| File: | modules/Text/xxml.cc |
| Location: | line 125, column 5 |
| Description: | Value stored to 'utt' is never read |
| 1 | /*************************************************************************/ |
| 2 | /* */ |
| 3 | /* Centre for Speech Technology Research */ |
| 4 | /* University of Edinburgh, UK */ |
| 5 | /* Copyright (c) 1996,1997 */ |
| 6 | /* All Rights Reserved. */ |
| 7 | /* */ |
| 8 | /* Permission is hereby granted, free of charge, to use and distribute */ |
| 9 | /* this software and its documentation without restriction, including */ |
| 10 | /* without limitation the rights to use, copy, modify, merge, publish, */ |
| 11 | /* distribute, sublicense, and/or sell copies of this work, and to */ |
| 12 | /* permit persons to whom this work is furnished to do so, subject to */ |
| 13 | /* the following conditions: */ |
| 14 | /* 1. The code must retain the above copyright notice, this list of */ |
| 15 | /* conditions and the following disclaimer. */ |
| 16 | /* 2. Any modifications must be clearly marked as such. */ |
| 17 | /* 3. Original authors' names are not deleted. */ |
| 18 | /* 4. The authors' names are not used to endorse or promote products */ |
| 19 | /* derived from this software without specific prior written */ |
| 20 | /* permission. */ |
| 21 | /* */ |
| 22 | /* THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK */ |
| 23 | /* DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING */ |
| 24 | /* ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT */ |
| 25 | /* SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE */ |
| 26 | /* FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES */ |
| 27 | /* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN */ |
| 28 | /* AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, */ |
| 29 | /* ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF */ |
| 30 | /* THIS SOFTWARE. */ |
| 31 | /* */ |
| 32 | /*************************************************************************/ |
| 33 | /* Author : Alan W Black */ |
| 34 | /* Date : August 1997 */ |
| 35 | /*-----------------------------------------------------------------------*/ |
| 36 | /* */ |
| 37 | /* There are just too many different versions of sgml-based mark up */ |
| 38 | /* and none of them are stable so this is allows arbitrary */ |
| 39 | /* of tags to Lisp functions so any of them can be implemented in Lisp */ |
| 40 | /* That is people can worry about the actual content later and do not */ |
| 41 | /* need to change the C++ */ |
| 42 | /* */ |
| 43 | /* Of course once I give you this functionality you'll just want more */ |
| 44 | /* */ |
| 45 | /*=======================================================================*/ |
| 46 | #include "EST_unix.h" |
| 47 | #include "festival.h" |
| 48 | #include "text.h" |
| 49 | #include "lexicon.h" |
| 50 | |
| 51 | using namespace std; |
| 52 | |
| 53 | static LISP xxml_get_attribute(const EST_String &remainder); |
| 54 | static char *xxml_process_line(const char *line); |
| 55 | static void tts_xxml_token(EST_Item *t); |
| 56 | static void tts_xxml_utt(LISP lutt); |
| 57 | |
| 58 | static LISP xxml_word_features = NIL((struct obj *) 0); |
| 59 | static LISP xxml_token_hooks = NIL((struct obj *) 0); |
| 60 | |
| 61 | void tts_file_xxml(LISP filename) |
| 62 | { |
| 63 | // For stml ssml rsml and jsml etc |
| 64 | // filename contains *output* from something like nsgml |
| 65 | EST_String inname = get_c_string(filename); |
| 66 | EST_String line, type, remainder; |
| 67 | EST_TokenStream ts; |
| 68 | LISP atts, element_defs; |
| 69 | LISP utt = NIL((struct obj *) 0); // for cummulation of tokens |
| 70 | |
| 71 | if (ts.open(inname) == -1) |
| 72 | { |
| 73 | cerr << "xxml: unable to open output from SGML parser" << endl; |
| 74 | festival_error()(errjmp_ok ? longjmp(*est_errjmp,1) : festival_tidy_up(),exit (-1)); |
| 75 | } |
| 76 | ts.set_WhiteSpaceChars(" \t\r\n"); |
| 77 | ts.set_SingleCharSymbols(""); |
