| 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 Searcher |
| 46 |
// |
| 47 |
// Douglas Thrift |
| 48 |
// |
| 49 |
// Searcher.cpp |
| 50 |
|
| 51 |
#include "Searcher.h" |
| 52 |
|
| 53 |
Searcher::Searcher(string& queryString) |
| 54 |
{ |
| 55 |
setCommon(); |
| 56 |
setAnd(false); |
| 57 |
setOr(false); |
| 58 |
|
| 59 |
this->queryString = queryString; |
| 60 |
|
| 61 |
setQuery(); |
| 62 |
} |
| 63 |
|
| 64 |
void Searcher::search(vector<string> indices) |
| 65 |
{ |
| 66 |
start = clock(); |
| 67 |
|
| 68 |
const string XMLTYPE = "<?xml version=\"1.0\" encoding=\"ISO-8859-1\" stan" |
| 69 |
+ string("dalone=\"no\"?>"); |
| 70 |
const string DOCTYPE = "<!DOCTYPE index SYSTEM \"index.dtd\">"; |
| 71 |
|
| 72 |
if (query.size() > 0) |
| 73 |
{ |
| 74 |
for (int index = 0; index < indices.size(); index++) |
| 75 |
{ |
| 76 |
ifstream fin(indices[index].c_str()); |
| 77 |
|
| 78 |
if (!fin.is_open()) |
| 79 |
{ |
| 80 |
cerr << program << ": Could not open index file: " |
| 81 |
<< indices[index] << "\n"; |
| 82 |
fin.clear(); |
| 83 |
|
| 84 |
continue; |
| 85 |
} |
| 86 |
|
| 87 |
string line; |
| 88 |
|
| 89 |
getline(fin, line); |
| 90 |
if (line != XMLTYPE) |
| 91 |
{ |
| 92 |
cerr << program << ": Invalid XML version declaration: " |
| 93 |
<< indices[index] << "\n"; |
| 94 |
fin.close(); |
| 95 |
fin.clear(); |
| 96 |
|
| 97 |
continue; |
| 98 |
} |
| 99 |
|
| 100 |
getline(fin, line); |
| 101 |
if (line != DOCTYPE) |
| 102 |
{ |
| 103 |
cerr << program << ": Invalid XML doctype: " << indices[index] |
| 104 |
<< "\n"; |
| 105 |
fin.close(); |
| 106 |
fin.clear(); |
| 107 |
continue; |
| 108 |
} |
| 109 |
|
| 110 |
getline(fin, line); |
| 111 |
if (line != "<index>") |
| 112 |
{ |
| 113 |
cerr << program << ": Invalid XML root element: " |
| 114 |
<< indices[index] << "\n"; |
| 115 |
fin.close(); |
| 116 |
fin.clear(); |
| 117 |
continue; |
| 118 |
} |
| 119 |
|
| 120 |
while (fin.good()) |
| 121 |
{ |
| 122 |
Page page; |
| 123 |
|
| 124 |
fin >> page; |
| 125 |
|
| 126 |
if (!page.empty()) search(page); |
| 127 |
} |
| 128 |
|
| 129 |
fin.close(); |
| 130 |
fin.clear(); |
| 131 |
} |
| 132 |
} |
| 133 |
|
| 134 |
finish = clock(); |
| 135 |
} |
| 136 |
|
| 137 |
void Searcher::search(Page& page) |
| 138 |
{ |
| 139 |
Ranker ranker(page); |
| 140 |
ranker.rank(query); |
| 141 |
|
| 142 |
if (ranker != 0) |
| 143 |
{ |
| 144 |
ranker.setSample(); |
| 145 |
pages.insert(ranker); |
| 146 |
} |
| 147 |
} |
| 148 |
|
| 149 |
void Searcher::setQuery() |
| 150 |
{ |
| 151 |
string queryStringNormal = queryString; |
| 152 |
normalize(queryStringNormal); |
| 153 |
|
| 154 |
if (debug) cerr << "queryString = " << queryString << "\n" |
| 155 |
<< "queryStringNormal = " << queryStringNormal << "\n"; |
| 156 |
|
| 157 |
unsigned begin = 0; |
| 158 |
bool capitalOr = false, quotes = false; |
| 159 |
while (begin < queryStringNormal.length()) |
| 160 |
{ |
| 161 |
string word; |
| 162 |
|
| 163 |
unsigned beginQuote = queryStringNormal.find('\"', begin); |
| 164 |
unsigned endQuote = beginQuote != string::npos ? |
| 165 |
queryStringNormal.find('\"', beginQuote + 1) : string::npos; |
| 166 |
unsigned space = queryStringNormal.find(' ', begin); |
| 167 |
|
| 168 |
if (beginQuote == begin) |
| 169 |
{ |
| 170 |
quotes = true; |
| 171 |
word = queryStringNormal.substr(beginQuote + 1, endQuote - |
| 172 |
beginQuote - 1); |
| 173 |
} |
| 174 |
else if (beginQuote == (begin + 8) && queryStringNormal.substr(begin, 8 |
| 175 |
) == "intitle:") |
| 176 |
{ |
| 177 |
quotes = true; |
| 178 |
word = "intitle:" + queryStringNormal.substr(beginQuote + 1, |
| 179 |
endQuote - beginQuote - 1); |
| 180 |
beginQuote = begin; |
| 181 |
} |
| 182 |
