topsites.c

   1 /*
   2  * SARG Squid Analysis Report Generator      http://sarg.sourceforge.net
   3  *                                                            1998, 2015
   4  *
   5  * SARG donations:
   6  *      please look at http://sarg.sourceforge.net/donations.php
   7  * Support:
   8  *     http://sourceforge.net/projects/sarg/forums/forum/363374
   9  * ---------------------------------------------------------------------
  10  *
  11  *  This program is free software; you can redistribute it and/or modify
  12  *  it under the terms of the GNU General Public License as published by
  13  *  the Free Software Foundation; either version 2 of the License, or
  14  *  (at your option) any later version.
  15  *
  16  *  This program is distributed in the hope that it will be useful,
  17  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19  *  GNU General Public License for more details.
  20  *
  21  *  You should have received a copy of the GNU General Public License
  22  *  along with this program; if not, write to the Free Software
  23  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  24  *
  25  */
  26
  27 #include "include/conf.h"
  28 #include "include/defs.h"
  29
  30 #ifdef ENABLE_DOUBLE_CHECK_DATA
  31 extern struct globalstatstruct globstat;
  32 #endif
  33
  34 void topsites(void)
  35 {
  36         FILE *fp_in, *fp_ou;
  37
  38         char *buf;
  39         char *url;
  40         char *ourl=NULL;
  41         char csort[4096];
  42         char general[MAXLEN];
  43         char general2[MAXLEN];
  44         char general3[MAXLEN];
  45         char sites[MAXLEN];
  46         char report[MAXLEN];
  47         char ouser[MAX_USER_LEN]="";
  48         const char *sortf;
  49         const char *sortt;
  50         long long int nacc;
  51         long long int nbytes;
  52         long long int ntime;
  53         long long int tnacc=0;
  54         long long int tnbytes=0;
  55         long long int tntime=0;
  56         long long int twork1=0, twork2=0, twork3=0;
  57 #ifdef ENABLE_DOUBLE_CHECK_DATA
  58         long long int ttnacc=0;
  59         long long int ttnbytes=0;
  60         long long int ttntime=0;
  61 #endif
  62         int nusers=0;
  63         int regs=0;
  64         int cstatus;
  65         int url_len;
  66         int ourl_size=0;
  67         struct getwordstruct gwarea;
  68         longline line;
  69         struct generalitemstruct item;
  70
  71         if(Privacy) {
  72                 if (debugz>=LogLevel_Process) debugaz(__FILE__,__LINE__,_("Top sites report not produced because privacy option is on\n"));
  73                 return;
  74         }
  75         if (debugz>=LogLevel_Process)
  76                 debuga(__FILE__,__LINE__,_("Creating top sites report...\n"));
  77
  78         sprintf(general,"%s/sarg-general",outdirname);
  79         sprintf(sites,"%s/sarg-sites",outdirname);
  80         sprintf(general2,"%s/sarg-general2",outdirname);
  81         sprintf(general3,"%s/sarg-general3",outdirname);
  82
  83         sprintf(report,"%s/topsites.html",outdirname);
  84
  85         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" -k 4,4 -k 1,1 -o \"%s\" \"%s\"",general2,general)>=sizeof(csort)) {
  86                 debuga(__FILE__,__LINE__,_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general,general2);
  87                 exit(EXIT_FAILURE);
  88         }
  89         cstatus=system(csort);
  90         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
  91                 debuga(__FILE__,__LINE__,_("sort command return status %d\n"),WEXITSTATUS(cstatus));
  92                 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
  93                 exit(EXIT_FAILURE);
  94         }
  95
  96         if((fp_in=fopen(general2,"r"))==NULL) {
  97                 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),general2,strerror(errno));
  98                 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
  99                 exit(EXIT_FAILURE);
 100         }
 101
 102         if((fp_ou=fopen(general3,"w"))==NULL) {
 103                 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),general3,strerror(errno));
 104                 exit(EXIT_FAILURE);
 105         }
 106
 107         if ((line=longline_create())==NULL) {
 108                 debuga(__FILE__,__LINE__,_("Not enough memory to read file \"%s\"\n"),general2);
 109                 exit(EXIT_FAILURE);
 110         }
 111
 112         while((buf=longline_read(fp_in,line))!=NULL) {
 113                 ger_read(buf,&item,general2);
 114                 if(item.total) continue;
 115
 116                 if(!regs) {
 117                         url_len=strlen(item.url);
 118                         if (!ourl || url_len>=ourl_size) {
 119                                 ourl_size=url_len+1;
 120                                 ourl=realloc(ourl,ourl_size);
 121                                 if (!ourl) {
 122                                         debuga(__FILE__,__LINE__,_("Not enough memory to store the url\n"));
 123                                         exit(EXIT_FAILURE);
 124                                 }
 125                         }
 126                         strcpy(ourl,item.url);
 127                         regs++;
 128                 }
 129
 130                 if(strcmp(item.url,ourl) != 0) {
 131                         /*
 132                         This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 133                         to print a long long int unless it is exactly 64-bits long.
