topsites.c

   1 /*
   2  * SARG Squid Analysis Report Generator      http://sarg.sourceforge.net
   3  *                                                            1998, 2013
   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) debugaz(_("Top sites report not produced because privacy option is on\n"));
  73                 return;
  74         }
  75
  76         sprintf(general,"%s/sarg-general",outdirname);
  77         sprintf(sites,"%s/sarg-sites",outdirname);
  78         sprintf(general2,"%s/sarg-general2",outdirname);
  79         sprintf(general3,"%s/sarg-general3",outdirname);
  80
  81         sprintf(report,"%s/topsites.html",outdirname);
  82
  83         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" -k 4,4 -k 1,1 -o \"%s\" \"%s\"",general2,general)>=sizeof(csort)) {
  84                 debuga(_("Command too long: "));
  85                 debuga_more("sort -t \"\t\" -k 4,4 -k 1,1 -o \"%s\" \"%s\"",general2,general);
  86                 exit(EXIT_FAILURE);
  87         }
  88         cstatus=system(csort);
  89         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
  90                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
  91                 debuga(_("sort command: %s\n"),csort);
  92                 exit(EXIT_FAILURE);
  93         }
  94
  95         if((fp_in=fopen(general2,"r"))==NULL) {
  96                 debugapos("topsites",_("Cannot open file \"%s\": %s\n"),general2,strerror(errno));
  97                 debuga(_("sort command: %s\n"),csort);
  98                 exit(EXIT_FAILURE);
  99         }
 100
 101         if((fp_ou=fopen(general3,"w"))==NULL) {
 102                 debugapos("topsites",_("Cannot open file \"%s\": %s\n"),general3,strerror(errno));
 103                 exit(EXIT_FAILURE);
 104         }
 105
 106         if ((line=longline_create())==NULL) {
 107                 debuga(_("Not enough memory to read file \"%s\"\n"),general2);
 108                 exit(EXIT_FAILURE);
 109         }
 110
 111         while((buf=longline_read(fp_in,line))!=NULL) {
 112                 ger_read(buf,&item,general2);
 113                 if(item.total) continue;
 114
 115                 if(!regs) {
 116                         url_len=strlen(item.url);
 117                         if (!ourl || url_len>=ourl_size) {
 118                                 ourl_size=url_len+1;
 119                                 ourl=realloc(ourl,ourl_size);
 120                                 if (!ourl) {
 121                                         debuga(_("Not enough memory to store the url\n"));
 122                                         exit(EXIT_FAILURE);
 123                                 }
 124                         }
 125                         strcpy(ourl,item.url);
 126                         regs++;
 127                 }
 128
 129                 if(strcmp(item.url,ourl) != 0) {
 130                         /*
 131                         This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 132                         to print a long long int unless it is exactly 64-bits long.
 133                         */
 134                         fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 135                         url_len=strlen(item.url);
 136                         if (url_len>=ourl_size) {
 137                                 ourl_size=url_len+1;
 138                                 ourl=realloc(ourl,ourl_size);
 139                                 if (!ourl) {
 140                                         debuga(_("Not enough memory to store the url\n"));
 141                                         exit(EXIT_FAILURE);
 142                                 }
 143                         }
 144                         strcpy(ourl,item.url);
 145                         strcpy(ouser,item.user);
 146                         tnacc=0;
 147                         tnbytes=0;
 148                         tntime=0;
 149                         nusers=1;
 150                 } else if (strcmp(item.user,ouser)!=0) {
 151                         strcpy(ouser,item.user);
 152                         nusers++;
 153                 }
 154
 155                 tnacc+=item.nacc;
 156                 tnbytes+=item.nbytes;
 157                 tntime+=item.nelap;
 158 #ifdef ENABLE_DOUBLE_CHECK_DATA
 159                 ttnacc+=item.nacc;
 160                 ttnbytes+=item.nbytes;
 161                 ttntime+=item.nelap;
 162 #endif
 163         }
 164         fclose(fp_in);
 165         longline_destroy(&line);
 166
 167         if (ourl) {
 168                 /*
 169                 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 170                 to print a long long int unless it is exactly 64-bits long.