| 78 | ts.set_PunctuationSymbols(""); |
| 79 | ts.set_PrePunctuationSymbols(""); |
| 80 | |
| 81 | element_defs = siod_get_lval("xxml_elements",NULL__null); |
| 82 | atts = NIL((struct obj *) 0); |
| 83 | |
| 84 | if (ts.peek() != get_c_string(car(car(element_defs)))) |
| 85 | { |
| 86 | cerr << "xxml parse error: " << get_c_string(filename) << |
| 87 | " Expected " << get_c_string(car(car(element_defs))) |
| 88 | << " but found " << ts.peek() << endl; |
| 89 | festival_error()(errjmp_ok ? longjmp(*est_errjmp,1) : festival_tidy_up(),exit (-1)); |
| 90 | } |
| 91 | while (ts.peek() != get_c_string(car(car(cdr(element_defs))))) |
| 92 | { |
| 93 | if (ts.eof()) |
| 94 | { |
| 95 | cerr << "xxml parse error: unexpected end of file \n"; |
| 96 | festival_error()(errjmp_ok ? longjmp(*est_errjmp,1) : festival_tidy_up(),exit (-1)); |
| 97 | } |
| 98 | line = (EST_String)ts.get_upto_eoln(); |
| 99 | type = line.at(0,1); |
| 100 | remainder = line.after(0); |
| 101 | if (type == "-") |
| 102 | { // Segments into utterances as it goes along |
| 103 | utt = xxml_get_tokens(remainder, |
| 104 | siod_get_lval("xxml_word_features",NULL__null), |
| 105 | utt); |
| 106 | } |
| 107 | else if (type == "A") // general attribute |
| 108 | { |
| 109 | atts = cons(xxml_get_attribute(remainder),atts); |
| 110 | } |
| 111 | else if ((type == "(") || (type == ")")) |
| 112 | { |
| 113 | utt = xxml_call_element_function(type+remainder,atts, |
| 114 | element_defs,utt); |
| 115 | atts = NIL((struct obj *) 0); |
| 116 | } |
| 117 | else |
| 118 | { |
| 119 | cerr << "xxml parse error: unexpected token found " |
| 120 | << line << endl; |
| 121 | festival_error()(errjmp_ok ? longjmp(*est_errjmp,1) : festival_tidy_up(),exit (-1)); |
| 122 | } |
| 123 | } |
| 124 | // Last call (should synthesize trailing tokens |
| 125 | utt = xxml_call_element_function(ts.get().string(),atts,element_defs,utt); |
Value stored to 'utt' is never read | |
| 126 | |
| 127 | ts.close(); |
| 128 | } |
| 129 | |
| 130 | LISP xxml_call_element_function(const EST_String &element, |
| 131 | LISP atts, LISP elements, LISP utt) |
| 132 | { |
| 133 | // Form the call to the defined element function, with the attributes |
| 134 | // and the utterance, returns the utterance |
| 135 | LISP def,l; |
| 136 | |
| 137 | def = siod_assoc_str(element,elements); |
| 138 | |
| 139 | if (def != NIL((struct obj *) 0)) |
| 140 | { |
| 141 | // You get two arguments, ATTLIST and UTTERANCE |
| 142 | l = cons( |
| 143 | make_param_lisp("ATTLIST", |
| 144 | cons(rintern("quote"),cons(atts,NIL((struct obj *) 0)))), |
| 145 | cons( |
| 146 | make_param_lisp("UTT", |
| 147 | cons(rintern("quote"),cons(utt,NIL((struct obj *) 0)))), |
| 148 | NIL((struct obj *) 0))); |
| 149 | return leval(cons(rintern("let"), |
| 150 | cons(l,cdr(cdr(def)))),NIL((struct obj *) 0)); |
| 151 | } |
| 152 | else // no definition to do nothing |
| 153 | return utt; |
| 154 | } |
| 155 | |
| 156 | static LISP xxml_get_attribute(const EST_String &remainder) |
| 157 | { |
| 158 | EST_TokenStream ts; |
| 159 | LISP tokens=NIL((struct obj *) 0),att=NIL((struct obj *) 0); |
| 160 | EST_String name; |
| 161 | EST_Token t; |
| 162 | |
| 163 | ts.open_string(remainder); |
| 164 | name = (EST_String)ts.get(); |
| 165 | if ((t=ts.get()) == "IMPLIED") |
| 166 | att = cons(rintern(name),cons(NIL((struct obj *) 0),NIL((struct obj *) 0))); |
| 167 | else if (t == "TOKEN") |
| 168 | { |
| 169 | EST_Token v = ts.get(); |
| 170 | att = cons(rintern(name),cons(cons(rintern(v.string()),NIL((struct obj *) 0)),NIL((struct obj *) 0))); |
| 171 | } |
| 172 | else if (t == "CDATA") |
| 173 | { |
| 174 | while (!ts.eof()) |
| 175 | tokens = cons(rintern(ts.get().string()),tokens); |
| 176 | att = cons(rintern(name),cons(reverse(tokens),NIL((struct obj *) 0))); |
| 177 | } |
| 178 | else |
| 179 | { |
| 180 | cerr << "XXML: unknow attribute type " << remainder << endl; |
