| 1 |
/* ============================================================================ |
| 2 |
* Douglas Thrift's Search Engine License |
| 3 |
* |
| 4 |
* Copyright (C) 2002-2003, Douglas Thrift. All Rights Reserved. |
| 5 |
* Redistribution and use in source and binary forms, with or without |
| 6 |
* modification, are permitted provided that the following conditions are met: |
| 7 |
* |
| 8 |
* 1. Redistributions of source code must retain the above copyright notice, |
| 9 |
* this list of conditions and the following disclaimer. |
| 10 |
* |
| 11 |
* 2. Redistributions in binary form must reproduce the above copyright notice, |
| 12 |
* this list of conditions and the following disclaimer in the documentation |
| 13 |
* and/or other materials provided with the distribution. |
| 14 |
* |
| 15 |
* 3. The end-user documentation included with the redistribution, if any, must |
| 16 |
* include the following acknowledgment: |
| 17 |
* |
| 18 |
* "This product includes software developed by Douglas Thrift |
| 19 |
* (http://computers.douglasthrift.net/searchengine/)." |
| 20 |
* |
| 21 |
* Alternately, this acknowledgment may appear in the software itself, if |
| 22 |
* and wherever such third-party acknowledgments normally appear. |
| 23 |
* |
| 24 |
* 4. The names "Douglas Thrift" and "Douglas Thrift's Search Engine" must not |
| 25 |
* be used to endorse or promote products derived from this software without |
| 26 |
* specific prior written permission. For written permission, please visit |
| 27 |
* http://www.douglasthrift.net/contact.cgi for contact information. |
| 28 |
* |
| 29 |
* 5. Products derived from this software may not be called "Douglas Thrift's |
| 30 |
* Search Engine", nor may "Douglas Thrift's Search Engine" appear in their |
| 31 |
* name, without prior written permission. |
| 32 |
* |
| 33 |
* THIS SOFTWARE IS PROVIDED "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, |
| 34 |
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND |
| 35 |
* FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE |
| 36 |
* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
| 37 |
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| 38 |
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, |
| 39 |
* OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF |
| 40 |
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING |
| 41 |
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, |
| 42 |
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 43 |
* ============================================================================ |
| 44 |
*/ |
| 45 |
// Douglas Thrift's Search Engine Main |
| 46 |
// |
| 47 |
// Douglas Thrift |
| 48 |
// |
| 49 |
// Search.cpp |
| 50 |
|
| 51 |
#include "Search.h" |
| 52 |
#include "Indexer.h" |
| 53 |
#include "Searcher.h" |
| 54 |
#include "Outputer.h" |
| 55 |
|
| 56 |
#ifndef _WIN32 |
| 57 |
#include <sys/utsname.h> |
| 58 |
#else |
| 59 |
#include <windows.h> |
| 60 |
#endif // _WIN32 |
