ViewVC Help
View File | Revision Log | Show Annotations | Download File | View Changeset | Root Listing
root/proj/trunk/Search/Outputter.cpp
Revision: 205
Committed: 2003-07-17T00:03:08-07:00 (21 years, 11 months ago) by douglas
Original Path: trunk/Search/Outputer.cpp
File size: 12897 byte(s)
Log Message:
Moved openssl() to Search.h and fixed <?version?> tag to use it.

File Contents

# User Rev Content
1 douglas 1 /* ============================================================================
2     * Douglas Thrift's Search Engine License
3     *
4 douglas 28 * Copyright (C) 2002-2003, Douglas Thrift. All Rights Reserved.
5 douglas 1 * 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 Outputer
46     //
47     // Douglas Thrift
48     //
49 douglas 205 // $Id: Outputer.cpp,v 1.8 2003/07/17 07:03:08 douglas Exp $
50 douglas 1
51     #include "Outputer.h"
52    
53 douglas 28 Outputer::Outputer(const string& headerFile, const string& bodyFile, const
54 douglas 15 string& footerFile, const string& notfoundFile, const string& pagesFile)
55 douglas 1 {
56     this->headerFile = headerFile;
57     this->bodyFile = bodyFile;
58     this->footerFile = footerFile;
59     this->notfoundFile = notfoundFile;
60     this->pagesFile = pagesFile;
61     }
62    
63     void Outputer::output(Searcher& searcher, unsigned page)
64     {
65     MultiSet pagesSet = searcher.getPages();
66     numWebpages = pagesSet.size();
67     numPages = (numWebpages + 9) / 10;
68     string query = searcher.getQueryString();
69     vector<string> common = searcher.getCommonUsed();
70    
71     MultiSetIterator itor = pagesSet.begin();
72    
73     for (int count = 0; count < page * 10 && itor != pagesSet.end(); count++)
74     {
75     itor++;
76     }
77    
78     for (int index = 0; index < 10 && itor != pagesSet.end(); index++, itor++)
79     {
80     webpages.push_back(*itor);
81     }
82    
83     this->query = searcher.getQuery().size() > 0;
84     results = webpages.size() > 0;
85     time = searcher.time();
86    
87     if (debug)
88     {
89 douglas 183 cerr << "query = " << this->query << "\n"
90     << "results = " << results << "\n"
91 douglas 1 << "time = " << duration() << "\n";
92     }
93    
94     entities(query, '&', "&amp;");
95     entities(query, '\"', "&quot;");
96     entities(query, '<', "&lt;");
97     entities(query, '>', "&gt;");
98    
99     string ignore = searcher.getIgnore();
100    
101     header(query, page, common, searcher.getAnd(), searcher.getOr(),
102     ignore);
103    
104     if (results)
105     {
106     body();
107     }
108     else if (this->query)
109     {
110     notfound(query, searcher.getQuery().size());
111     }
112    
113     footer(query, page, common, searcher.getAnd(), searcher.getOr(),
114     ignore);
115     }
116    
117 douglas 15 void Outputer::header(const string& query, unsigned page, vector<string>
118     common, bool and_, bool or_, const string& ignore)
119 douglas 1 {
120     ifstream fin(headerFile.c_str());
121    
122     string line;
123     while (fin.good())
124     {
125     getline(fin, line);
126    
127     conditional(line, fin, "<?ifquery?>", this->query);
128     conditional(line, fin, "<?ifresults?>", results);
129     conditional(line, fin, "<?ifor?>", or_);
130     conditional(line, fin, "<?ifand?>", and_);
131     conditional(line, fin, "<?ifignore?>", ignore != "");
132     conditional(line, fin, "<?ifcommon?>", common.size() == 1);
133     conditional(line, fin, "<?ifmanycommon?>", common.size() > 1);
134    
135 douglas 43 tag(line, "<?version?>", programName + ' ' + programVersion + ' ' +
136     platform());
137 douglas 1 tag(line, "<?query?>", query);
138 douglas 15 tag(line, "<?range?>", range(page));
139     tag(line, "<?total?>", total());
140     tag(line, "<?time?>", duration());
141     tag(line, "<?pages?>", pages(query, page));
142 douglas 1 tag(line, "<?ignore?>", ignore);
143     tag(line, "<?common?>", common[0]);
144 douglas 15 tag(line, "<?manycommon?>", manycommon(common));
145 douglas 1
146 douglas 30 cout << line << (fin.good() ? "\n" : "");
147 douglas 1 }
148    
149     fin.close();
150     }
151    
152     void Outputer::body()
153     {
154     for (int index = 0; index < webpages.size(); index++)
155     {
156     Ranker webpage = webpages[index];
157     string title = webpage.getTitle();
158     if (title == "")
159     {
160     title = webpage.getURL();
161     entities(title, '&', "&amp;");
162     entities(title, '\"', "&quot;");
163     entities(title, '<', "&lt;");
164     entities(title, '>', "&gt;");
165     }
166     string address = webpage.getURL();
167     string sample = webpage.getSample();
168     string description = webpage.getDescription();
169    
170     char* csize = new char[1024];