else if (beginQuote == (begin + 6) && queryStringNormal.substr(begin, 6 |
| 183 |
) == "inurl:") |
| 184 |
{ |
| 185 |
quotes = true; |
| 186 |
word = "inurl:" + queryStringNormal.substr(beginQuote + 1, endQuote |
| 187 |
- beginQuote - 1); |
| 188 |
beginQuote = begin; |
| 189 |
} |
| 190 |
else if (beginQuote == (begin + 7) && queryStringNormal.substr(begin, 7 |
| 191 |
) == "intext:") |
| 192 |
{ |
| 193 |
quotes = true; |
| 194 |
word = "intext:" + queryStringNormal.substr(beginQuote + 1, |
| 195 |
endQuote - beginQuote - 1); |
| 196 |
beginQuote = begin; |
| 197 |
} |
| 198 |
else |
| 199 |
{ |
| 200 |
quotes = false; |
| 201 |
word = queryStringNormal.substr(begin, space - begin); |
| 202 |
} |
| 203 |
|
| 204 |
if (word != "") |
| 205 |
{ |
| 206 |
string lowerWord(word.length(), ' '); |
| 207 |
|
| 208 |
for (unsigned index = 0; index < word.length(); index++) |
| 209 |
{ |
| 210 |
lowerWord[index] = tolower(word[index]); |
| 211 |
} |
| 212 |
|
| 213 |
if (debug) cerr << "word = " << word << "\n" |
| 214 |
<< "lowerWord = " << lowerWord << "\n"; |
| 215 |
|
| 216 |
if (word == "OR" && !quotes) |
| 217 |
{ |
| 218 |
capitalOr = true; |
| 219 |
query[query.size() - 1] += " OR"; |
| 220 |
} |
| 221 |
else if (lowerWord == "or" && !quotes) |
| 222 |
{ |
| 223 |
setOr(true); |
| 224 |
} |
| 225 |
else if (lowerWord == "and" && !quotes) |
| 226 |
{ |
| 227 |
setAnd(true); |
| 228 |
} |
| 229 |
else if (common.find(lowerWord) != common.end() && !quotes) |
| 230 |
{ |
| 231 |
commonUsed.push_back(word); |
| 232 |
} |
| 233 |
else |
| 234 |
{ |
| 235 |
if (capitalOr) capitalOr = false; |
| 236 |
if (query.size() < 10) |
| 237 |
{ |
| 238 |
query.push_back(lowerWord); |
| 239 |
} |
| 240 |
else |
| 241 |
{ |
| 242 |
setIgnore(lowerWord); |
| 243 |
|
| 244 |
if (debug) cerr << "ignore = " << ignore << "\n"; |
| 245 |
|
| 246 |
break; |
| 247 |
} |
| 248 |
} |
| 249 |
} |
| 250 |
|
| 251 |
if (endQuote == string::npos && space == string::npos) |
| 252 |
{ |
| 253 |
begin = string::npos; |
| 254 |
} |
| 255 |
else if (beginQuote == begin && endQuote != string::npos) |
| 256 |
{ |
| 257 |
begin = endQuote + 1; |
| 258 |
} |
| 259 |
else |
| 260 |
{ |
| 261 |
begin = space + 1; |
| 262 |
} |
| 263 |
} |
| 264 |
|
| 265 |
if (debug) |
| 266 |
{ |
| 267 |
cerr << "query = {\n"; |
| 268 |
|
| 269 |
for (unsigned index = 0; index < query.size(); index++) |
| 270 |
{ |
| 271 |
cerr << " [" << index << "] = " << query[index] << "\n"; |
| 272 |
} |
| 273 |
|
| 274 |
cerr << "}\n"; |
| 275 |
} |
| 276 |
} |
| 277 |
|
| 278 |
void Searcher::setCommon() |
| 279 |
{ |
| 280 |
common.insert("&"); |
| 281 |
common.insert("a"); |
| 282 |
common.insert("about"); |
| 283 |
common.insert("an"); |
| 284 |
// "and" has its own special message |
| 285 |
// common.insert("and"); |
| 286 |
common.insert("are"); |
| 287 |
common.insert("as"); |
| 288 |
common.insert("at"); |
| 289 |
common.insert("be"); |
| 290 |
common.insert("by"); |
| 291 |
common.insert("com"); |
| 292 |
common.insert("from"); |
| 293 |
common.insert("how"); |
| 294 |
common.insert("i"); |
| 295 |
common.insert("in"); |
| 296 |
common.insert("is"); |
| 297 |
common.insert("it"); |
| 298 |
common.insert("of"); |
| 299 |
common.insert("on"); |
| 300 |
// "or" has its own special message |
| 301 |
// common.insert("or"); |
| 302 |
common.insert("that"); |
| 303 |
common.insert("the"); |
| 304 |
common.insert("this"); |
| 305 |
common.insert("to"); |
| 306 |
common.insert("was"); |
| 307 |
common.insert("what"); |
| 308 |
common.insert("when"); |
| 309 |
common.insert("where"); |
| 310 |
common.insert("which"); |
| 311 |
common.insert("who"); |
| 312 |
common.insert("why"); |
| 313 |
common.insert("will"); |
| 314 |
common.insert("with"); |
| 315 |
} |
