODE_POINTER))
        return(0==0)
    if ch == *word:
      word += 1
      if (edge = dict+(*edge & M_NODE_POINTER)) == dict: break # End of dict 
      continue # edge moved down a level 
    if ((*edge++) & M_END_OF_NODE) != 0: break # End of options 
  return(0!=0)

# Support for typographical conventions used (see document entitled 'COMMENT') 

static convert[256]	# canonicalize ISO Latin1

GOOD	= (0==0)
BAD	= (0!=0)

def reformat(tiles):
   static word[32]
   wordp = word

   while True:
      c = *tiles += 1

      if c == '4': c = LATIN_SMALL_LETTER_N_WITH_TILDE

      *wordp = c; if c == '\0': return word

      if c == '1':
        *wordp++ = 'c'; *wordp = 'h';}
      elif c == '2':
        *wordp++ = 'l'; *wordp = 'l';}
      elif c == '3':
        *wordp++ = 'r'; *wordp = 'r';}
      wordp += 1

def standardize(inword, outword):
   unsigned c

   while True:
      c = (unsigned int) (*inword++) & 255U
      if c == '\0': break

      c = convert[c]		# remove accents, convert to lower, leave
				# tilde on n.
      if c == '?') return (BAD)

      if (c == 'c') and (convert[(unsigned int) (*inword) & 255U] == 'h'):	# ch
	 c = '1'; inword += 1
      } elif (c == 'l') and (convert[(unsigned int) (*inword) & 255U] == 'l'):	# ll
	 c = '2'; inword += 1
      } elif (c == 'r') and (convert[(unsigned int) (*inword) & 255U] == 'r'):	# rr
	 c = '3'; inword += 1

      *outword++ = (char) c
   *outword = '\0'
   return (GOOD)


def main(argc, argv):
   word = ( '\0' )

   for i in range(0, 256): convert[i] = '?' # rejected unless explicitly specified
   for i in range('a', 'z'+1): convert[i] = i
   for i in range('A', 'Z'+1): convert[i] = tolower (i)

   # K and W not present in Spanish Scrabble
   convert['k'] = '?'; convert['w'] = '?'
   convert['K'] = '?'; convert['W'] = '?'

   # capital tilde N => lower tilde n, encoded here in 7-bit ascii as '4'
   # because the dawg code may not be 8-bit clean (signed char problem? - debug later...)

   convert[LATIN_CAPITAL_LETTER_N_WITH_TILDE] = '4'
   convert[LATIN_SMALL_LETTER_N_WITH_TILDE] = '4'

   # remove all accents and convert to lowercase

   convert[LATIN_SMALL_LETTER_A_WITH_ACUTE] = 'a'
   convert[LATIN_SMALL_LETTER_E_WITH_ACUTE] = 'e'
   convert[LATIN_SMALL_LETTER_I_WITH_ACUTE] = 'i'
   convert[LATIN_SMALL_LETTER_O_WITH_ACUTE] = 'o'
   convert[LATIN_SMALL_LETTER_U_WITH_ACUTE] = 'u'

   convert[LATIN_SMALL_LETTER_U_WITH_DIAERESIS] = 'u'
   convert[LATIN_SMALL_LETTER_E_WITH_GRAVE] = 'e'
   convert[LATIN_SMALL_LETTER_I_WITH_CIRCUMFLEX] = 'i'

   convert[LATIN_CAPITAL_LETTER_A_WITH_ACUTE] = 'a'
   convert[LATIN_CAPITAL_LETTER_E_WITH_ACUTE] = 'e'
   convert[LATIN_CAPITAL_LETTER_I_WITH_ACUTE] = 'i'
   convert[LATIN_CAPITAL_LETTER_O_WITH_ACUTE] = 'o'
   convert[LATIN_CAPITAL_LETTER_U_WITH_ACUTE] = 'u'

   convert[LATIN_CAPITAL_LETTER_U_WITH_DIAERESIS] = 'u'
   convert[LATIN_CAPITAL_LETTER_E_WITH_GRAVE] = 'e'
   convert[LATIN_CAPITAL_LETTER_I_WITH_CIRCUMFLEX] = 'i'