| 61 |
|
| 62 |
string program; |
| 63 |
string programName = "Douglas Thrift's Search Engine"; |
| 64 |
string programVersion = "1.1alpha2"; |
| 65 |
bool debug = false; |
| 66 |
|
| 67 |
int main(int argc, char* argv[]) |
| 68 |
{ |
| 69 |
program = argv[0]; |
| 70 |
|
| 71 |
bool indexMode = false; |
| 72 |
string indexURL; |
| 73 |
set<string> indexDomains; |
| 74 |
set<string> indexRestrictions; |
| 75 |
|
| 76 |
unsigned page = 1; |
| 77 |
string query; |
| 78 |
|
| 79 |
vector<string> indices; |
| 80 |
|
| 81 |
string header = "header.html"; |
| 82 |
string body = "body.html"; |
| 83 |
string footer = "footer.html"; |
| 84 |
string notfound = "notfound.html"; |
| 85 |
string pages = "pages.html"; |
| 86 |
|
| 87 |
for (int index = 1; index < argc; index++) |
| 88 |
{ |
| 89 |
string arg(argv[index]); |
| 90 |
|
| 91 |
if (arg == "-help") |
| 92 |
{ |
| 93 |
usage(); |
| 94 |
return 0; |
| 95 |
} |
| 96 |
else if (arg == "-version") |
| 97 |
{ |
| 98 |
version(); |
| 99 |
return 0; |
| 100 |
} |
| 101 |
else if (arg == "-license") |
| 102 |
{ |
| 103 |
license(); |
| 104 |
return 0; |
| 105 |
} |
| 106 |
else if (arg == "-P") |
| 107 |
{ |
| 108 |
if (++index < argc) |
| 109 |
{ |
| 110 |
page = strtoul(argv[index],0,0); |
| 111 |
} |
| 112 |
else |
| 113 |
{ |
| 114 |
cerr << program << ": Bad arguments\n"; |
| 115 |
usage(); |
| 116 |
return 1; |
| 117 |
} |
| 118 |
} |
| 119 |
else if (arg == "-i") |
| 120 |
{ |
| 121 |
indexMode = true; |
| 122 |
|
| 123 |
if (++index < argc) |
| 124 |
{ |
| 125 |
indexURL = argv[index]; |
| 126 |
} |
| 127 |
else |
| 128 |
{ |
| 129 |
cerr << program << ": Bad arguments\n"; |
| 130 |
usage(); |
| 131 |
return 1; |
| 132 |
} |
| 133 |
} |
| 134 |
else if (arg == "-d") |
| 135 |
{ |
| 136 |
if (++index < argc) |
| 137 |
{ |
| 138 |
indexDomains.insert(argv[index]); |
| 139 |
} |
| 140 |
else |
| 141 |
{ |
| 142 |
cerr << program << ": Bad arguments\n"; |
| 143 |
usage(); |
| 144 |
return 1; |
| 145 |
} |
| 146 |
} |
| 147 |
else if (arg == "-r") |
| 148 |
{ |
| 149 |
if (++index < argc) |
| 150 |
{ |
| 151 |
indexRestrictions.insert(argv[index]); |
| 152 |
} |
| 153 |
else |
| 154 |
{ |
| 155 |
cerr << program << ": Bad arguments\n"; |
| 156 |
usage(); |
| 157 |
return 1; |
| 158 |
} |
| 159 |
} |
| 160 |
else if (arg == "-h") |
| 161 |
{ |
| 162 |
if (++index < argc) |
| 163 |
{ |
| 164 |
header = argv[index]; |
| 165 |
} |
| 166 |
else |
| 167 |
{ |
| 168 |
cerr << program << ": Bad arguments\n"; |
| 169 |
usage(); |
| 170 |
return 1; |
| 171 |
} |
| 172 |
} |
| 173 |
else if (arg == "-b") |
| 174 |
{ |
| 175 |
if (++index < argc) |
| 176 |
{ |
| 177 |
body = argv[index]; |
| 178 |
} |
| 179 |
else |
| 180 |
{ |
| 181 |
cerr << program << ": Bad arguments\n"; |
| 182 |
usage(); |
| 183 |
return 1; |
| 184 |
} |
| 185 |
} |
| 186 |
else if (arg == "-f") |
| 187 |
{ |
| 188 |
if (++index < argc) |
| 189 |
{ |
| 190 |
footer = argv[index]; |
| 191 |
} |
| 192 |
else |
| 193 |
{ |
| 194 |