 171                 */
 172                 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 173                 free(ourl);
 174         }
 175
 176         fclose(fp_ou);
 177
 178 #ifdef ENABLE_DOUBLE_CHECK_DATA
 179         if (ttnacc!=globstat.nacc || ttnbytes!=globstat.nbytes || ttntime!=globstat.elap) {
 180                 debuga(_("Total statistics mismatch when reading %s to produce the top sites\n"),general2);
 181                 exit(EXIT_FAILURE);
 182         }
 183 #endif
 184
 185         if (!KeepTempLog && unlink(general2)) {
 186                 debuga(_("Cannot delete \"%s\": %s\n"),general2,strerror(errno));
 187                 exit(EXIT_FAILURE);
 188         }
 189
 190         if((TopsitesSort & TOPSITE_SORT_CONNECT) != 0) {
 191                 sortf="-k 1,1 -k 2,2";
 192         } else if((TopsitesSort & TOPSITE_SORT_BYTES) != 0) {
 193                 sortf="-k 2,2 -k 1,1";
 194         } else if((TopsitesSort & TOPSITE_SORT_TIME) != 0) {
 195                 sortf="-k 3,3";
 196         } else if((TopsitesSort & TOPSITE_SORT_USER) != 0) {
 197                 sortf="-k 4,4 -k 1,1 -k 2,2";
 198         } else {
 199                 sortf="-k 2,2 -k 1,1"; //default is BYTES
 200         }
 201         if((TopsitesSort & TOPSITE_SORT_REVERSE) != 0) {
 202                 sortt="-r";
 203         } else {
 204                 sortt="";
 205         }
 206
 207         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3)>=sizeof(csort)) {
 208                 debuga(_("Command too long: "));
 209                 debuga_more("sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3);
 210                 exit(EXIT_FAILURE);
 211         }
 212         cstatus=system(csort);
 213         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 214                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 215                 debuga(_("sort command: %s\n"),csort);
 216                 exit(EXIT_FAILURE);
 217         }
 218         if((fp_in=fopen(sites,"r"))==NULL) {
 219                 debugapos("topsites",_("Cannot open file \"%s\": %s\n"),sites,strerror(errno));
 220                 debuga(_("sort command: %s\n"),csort);
 221                 exit(EXIT_FAILURE);
 222         }
 223
 224         if (!KeepTempLog && unlink(general3)) {
 225                 debuga(_("Cannot delete \"%s\": %s\n"),general3,strerror(errno));
 226                 exit(EXIT_FAILURE);
 227         }
 228
 229         if((fp_ou=fopen(report,"w"))==NULL) {
 230                 debugapos("topsites",_("Cannot open file \"%s\": %s\n"),report,strerror(errno));
 231                 exit(EXIT_FAILURE);
 232         }
 233
 234         write_html_header(fp_ou,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Top sites"),HTML_JS_SORTTABLE);
 235         fputs("<tr><td class=\"header_c\">",fp_ou);
 236         fprintf(fp_ou,_("Period: %s"),period.html);
 237         fputs("</td></tr>\n",fp_ou);
 238         fputs("<tr><th class=\"header_c\">",fp_ou);
 239         fprintf(fp_ou,_("Top %d sites"),TopSitesNum);
 240         fputs("</th></tr>\n",fp_ou);
 241         close_html_header(fp_ou);
 242
 243         fputs("<div class=\"report\"><table cellpadding=\"1\" cellspacing=\"2\"",fp_ou);
 244         if (SortTableJs[0]) fputs(" class=\"sortable\"",fp_ou);
 245         fputs(">\n",fp_ou);
 246         fprintf(fp_ou,"<thead><tr><th class=\"header_l\">%s</th><th class=\"header_l",
 247         /* TRANSLATORS: This is a column header showing the position of the entry in the sorted list. */
 248         _("NUM"));
 249         if (SortTableJs[0]) fputs(" sorttable_alpha",fp_ou);
 250         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",
 251         /* TRANSLATORS: This is a column header showing the URL of the visited sites. */
 252         _("ACCESSED SITE"),