   { # WEB
      static s, line[32]
      c = 0

      if (getenv ("REQUEST_METHOD") != NULL) {	# Invoked from CGI 
	  pairs = 0

          # WEB 
	  fprintf(stdout, "Content-type: text/html\r\n")
	  fprintf(stdout, "\r\n")

	  fprintf(stdout, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\r\n")
	  fprintf(stdout, "    \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\r\n")
	  fprintf(stdout, "<html xmlns=\"http://www.w3.org/1999/xhtml\">\r\n")
	  fprintf(stdout, "<head>\r\n")
	  fprintf(stdout, "<meta http-equiv=\"Content-Type\" content=\"text/html;charset=iso-8859-1\" />\r\n")
	  fprintf(stdout, "<title>\r\n")
	  fprintf(stdout, "%s %s\r\n" % (TITLE, Version))
	  fprintf(stdout, "</title>\r\n")
	  fprintf(stdout, "</head>\r\n")
	  fprintf(stdout, "<body>\r\n")
	  fprintf(stdout, "<h1>%s %s</h1>\r\n" % (TITLE, Version))

          # display form for user to enter word
	  fprintf(stdout, "<form action=\"/cgi-bin/%s\" method=\"post\">\r\n" % (PROGNAME))
          fprintf(stdout, "%s: \r\n" % (ENTER_WORD))
	  fprintf(stdout, "<input type=\"text\" name=\"%s\" />\r\n" % (WORD))
	  fprintf(stdout, "<input type=\"submit\" value=\"%s\" />\r\n" % (SUBMIT))
	  fprintf(stdout, "<BR>\r\n")
	  fprintf(stdout, "<BR>&iquest;Mostrar la palabra origen? ")
	  if (strcmp(os.getenv("REQUEST_METHOD"), "POST") == 0) pairs = getpairs(&name, &value)
	  if pairs > 0:
	    for i in range(0, pairs): if (strcmp(name[i], "showroot") == 0) {showroot = (*value[i] == 'Y'); break;}
	  fprintf(stdout, "<input type=\"radio\" name=\"showroot\" value=\"Y\"%s> SI" % (showroot ? " checked" : "")
	  fprintf(stdout, "<input type=\"radio\" name=\"showroot\" value=\"N\"%s> NO<BR>\r\n" % (showroot ? "" : " checked")
	  fprintf(stdout, "</form>\r\n")

	  if (strcmp(os.getenv("REQUEST_METHOD"), "POST") == 0) {
              GIVEN_ARG = NULL

	      # word supplied, check it

	      if pairs > 0:

# TO DO: conversions into and out of quoted web format; hex entity refs for accented chars

		for i in range(0, pairs): if strcmp(name[i], WORD) == 0:GIVEN_ARG = value[i]; break;}
		for i in range(0, pairs): if (strcmp(name[i], "showroot") == 0) {showroot = *value[i] == 'Y'; break;}

                if (GIVEN_ARG != NULL) and (not standardize (GIVEN_ARG, word)):
		    fprintf(stdout, "<font color=\"red\">%s: %s</font>\n\n" % (INVALID_ENTRY, GIVEN_ARG))
	        } elif (GIVEN_ARG == NULL) or (strlen (GIVEN_ARG) > 30) or (strlen (word) < 2) or (strlen (word) > 15):
	            fprintf(stdout, "<font color=\"red\">%s: %s</font>\n" % (WRONG_SIZE,
		            (GIVEN_ARG == NULL ? NO_ARG_GIVEN : GIVEN_ARG)))
	        else:
	           if dawg_checkeq(_token, word):
		       fprintf(stdout, "<font color=\"green\">%s: %s</font>\n" % (GOOD_WORD, reformat(word)))
	           else:
		       fprintf(stdout, "<font color=\"red\">%s: %s</font>\n" % (BAD_WORD, reformat(word)))
		for i in range(pairs):
		    if name[i] != NULL) free(name[i]:
		    if value[i] != NULL) free(value[i]:
		if name != NULL) free(name:
		if value != NULL) free(value:
	      else:
	          fprintf(stdout, "Form does not seem to include the required field '%s'.\r\n" % (WORD))

	  fprintf(stdout, "</body>\r\n")
	  fprintf(stdout, "</html>\r\n")

   exit (0)