cerr << program << ": Bad arguments\n"; |
| 195 |
usage(); |
| 196 |
return 1; |
| 197 |
} |
| 198 |
} |
| 199 |
else if (arg == "-n") |
| 200 |
{ |
| 201 |
if (++index < argc) |
| 202 |
{ |
| 203 |
notfound = argv[index]; |
| 204 |
} |
| 205 |
else |
| 206 |
{ |
| 207 |
cerr << program << ": Bad arguments\n"; |
| 208 |
usage(); |
| 209 |
return 1; |
| 210 |
} |
| 211 |
} |
| 212 |
else if (arg == "-p") |
| 213 |
{ |
| 214 |
if (++index < argc) |
| 215 |
{ |
| 216 |
pages = argv[index]; |
| 217 |
} |
| 218 |
else |
| 219 |
{ |
| 220 |
cerr << program << ": Bad arguments\n"; |
| 221 |
usage(); |
| 222 |
return 1; |
| 223 |
} |
| 224 |
} |
| 225 |
else if (arg == "-e") |
| 226 |
{ |
| 227 |
if (++index < argc) |
| 228 |
{ |
| 229 |
// set from email |
| 230 |
} |
| 231 |
else |
| 232 |
{ |
| 233 |
cerr << program << ": Bad arguments\n"; |
| 234 |
usage(); |
| 235 |
return 1; |
| 236 |
} |
| 237 |
} |
| 238 |
else if (arg == "-D") |
| 239 |
{ |
| 240 |
debug = true; |
| 241 |
} |
| 242 |
else |
| 243 |
{ |
| 244 |
indices.push_back(arg); |
| 245 |
} |
| 246 |
} |
| 247 |
|
| 248 |
if (indices.size() < 1) |
| 249 |
{ |
| 250 |
usage(); |
| 251 |
return 0; |
| 252 |
} |
| 253 |
|
| 254 |
if (indexMode) |
| 255 |
{ |
| 256 |
if (indices.size() > 1) |
| 257 |
{ |
| 258 |
cerr << program << ": Too many indices, can only build one index" |
| 259 |
<< " at a time\n"; |
| 260 |
return 1; |
| 261 |
} |
| 262 |
|
| 263 |
if (indexDomains.size() < 1) |
| 264 |
{ |
| 265 |
cerr << program << ": Must specify at least one domain\n"; |
| 266 |
return 1; |
| 267 |
} |
| 268 |
|
| 269 |
Indexer indexer(indices[0], indexDomains, indexRestrictions); |
| 270 |
|
| 271 |
indexer.index(indexURL); |
| 272 |
} |
| 273 |
else |
| 274 |
{ |
| 275 |
string line; |
| 276 |
getline(cin, line); |
| 277 |
query = line; |
| 278 |
|
| 279 |
Searcher searcher(query); |
| 280 |
|
| 281 |
searcher.search(indices); |
| 282 |
|
| 283 |
Outputer outputer(header, body, footer, notfound, |
| 284 |
pages); |
| 285 |
|
| 286 |
outputer.output(searcher, page < 1 ? page : --page); |
| 287 |
} |
| 288 |
|
| 289 |
return 0; |
| 290 |
} |
| 291 |
|
| 292 |
string agent(bool version) |
| 293 |
{ |
| 294 |
string agent = programName + (version ? ('/' + programVersion) : ""); |
| 295 |
|
| 296 |
return agent; |
| 297 |
} |
| 298 |
|
| 299 |
string platform() |
| 300 |
{ |
| 301 |
string platform; |
| 302 |
string os; |
| 303 |
string version; |
| 304 |
string architecture; |
| 305 |
string marketing; |
| 306 |
|
| 307 |
#ifdef _WIN32 |
| 308 |
OSVERSIONINFO* computer = new OSVERSIONINFO; |
| 309 |
computer->dwOSVersionInfoSize = sizeof(OSVERSIONINFO); |
| 310 |
GetVersionEx(computer); |
| 311 |
|
| 312 |
os = computer->dwPlatformId == VER_PLATFORM_WIN32_NT ? "Windows NT" : |
| 313 |
"Windows"; |
| 314 |
unsigned major = computer->dwMajorVersion; |
| 315 |
unsigned minor = computer->dwMinorVersion; |
| 316 |
|
| 317 |
delete computer; |
| 318 |
|
| 319 |
SYSTEM_INFO* system = new SYSTEM_INFO; |