 134                         */
 135                         fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 136                         url_len=strlen(item.url);
 137                         if (url_len>=ourl_size) {
 138                                 ourl_size=url_len+1;
 139                                 ourl=realloc(ourl,ourl_size);
 140                                 if (!ourl) {
 141                                         debuga(__FILE__,__LINE__,_("Not enough memory to store the url\n"));
 142                                         exit(EXIT_FAILURE);
 143                                 }
 144                         }
 145                         strcpy(ourl,item.url);
 146                         strcpy(ouser,item.user);
 147                         tnacc=0;
 148                         tnbytes=0;
 149                         tntime=0;
 150                         nusers=1;
 151                 } else if (strcmp(item.user,ouser)!=0) {
 152                         strcpy(ouser,item.user);
 153                         nusers++;
 154                 }
 155
 156                 tnacc+=item.nacc;
 157                 tnbytes+=item.nbytes;
 158                 tntime+=item.nelap;
 159 #ifdef ENABLE_DOUBLE_CHECK_DATA
 160                 ttnacc+=item.nacc;
 161                 ttnbytes+=item.nbytes;
 162                 ttntime+=item.nelap;
 163 #endif
 164         }
 165         if (fclose(fp_in)==EOF) {
 166                 debuga(__FILE__,__LINE__,_("Read error in \"%s\": %s\n"),general2,strerror(errno));
 167                 exit(EXIT_FAILURE);
 168         }
 169         longline_destroy(&line);
 170
 171         if (ourl) {
 172                 /*
 173                 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 174                 to print a long long int unless it is exactly 64-bits long.
 175                 */
 176                 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 177                 free(ourl);
 178         }
 179
 180         if (fclose(fp_ou)==EOF) {
 181                 debuga(__FILE__,__LINE__,_("Write error in \"%s\": %s\n"),general3,strerror(errno));
 182                 exit(EXIT_FAILURE);
 183         }
 184
 185 #ifdef ENABLE_DOUBLE_CHECK_DATA
 186         if (ttnacc!=globstat.nacc || ttnbytes!=globstat.nbytes || ttntime!=globstat.elap) {
 187                 debuga(__FILE__,__LINE__,_("Total statistics mismatch when reading \"%s\" to produce the top sites\n"),general2);
 188                 exit(EXIT_FAILURE);
 189         }
 190 #endif
 191
 192         if (!KeepTempLog && unlink(general2)) {
 193                 debuga(__FILE__,__LINE__,_("Cannot delete \"%s\": %s\n"),general2,strerror(errno));
 194                 exit(EXIT_FAILURE);
 195         }
 196
 197         if((TopsitesSort & TOPSITE_SORT_CONNECT) != 0) {
 198                 sortf="-k 1,1 -k 2,2";
 199         } else if((TopsitesSort & TOPSITE_SORT_BYTES) != 0) {
 200                 sortf="-k 2,2 -k 1,1";
 201         } else if((TopsitesSort & TOPSITE_SORT_TIME) != 0) {
 202                 sortf="-k 3,3";
 203         } else if((TopsitesSort & TOPSITE_SORT_USER) != 0) {
 204                 sortf="-k 4,4 -k 1,1 -k 2,2";
 205         } else {
 206                 sortf="-k 2,2 -k 1,1"; //default is BYTES
 207         }
 208         if((TopsitesSort & TOPSITE_SORT_REVERSE) != 0) {
 209                 sortt="-r";
 210         } else {
 211                 sortt="";
 212         }
 213
 214         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3)>=sizeof(csort)) {
 215                 debuga(__FILE__,__LINE__,_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general3,sites);
 216                 exit(EXIT_FAILURE);
 217         }
 218         cstatus=system(csort);
 219         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 220                 debuga(__FILE__,__LINE__,_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 221                 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
 222                 exit(EXIT_FAILURE);
 223         }
 224         if((fp_in=fopen(sites,"r"))==NULL) {
 225                 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),sites,strerror(errno));
 226                 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
 227                 exit(EXIT_FAILURE);
 228         }
 229
 230         if (!KeepTempLog && unlink(general3)) {
 231                 debuga(__FILE__,__LINE__,_("Cannot delete \"%s\": %s\n"),general3,strerror(errno));
 232                 exit(EXIT_FAILURE);
 233         }
 234
 235         if((fp_ou=fopen(report,"w"))==NULL) {
 236                 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),report,strerror(errno));
 237                 exit(EXIT_FAILURE);
 238         }
 239
 240         write_html_header(fp_ou,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Top sites"),HTML_JS_SORTTABLE);
 241         fputs("<tr><td class=\"header_c\">",fp_ou);