171     sprintf(csize, "%.0fk", (double(webpage.getSize()) / double(1024)));
172    
173     string size = csize;
174    
175     delete [] csize;
176    
177     entities(address, '&', "&amp;");
178     entities(address, '\"', "&quot;");
179     entities(address, '<', "&lt;");
180     entities(address, '>', "&gt;");
181    
182     ifstream fin(bodyFile.c_str());
183    
184     string line;
185     while (fin.good())
186     {
187     getline(fin, line);
188    
189     conditional(line, fin, "<?ifdescription?>", description != "");
190    
191     tag(line, "<?address?>", address);
192     tag(line, "<?title?>", title);
193     tag(line, "<?sample?>", sample);
194     tag(line, "<?description?>", description);
195     tag(line, "<?size?>", size);
196    
197 douglas 30 cout << line << (fin.good() ? "\n" : "");
198 douglas 1 }
199    
200     fin.close();
201     }
202     }
203    
204 douglas 15 void Outputer::footer(const string& query, unsigned page, vector<string>
205     common, bool and_, bool or_, const string& ignore)
206 douglas 1 {
207     ifstream fin(footerFile.c_str());
208    
209     string line;
210     while (fin.good())
211     {
212     getline(fin, line);
213    
214     conditional(line, fin, "<?ifquery?>", this->query);
215     conditional(line, fin, "<?ifresults?>", results);
216     conditional(line, fin, "<?ifor?>", or_);
217     conditional(line, fin, "<?ifand?>", and_);
218     conditional(line, fin, "<?ifignore?>", ignore != "");
219     conditional(line, fin, "<?ifcommon?>", common.size() == 1);
220     conditional(line, fin, "<?ifmanycommon?>", common.size() > 1);
221    
222 douglas 205 #ifndef _OpenSSL_
223 douglas 43 tag(line, "<?version?>", programName + ' ' + programVersion + ' ' +
224     platform());
225 douglas 205 #else
226     tag(line, "<?version?>", programName + ' ' + programVersion + ' ' +
227     platform() + ' ' + openssl());
228     #endif
229 douglas 1 tag(line, "<?query?>", query);
230 douglas 15 tag(line, "<?range?>", range(page));
231     tag(line, "<?total?>", total());
232     tag(line, "<?time?>", duration());
233     tag(line, "<?pages?>", pages(query, page));
234 douglas 1 tag(line, "<?ignore?>", ignore);
235     tag(line, "<?common?>", common[0]);
236 douglas 15 tag(line, "<?manycommon?>", manycommon(common));
237 douglas 1
238 douglas 30 cout << line << (fin.good() ? "\n" : "");
239 douglas 1 }
240    
241     fin.close();
242     }
243    
244 douglas 15 void Outputer::notfound(const string& query, unsigned keywords)
245 douglas 1 {
246     ifstream fin(notfoundFile.c_str());
247    
248     string line;
249     while (fin.good())
250     {
251     getline(fin, line);
252    
253     conditional(line, fin, "<?ifmany?>", keywords > 1);
254    
255     tag(line, "<?query?>", query);
256    
257 douglas 30 cout << line << (fin.good() ? "\n" : "");
258 douglas 1 }
259    
260     fin.close();
261     }
262    
263     string Outputer::pages(string query, unsigned page)
264     {
265     entities(query, "&lt;", '<');
266     entities(query, "&gt;", '>');
267     entities(query, "&quot;", '\"');
268     entities(query, "&amp;", '&');
269    
270     entities(query, '%', "%25");
271     entities(query, '\t', "%09");
272     entities(query, ' ', "%20");
273     entities(query, '\"', "%22");
274     entities(query, '#', "%23");
275     entities(query, '$', "%24");
276     entities(query, '&', "%26");
277     entities(query, '\'', "%27");
278     entities(query, '+', "%2B");
279     entities(query, ',', "%2C");
280     entities(query, '/', "%2F");
281     entities(query, ':', "%3A");
282     entities(query, ';', "%3B");
283     entities(query, '<', "%3C");
284     entities(query, '=', "%3D");
285     entities(query, '>', "%3E");
286     entities(query, '?', "%3F");
287     entities(query, '@', "%40");
288     entities(query, '[', "%5B");
289     entities(query, ']', "%5D");
290     entities(query, '\\', "%5C");
291     entities(query, '^', "%5E");
292     entities(query, '`', "%60");
293     entities(query, '{', "%7B");
294     entities(query, '|', "%7C");
295     entities(query, '}', "%7D");
296     entities(query, '~', "%7E");
297    
298     string lines;
299    
300     ifstream fin(pagesFile.c_str());
301    
302     string line;
303     while (fin.good())
304     {
305     getline(fin, line);
306     conditional(line, fin, "<?ifprevious?>", page >= 1);
307     conditional(line, fin, "<?ifpage?>", false);
308     conditional(line, fin, "<?ifnum?>", false);
309     conditional(line, fin, "<?ifnext?>", false);
310    
311     char* cprevious = new char[1024];
312    
313     sprintf(cprevious, "%u", page);
314    
315     string previous = cprevious;
316    
317     delete [] cprevious;
318    
319     tag(line, "<?query?>", query);
320     tag(line, "<?previous?>", previous);
321    