| 320 |
GetSystemInfo(system); |
| 321 |
|
| 322 |
switch (system->wProcessorArchitecture) |
| 323 |
{ |
| 324 |
case PROCESSOR_ARCHITECTURE_INTEL: |
| 325 |
architecture = "ix86"; |
| 326 |
break; |
| 327 |
case PROCESSOR_ARCHITECTURE_MIPS: |
| 328 |
architecture = "mips"; |
| 329 |
break; |
| 330 |
case PROCESSOR_ARCHITECTURE_ALPHA: |
| 331 |
architecture = "alpha"; |
| 332 |
break; |
| 333 |
case PROCESSOR_ARCHITECTURE_PPC: |
| 334 |
architecture = "ppc"; |
| 335 |
break; |
| 336 |
case PROCESSOR_ARCHITECTURE_IA64: |
| 337 |
architecture = "ia64"; |
| 338 |
break; |
| 339 |
case PROCESSOR_ARCHITECTURE_IA32_ON_WIN64: |
| 340 |
architecture = "ix86_on_win64"; |
| 341 |
break; |
| 342 |
case PROCESSOR_ARCHITECTURE_AMD64: |
| 343 |
architecture = "amd64"; |
| 344 |
break; |
| 345 |
default: |
| 346 |
architecture = "unknown"; |
| 347 |
break; |
| 348 |
} |
| 349 |
|
| 350 |
char* cversion = new char[1024]; |
| 351 |
sprintf(cversion, "%u.%u", major, minor); |
| 352 |
version = cversion; |
| 353 |
|
| 354 |
delete [] cversion; |
| 355 |
|
| 356 |
if (major == 4 && minor <= 3 && os != "Windows NT") |
| 357 |
{ |
| 358 |
marketing = " [Windows 95]"; |
| 359 |
} |
| 360 |
else if (major == 4 && minor == 10 && os != "Windows NT") |
| 361 |
{ |
| 362 |
marketing = " [Windows 98]"; |
| 363 |
} |
| 364 |
else if (major == 5 && minor == 0 && os == "Windows NT") |
| 365 |
{ |
| 366 |
marketing = " [Windows 2000]"; |
| 367 |
} |
| 368 |
else if (major == 4 && minor == 90 && os != "Windows NT") |
| 369 |
{ |
| 370 |
marketing = " [Windows ME]"; |
| 371 |
} |
| 372 |
else if (major == 5 && minor == 1 && os == "Windows NT") |
| 373 |
{ |
| 374 |
marketing = " [Windows XP]"; |
| 375 |
} |
| 376 |
else if (major == 5 && minor == 2 && os == "Windows NT") |
| 377 |
{ |
| 378 |
marketing = " [Windows .NET Server]"; |
| 379 |
} |
| 380 |
#else // _WIN32 |
| 381 |
struct utsname* computer = new struct utsname; |
| 382 |
uname(computer); |
| 383 |
|
| 384 |
os = computer->sysname; |
| 385 |
version = computer->release; |
| 386 |
architecture = computer->machine; |
| 387 |
|
| 388 |
delete computer; |
| 389 |
#endif // _WIN32 |
| 390 |
|
| 391 |
platform = "(" + os + " " + version + marketing + " " + architecture + ")"; |
| 392 |
|
| 393 |
return platform; |
| 394 |
} |
| 395 |
|
| 396 |
void usage() |
| 397 |
{ |
| 398 |
string tab(8 + program.length(), ' '); |
| 399 |
|
| 400 |
cout << "Usage: " << program << " [index ...] [-P page] [-h header] [-b bo" |
| 401 |
<< "dy]\n" |
| 402 |
<< tab << "[-f footer] [-n notfound] [-p pages]\n" |
| 403 |
<< tab << "[-i begin] [-d domain ...] [-r restriction ...]\n" |
| 404 |
<< tab << "[-D] [-version] [-help]\n" |
| 405 |
<< "Options:\n" |
| 406 |
<< " index Index file to use (can only use one file for i" |
| 407 |
<< "ndexing)\n" |
| 408 |
<< " -P page Page of search to display (defaults to 1)\n" |
| 409 |
<< " -h header Header template to use (defaults to header.htm" |
| 410 |
<< "l)\n" |
| 411 |