 242         fprintf(fp_ou,_("Period: %s"),period.html);
 243         fputs("</td></tr>\n",fp_ou);
 244         fputs("<tr><th class=\"header_c\">",fp_ou);
 245         fprintf(fp_ou,_("Top %d sites"),TopSitesNum);
 246         fputs("</th></tr>\n",fp_ou);
 247         close_html_header(fp_ou);
 248
 249         fputs("<div class=\"report\"><table cellpadding=\"1\" cellspacing=\"2\"",fp_ou);
 250         if (SortTableJs[0]) fputs(" class=\"sortable\"",fp_ou);
 251         fputs(">\n",fp_ou);
 252         fprintf(fp_ou,"<thead><tr><th class=\"header_l\">%s</th><th class=\"header_l",
 253         /* TRANSLATORS: This is a column header showing the position of the entry in the sorted list. */
 254         _("NUM"));
 255         if (SortTableJs[0]) fputs(" sorttable_alpha",fp_ou);
 256         fprintf(fp_ou,"\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th></tr></thead>\n",
 257         /* TRANSLATORS: This is a column header showing the URL of the visited sites. */
 258         _("ACCESSED SITE"),
 259         /* TRANSLATORS: This is a column header showing the number of connections to a visited site. */
 260         _("CONNECT"),
 261         /* TRANSLATORS: This is a column header showing the number of transfered bytes. */
 262         _("BYTES"),
 263         /* TRANSLATORS: This is a column header showing the time spent by the proxy processing the requests. */
 264         pgettext("duration","TIME"),
 265         /* TRANSLATORS: This is a column header showing the number of users who visited a sites. */
 266         _("USERS"));
 267
 268         regs=0;
 269         ntopsites = 0;
 270
 271         if ((line=longline_create())==NULL) {
 272                 debuga(__FILE__,__LINE__,_("Not enough memory to read file \"%s\"\n"),sites);
 273                 exit(EXIT_FAILURE);
 274         }
 275
 276         while(regs<TopSitesNum && (buf=longline_read(fp_in,line))!=NULL) {
 277                 getword_start(&gwarea,buf);
 278                 if (getword_atoll(&nacc,&gwarea,'\t')<0) {
 279                         debuga(__FILE__,__LINE__,_("Invalid record in file \"%s\"\n"),sites);
 280                         exit(EXIT_FAILURE);
 281                 }
 282                 if (nacc == 0) continue;
 283                 if (getword_atoll(&nbytes,&gwarea,'\t')<0 || getword_atoll(&ntime,&gwarea,'\t')<0) {
 284                         debuga(__FILE__,__LINE__,_("Invalid record in file \"%s\"\n"),sites);
 285                         exit(EXIT_FAILURE);
 286                 }
 287                 if (getword_atoi(&nusers,&gwarea,'\t')<0) {
 288                         debuga(__FILE__,__LINE__,_("Invalid number of users in file \"%s\"\n"),sites);
 289                         exit(EXIT_FAILURE);
 290                 }
 291                 if (getword_ptr(buf,&url,&gwarea,'\t')<0) {
 292                         debuga(__FILE__,__LINE__,_("Invalid url in file \"%s\"\n"),sites);
 293                         exit(EXIT_FAILURE);
 294                 }
 295
 296                 twork1=nacc;
 297                 twork2=nbytes;
 298                 twork3=ntime;
 299
 300                 fprintf(fp_ou,"<tr><td class=\"data\">%d</td><td class=\"data2\">",++regs);
 301
 302                 if(BlockIt[0] != '\0' && url[0]!=ALIAS_PREFIX) {
 303                         fprintf(fp_ou,"<a href=\"%s%s?url=\"",wwwDocumentRoot,BlockIt);
 304                         output_html_url(fp_ou,url);
 305                         fputs("\"><img src=\"../images/sarg-squidguard-block.png\"></a>&nbsp;",fp_ou);
 306                 }
 307
 308                 output_html_link(fp_ou,url,100);
 309                 fputs("</td><td class=\"data\"",fp_ou);
 310                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork1);
 311                 fprintf(fp_ou,">%s</td>",fixnum(twork1,1));
 312                 fputs("<td class=\"data\"",fp_ou);
 313                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork2);
 314                 fprintf(fp_ou,">%s</td>",fixnum(twork2,1));
 315                 fputs("<td class=\"data\"",fp_ou);
 316                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork3);
 317                 fprintf(fp_ou,">%s</td>",fixtime(twork3));
 318                 fputs("<td class=\"data\"",fp_ou);
 319                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%d\"",nusers);
 320                 fprintf(fp_ou,">%s</td></tr>\n",fixnum(nusers,1));
 321         }
 322         if (fclose(fp_in)==EOF) {
 323                 debuga(__FILE__,__LINE__,_("Read error in \"%s\": %s\n"),sites,strerror(errno));
 324                 exit(EXIT_FAILURE);
 325         }
 326         longline_destroy(&line);
 327
 328         fputs("</table></div>\n",fp_ou);
 329         write_html_trailer(fp_ou);
 330         if (fclose(fp_ou)==EOF) {
 331                 debuga(__FILE__,__LINE__,_("Write error in \"%s\": %s\n"),report,strerror(errno));
 332                 exit(EXIT_FAILURE);
 333         }
 334
 335         return;
 336 }