322 douglas 30 lines += line + (fin.good() ? "\n" : "");
323 douglas 1 }
324    
325     fin.close();
326     fin.clear();
327    
328     for (int index = 0; index < numPages; index++)
329     {
330     fin.open(pagesFile.c_str());
331    
332     while (fin.good())
333     {
334     getline(fin, line);
335     if (index == page)
336     {
337     conditional(line, fin, "<?ifprevious?>", false);
338     conditional(line, fin, "<?ifpage?>", true);
339     conditional(line, fin, "<?ifnum?>", false);
340     conditional(line, fin, "<?ifnext?>", false);
341    
342     char* cpage = new char[1024];
343    
344     sprintf(cpage, "%u", (index + 1));
345    
346     string spage = cpage;
347    
348     delete [] cpage;
349    
350     tag(line, "<?page?>", spage);
351     }
352     else
353     {
354     conditional(line, fin, "<?ifprevious?>", false);
355     conditional(line, fin, "<?ifpage?>", false);
356     conditional(line, fin, "<?ifnum?>", true);
357     conditional(line, fin, "<?ifnext?>", false);
358    
359     char* cnum = new char[1024];
360    
361     sprintf(cnum, "%u", (index + 1));
362    
363     string num = cnum;
364    
365     delete [] cnum;
366    
367     tag(line, "<?query?>", query);
368     tag(line, "<?num?>", num);
369     }
370    
371 douglas 30 lines += line + (fin.good() ? "\n" : "");
372 douglas 1 }
373    
374     fin.close();
375     fin.clear();
376     }
377    
378     fin.open(pagesFile.c_str());
379    
380     while (fin.good())
381     {
382     getline(fin, line);
383     conditional(line, fin, "<?ifprevious?>", false);
384     conditional(line, fin, "<?ifpage?>", false);
385     conditional(line, fin, "<?ifnum?>", false);
386     conditional(line, fin, "<?ifnext?>", page + 2 <= numPages);
387    
388     char* cnext = new char[1024];
389    
390     sprintf(cnext, "%u", (page + 2));
391    
392     string next = cnext;
393    
394     delete [] cnext;
395    
396     tag(line, "<?query?>", query);
397     tag(line, "<?next?>", next);
398    
399 douglas 30 lines += line + (fin.good() ? "\n" : "");
400 douglas 1 }
401    
402     fin.close();
403    
404     return lines;
405     }
406    
407     string Outputer::range(unsigned page)
408     {
409     unsigned bottom = page * 10 + 1;
410     unsigned top = numWebpages > page * 10 + 10 ? page * 10 + 10 : numWebpages;
411    
412     char* cbottom = new char[1024];
413     char* ctop = new char[1024];
414    
415     sprintf(cbottom, "%u", bottom);
416     sprintf(ctop, "%u", top);
417    
418     string range = string(cbottom) + " - " + ctop;
419    
420     delete [] cbottom;
421     delete [] ctop;
422    
423     return range;
424     }
425    
426     string Outputer::total()
427     {
428     char* ctotal = new char[1024];
429    
430     sprintf(ctotal, "%u", numWebpages);
431    
432     string total = ctotal;
433    
434     delete [] ctotal;
435    
436     return total;
437     }
438    
439     string Outputer::duration()
440     {
441     char* ctime = new char[1024];
442     sprintf(ctime, "%.2f", time);
443    
444     string duration = ctime;
445    
446     delete [] ctime;
447    
448     return duration;
449     }
450    
451     string Outputer::manycommon(vector<string> common)
452     {
453     string line;
454    
455     for (int index = 0; index < common.size(); index++)
456     {
457     line += common[index];
458    
459     if (index != common.size() - 1) line += ' ';
460     }
461    
462     return line;
463     }
464    
465 douglas 15 void Outputer::tag(string& line, char* tag, const string& replacement)
466 douglas 1 {
467     int begin = 0;
468     while (begin < line.length())
469     {
470     int spot = line.find(tag, begin);
471    
472     if (spot != string::npos)
473     {
474     line.replace(spot, strlen(tag), replacement);
475     }
476     else
477     {
478     break;
479     }
480    
481     begin = spot + replacement.length();
482     }
483     }
484    
485     void Outputer::conditional(string& line, ifstream& fin, char* tag, bool
486     condition)
487     {
488     unsigned begin = 0;
489     while (begin < line.length())
490     {
491     unsigned start = line.find(tag, begin);
492     unsigned finish = line.find("<?endif?>", start);
493    
494     if (start == string::npos) break;
495    
496     string next;
497     while (finish == string::npos)
498     {
499     getline(fin, next);
500     line += '\n' + next;
501     finish = line.find("<?endif?>", start);
502     }
503    
504     if (condition)
505     {
506     line.erase(start, strlen(tag));
507     line.erase(finish - strlen(tag), 9);
508    
509     begin = finish - strlen(tag) - 9;
510     }
511     else
512     {
513     line.erase(start, finish - start + 9);
514    
515     begin = start;
516     }
517     }
518     }