<< " -b body Body template to use (defaults to body.html)\n" |
| 412 |
<< " -f footer Footer template to use (defaults to footer.htm" |
| 413 |
<< "l)\n" |
| 414 |
<< " -n notfound Not found template to use (defaults to notfoun" |
| 415 |
<< "d.html)\n" |
| 416 |
<< " -p pages Pages template to use (defaults to pages.html)" |
| 417 |
<< "\n" |
| 418 |
<< " -i begin URL to begin indexing (causes indexing rather " |
| 419 |
<< "than search)\n" |
| 420 |
<< " -d domain Domain to include in indexing\n" |
| 421 |
<< " -r restriction URL to restrict from indexing\n" |
| 422 |
<< " -D Display debug information\n" |
| 423 |
<< " -version Display version information and exit\n" |
| 424 |
<< " -license Display license information and exit\n" |
| 425 |
<< " -help Display this message and exit\n"; |
| 426 |
} |
| 427 |
|
| 428 |
void version() |
| 429 |
{ |
| 430 |
cout << programName << " " << programVersion << " "<< platform() << "\n\n" |
| 431 |
<< " Copyright (C) 2002-2003, Douglas Thrift. All Rights Reserved.\n" |
| 432 |
<< "\n" |
| 433 |
<< " This product includes software developed by Douglas Thrift\n" |
| 434 |
<< " (http://computers.douglasthrift.net/searchengine/).\n"; |
| 435 |
} |
| 436 |
|
| 437 |
void license() |
| 438 |
{ |
| 439 |
cout << "License:\n" |
| 440 |
<< " Douglas Thrift's Search Engine License\n\n" |
| 441 |
<< " Copyright (C) 2002-2003, Douglas Thrift. All Rights Reserved.\n" |
| 442 |
<< "\n" |
| 443 |
<< " Redistribution and use in source and binary forms, with or with" |
| 444 |
<< "out\n" |
| 445 |
<< " modification, are permitted provided that the following conditi" |
| 446 |
<< "ons are met:\n\n" |
| 447 |
<< " 1. Redistributions of source code must retain the above copyrig" |
| 448 |
<< "ht notice,\n" |
| 449 |
<< " this list of conditions and the following disclaimer.\n\n" |
| 450 |
<< " 2. Redistributions in binary form must reproduce the above copy" |
| 451 |
<< "right notice,\n" |
| 452 |
<< " this list of conditions and the following disclaimer in the " |
| 453 |
<< "documentation\n" |
| 454 |
<< " and/or other materials provided with the distribution.\n\n" |
| 455 |
<< " 3. The end-user documentation included with the redistribution," |
| 456 |
<< " if any, must\n" |
| 457 |
<< " include the following acknowledgment:\n\n" |
| 458 |
<< " \"This product includes software developed by Douglas Thr" |
| 459 |
<< "ift\n" |
| 460 |
<< " (http://computers.douglasthrift.net/searchengine/).\"\n\n" |
| 461 |
<< " Alternately, this acknowledgment may appear in the software " |
| 462 |
<< "itself, if\n" |
| 463 |
<< " and wherever such third-party acknowledgments normally appea" |
| 464 |
<< "r.\n\n" |
| 465 |
<< " 4. The names \"Douglas Thrift\" and \"Douglas Thrift\'s Search " |
| 466 |
<< "Engine\" must not\n" |
| 467 |
<< " be used to endorse or promote products derived from this sof" |
| 468 |
<< "tware without\n" |
| 469 |
<< " specific prior written permission. For written permission, p" |
| 470 |
<< "lease visit\n" |
| 471 |