| 181 | festival_error()(errjmp_ok ? longjmp(*est_errjmp,1) : festival_tidy_up(),exit (-1)); |
| 182 | } |
| 183 | |
| 184 | ts.close(); |
| 185 | return att; |
| 186 | } |
| 187 | |
| 188 | static char *xxml_process_line(const char *line) |
| 189 | { |
| 190 | // STML (sgml) data line have a number of special escape characters |
| 191 | // this undoes them, namely "\\n" to "\n" |
| 192 | char *procline = walloc(char,strlen(line)+1)((char *)safe_walloc(sizeof(char)*(strlen(line)+1))); |
| 193 | int i,j; |
| 194 | |
| 195 | for (i=j=0; line[i] != '\0'; j++,i++) |
| 196 | { |
| 197 | if (line[i] == '\\') |
| 198 | { |
| 199 | i++; |
| 200 | if (line[i] == 'n') |
| 201 | procline[j] = '\n'; |
| 202 | else if (line[i] == '\\') |
| 203 | procline[j] = '\\'; |
| 204 | else if ((line[i] == '0') || // its an octal number |
| 205 | (line[i] == '1')) |
| 206 | { |
| 207 | int k,oct = 0; |
| 208 | for (k=0; k < 3; k++,i++) |
| 209 | oct = (oct*8)+(line[i]-'0'); |
| 210 | procline[j] = oct; |
| 211 | i--; |
| 212 | } |
| 213 | else |
| 214 | { |
| 215 | procline[j] = line[i]; // no change |
| 216 | i--; |
| 217 | } |
| 218 | } |
| 219 | else |
| 220 | procline[j] = line[i]; // no change |
| 221 | } |
| 222 | procline[j] = '\0'; |
| 223 | return procline; |
| 224 | } |
| 225 | |
| 226 | static void tts_xxml_token(EST_Item *t) |
| 227 | { |
| 228 | // Add xxml_word features to t |
| 229 | LISP a; |
| 230 | |
| 231 | for (a=xxml_word_features; a != NIL((struct obj *) 0); a=cdr(a)) |
| 232 | if ((car(cdr(car(a))) != NIL((struct obj *) 0)) && |
| 233 | (!streq("NAME",get_c_string(car(car(a))))(strcmp("NAME",get_c_string(car(car(a))))==0))) |
| 234 | { |
| 235 | if (cdr(cdr(car(a))) == NIL((struct obj *) 0)) |
| 236 | t->set(get_c_string(car(car(a))), |
| 237 | get_c_string(car(cdr(car(a))))); |
| 238 | else |
| 239 | { |
| 240 | // Its more complex than a single atom so save the list |
| 241 | t->set(get_c_string(car(car(a))), |
| 242 | siod_sprint(car(cdr(car(a))))); |
| 243 | } |
| 244 | } |
| 245 | |
| 246 | apply_hooks(xxml_token_hooks,siod(t)); |
| 247 | } |
| 248 | |
| 249 | LISP xxml_get_tokens(const EST_String &line,LISP feats,LISP utt) |
| 250 | { |
| 251 | // Read from here until end of line collects all the tokens |
| 252 | // Note tokens are in reverse order until they are made into an |
| 253 | // utterance |
| 254 | EST_TokenStream ls; |
| 255 | EST_Token t; |
| 256 | LISP eou_tree; |
| 257 | char *processed_line; |
| 258 | processed_line = xxml_process_line(line); |
| 259 | ls.open_string(processed_line); |
| 260 | ls.set_SingleCharSymbols( |
| 261 | get_c_string(siod_get_lval("token.singlecharsymbols", |
| 262 | "token.singlecharsymbols unset"))); |
| 263 | ls.set_PunctuationSymbols( |
| 264 | get_c_string(siod_get_lval("token.punctuation", |
| 265 | "token.punctuation unset"))); |
| 266 | ls.set_PrePunctuationSymbols( |
| 267 | get_c_string(siod_get_lval("token.prepunctuation", |
| 268 | "token.prepunctuation unset"))); |
| 269 | ls.set_WhiteSpaceChars( |
| 270 | get_c_string(siod_get_lval("token.whitespace", |
| 271 | "token.whitespace unset"))); |
| 272 | |
| 273 | eou_tree = siod_get_lval("eou_tree","No end of utterance tree set"); |
| 274 | |
| 275 | xxml_word_features = feats; |
| 276 | xxml_token_hooks = siod_get_lval("xxml_token_hooks",NULL__null); |
| 277 | |
| 278 | // Segment and synth as much as appropriate |
| 279 | utt = tts_chunk_stream(ls,tts_xxml_token,tts_xxml_utt,eou_tree,utt); |
| 280 | |
| 281 | return utt; |
| 282 | } |
| 283 | |
| 284 | static void tts_xxml_utt(LISP lutt) |
| 285 | { |
| 286 | // Build and utterance with these tokens and apply xxml synth function |
| 287 | |
| 288 | if ((lutt == NIL((struct obj *) 0)) || |
| 289 | (get_c_utt(lutt)(utterance(lutt))->relation("Token")->length() == 0)) |
| 290 | return; // in this case do nothing. |
| 291 | |
| 292 | leval(cons(rintern("xxml_synth"), |
| 293 | cons(quote(lutt),NIL((struct obj *) 0))),NIL((struct obj *) 0)); |
| 294 | } |
| 295 |