 253         /* TRANSLATORS: This is a column header showing the number of connections to a visited site. */
 254         _("CONNECT"),
 255         /* TRANSLATORS: This is a column header showing the number of transfered bytes. */
 256         _("BYTES"),
 257         /* TRANSLATORS: This is a column header showing the time spent by the proxy processing the requests. */
 258         _("TIME"),
 259         /* TRANSLATORS: This is a column header showing the number of users who visited a sites. */
 260         _("USERS"));
 261
 262         regs=0;
 263         ntopsites = 0;
 264
 265         if ((line=longline_create())==NULL) {
 266                 debuga(_("Not enough memory to read file \"%s\"\n"),sites);
 267                 exit(EXIT_FAILURE);
 268         }
 269
 270         while(regs<TopSitesNum && (buf=longline_read(fp_in,line))!=NULL) {
 271                 getword_start(&gwarea,buf);
 272                 if (getword_atoll(&nacc,&gwarea,'\t')<0) {
 273                         debuga(_("Invalid number of accesses in file \"%s\"\n"),sites);
 274                         exit(EXIT_FAILURE);
 275                 }
 276                 if (nacc == 0) continue;
 277                 if (getword_atoll(&nbytes,&gwarea,'\t')<0) {
 278                         debuga(_("Invalid number of bytes in file \"%s\"\n"),sites);
 279                         exit(EXIT_FAILURE);
 280                 }
 281                 if (getword_atoll(&ntime,&gwarea,'\t')<0) {
 282                         debuga(_("Invalid elapsed time in file \"%s\"\n"),sites);
 283                         exit(EXIT_FAILURE);
 284                 }
 285                 if (getword_atoi(&nusers,&gwarea,'\t')<0) {
 286                         debuga(_("Invalid number of users in file \"%s\"\n"),sites);
 287                         exit(EXIT_FAILURE);
 288                 }
 289                 if (getword_ptr(buf,&url,&gwarea,'\t')<0) {
 290                         debuga(_("Invalid url in file \"%s\"\n"),sites);
 291                         exit(EXIT_FAILURE);
 292                 }
 293
 294                 twork1=nacc;
 295                 twork2=nbytes;
 296                 twork3=ntime;
 297
 298                 fprintf(fp_ou,"<tr><td class=\"data\">%d</td><td class=\"data2\">",++regs);
 299
 300                 if(BlockIt[0] != '\0' && url[0]!=ALIAS_PREFIX) {
 301                         fprintf(fp_ou,"<a href=\"%s%s?url=\"",wwwDocumentRoot,BlockIt);
 302                         output_html_url(fp_ou,url);
 303                         fputs("\"><img src=\"../images/sarg-squidguard-block.png\"></a>&nbsp;",fp_ou);
 304                 }
 305
 306                 output_html_link(fp_ou,url,100);
 307                 fputs("</td><td class=\"data\"",fp_ou);
 308                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork1);
 309                 fprintf(fp_ou,">%s</td>",fixnum(twork1,1));
 310                 fputs("<td class=\"data\"",fp_ou);
 311                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork2);
 312                 fprintf(fp_ou,">%s</td>",fixnum(twork2,1));
 313                 fputs("<td class=\"data\"",fp_ou);
 314                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork3);
 315                 fprintf(fp_ou,">%s</td>",fixtime(twork3));
 316                 fputs("<td class=\"data\"",fp_ou);
 317                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%d\"",nusers);
 318                 fprintf(fp_ou,">%s</td></tr>\n",fixnum(nusers,1));
 319         }
 320         fclose(fp_in);
 321         longline_destroy(&line);
 322
 323         fputs("</table></div>\n",fp_ou);
 324         if (write_html_trailer(fp_ou)<0)
 325                 debuga(_("Write error in file \"%s\"\n"),report);
 326         if (fclose(fp_ou)==EOF)
 327                 debuga(_("Failed to close file \"%s\": %s\n"),report,strerror(errno));
 328
 329         return;
 330 }