<< " http://www.douglasthrift.net/contact.cgi for contact inform" |
| 472 |
<< "ation.\n\n" |
| 473 |
<< " 5. Products derived from this software may not be called \"Doug" |
| 474 |
<< "las Thrift\'s\n" |
| 475 |
<< " Search Engine\", nor may \"Douglas Thrift\'s Search Engine\"" |
| 476 |
<< " appear in their\n" |
| 477 |
<< " name, without prior written permission.\n\n" |
| 478 |
<< " THIS SOFTWARE IS PROVIDED \"AS IS\" AND ANY EXPRESS OR IMPLIED " |
| 479 |
<< "WARRANTIES,\n" |
| 480 |
<< " INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHA" |
| 481 |
<< "NTABILITY AND\n" |
| 482 |
<< " FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SH" |
| 483 |
<< "ALL THE\n" |
| 484 |
<< " COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIR" |
| 485 |
<< "ECT,\n" |
| 486 |
<< " INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLU" |
| 487 |
<< "DING, BUT NOT\n" |
| 488 |
<< " LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS O" |
| 489 |
<< "F USE, DATA,\n" |
| 490 |
<< " OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY" |
| 491 |
<< " THEORY OF\n" |
| 492 |
<< " LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCL" |
| 493 |
<< "UDING\n" |
| 494 |
<< " NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF T" |
| 495 |
<< "HIS SOFTWARE,\n" |
| 496 |
<< " EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.\n"; |
| 497 |
} |
| 498 |
|
| 499 |
void entities(string& line, char character, char* entity) |
| 500 |
{ |
| 501 |
int begin = 0; |
| 502 |
|
| 503 |
while (begin < line.length()) |
| 504 |
{ |
| 505 |
int spot = line.find(character, begin); |
| 506 |
|
| 507 |
int end = spot + 1; |
| 508 |
|
| 509 |
if (spot != string::npos) |
| 510 |
{ |
| 511 |
line.replace(spot, 1, entity); |
| 512 |
} |
| 513 |
else |
| 514 |
{ |
| 515 |
break; |
| 516 |
} |
| 517 |
|
| 518 |
begin = end; |
| 519 |
} |
| 520 |
} |
| 521 |
|
| 522 |
void entities(string& line, char* entity, char character) |
| 523 |
{ |
| 524 |
int begin = 0; |
| 525 |
|
| 526 |
while (begin < line.length()) |
| 527 |
{ |
| 528 |
int spot = line.find(entity, begin); |
| 529 |
|
| 530 |
int end = spot + 1; |
| 531 |
|
| 532 |
if (spot != string::npos) |
| 533 |
{ |
| 534 |
line.replace(spot, strlen(entity), 1, character); |
| 535 |
} |
| 536 |
else |
| 537 |
{ |
| 538 |
break; |
| 539 |
} |
| 540 |
|
| 541 |
begin = end; |
| 542 |
} |
| 543 |
} |
| 544 |
|
| 545 |
void normalize(string& abbynormal) |
| 546 |
{ |
| 547 |
for (unsigned index = 0; index < abbynormal.length(); index++) |
| 548 |
{ |
| 549 |
if (isspace(abbynormal[index])) |
| 550 |
{ |
| 551 |
unsigned next = index + 1; |
| 552 |
while (isspace(abbynormal[next])) |
| 553 |
{ |
| 554 |
next++; |
| 555 |
} |
| 556 |
abbynormal.replace(index, next - index, 1, abbynormal[index]); |
| 557 |
} |
| 558 |
} |
| 559 |
|
| 560 |
if (isspace(abbynormal[0])) abbynormal.erase(0, 1); |
| 561 |
if (isspace(abbynormal[abbynormal.length() - 1])) |
| 562 |
abbynormal.erase(abbynormal.length() - 1, 1); |
| 563 |
} |
