topsites.c

   1 /*
   2  * SARG Squid Analysis Report Generator      http://sarg.sourceforge.net
   3  *                                                            1998, 2011
   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[255];
  42         char general[MAXLEN];
  43         char general2[MAXLEN];
  44         char general3[MAXLEN];
  45         char sites[MAXLEN];
  46         char report[MAXLEN];
  47         const char *sortf;
  48         const char *sortt;
  49         long long int nacc;
  50         long long int nbytes;
  51         long long int ntime;
  52         long long int tnacc=0;
  53         long long int tnbytes=0;
  54         long long int tntime=0;
  55         long long int twork1=0, twork2=0, twork3=0;
  56 #ifdef ENABLE_DOUBLE_CHECK_DATA
  57         long long int ttnacc=0;
  58         long long int ttnbytes=0;
  59         long long int ttntime=0;
  60 #endif
  61         int regs=0;
  62         int cstatus;
  63         int url_len;
  64         int ourl_size=0;
  65         struct getwordstruct gwarea;
  66         longline line;
  67         struct generalitemstruct item;
  68
  69         if(Privacy) {
  70                 if (debugz) debugaz(_("Top sites report not produced because privacy option is on\n"));
  71                 return;
  72         }
  73
  74         sprintf(general,"%s/sarg-general",outdirname);
  75         sprintf(sites,"%s/sarg-sites",outdirname);
  76         sprintf(general2,"%s/sarg-general2",outdirname);
  77         sprintf(general3,"%s/sarg-general3",outdirname);
  78
  79         if ((ReportType & REPORT_TYPE_TOPUSERS) == 0)
  80                 sprintf(report,"%s/index.html",outdirname);
  81         else
  82                 sprintf(report,"%s/topsites.html",outdirname);
  83
  84         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" -k 4,4 -o \"%s\" \"%s\"",general2,general)>=sizeof(csort)) {
  85                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general,general2);
  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                 debuga(_("(topsites) Cannot open log file %s\n"),general2);
  97                 debuga(_("sort command: %s\n"),csort);
  98                 exit(EXIT_FAILURE);
  99         }
 100
 101         if((fp_ou=fopen(general3,"w"))==NULL) {
 102                 debuga(_("(topsites) Cannot open log file %s\n"),general3);
 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%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,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                         tnacc=0;
 146                         tnbytes=0;
 147                         tntime=0;
 148                 }
 149
 150                 tnacc+=item.nacc;
 151                 tnbytes+=item.nbytes;
 152                 tntime+=item.nelap;
 153 #ifdef ENABLE_DOUBLE_CHECK_DATA
 154                 ttnacc+=item.nacc;
 155                 ttnbytes+=item.nbytes;
 156                 ttntime+=item.nelap;
 157 #endif
 158         }
 159         fclose(fp_in);
 160         longline_destroy(&line);
 161
 162         if (ourl) {
 163                 /*
 164                 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 165                 to print a long long int unless it is exactly 64-bits long.
 166                 */
 167                 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,ourl);
 168                 free(ourl);
 169         }
 170
 171         fclose(fp_ou);
 172
 173 #ifdef ENABLE_DOUBLE_CHECK_DATA
 174         if (ttnacc!=globstat.nacc || ttnbytes!=globstat.nbytes || ttntime!=globstat.elap) {
 175                 debuga(_("Total statistics mismatch when reading %s to produce the top sites\n"),general2);
 176                 exit(EXIT_FAILURE);
 177         }
 178 #endif
 179
 180         if (unlink(general2)) {
 181                 debuga(_("Cannot delete %s - %s\n"),general2,strerror(errno));
 182                 exit(EXIT_FAILURE);
 183         }
 184
 185         if((TopsitesSort & TOPSITE_SORT_CONNECT) != 0) {
 186                 sortf="-k 1,1 -k 2,2";
 187         } else if((TopsitesSort & TOPSITE_SORT_BYTES) != 0) {
 188                 sortf="-k 2,2 -k 1,1";
 189         } else if((TopsitesSort & TOPSITE_SORT_TIME) != 0) {
 190                 sortf="-k 3,3";
 191         } else {
 192                 sortf="-k 2,2 -k 1,1"; //default is BYTES
 193         }
 194         if((TopsitesSort & TOPSITE_SORT_REVERSE) != 0) {
 195                 sortt="-r";
 196         } else {
 197                 sortt="";
 198         }
 199
 200         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3)>=sizeof(csort)) {
 201                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general3,sites);
 202                 exit(EXIT_FAILURE);
 203         }
 204         cstatus=system(csort);
 205         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 206                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 207                 debuga(_("sort command: %s\n"),csort);
 208                 exit(EXIT_FAILURE);
 209         }
 210         if((fp_in=fopen(sites,"r"))==NULL) {
 211                 debuga(_("(topsites) Cannot open log file %s\n"),sites);
 212                 debuga(_("sort command: %s\n"),csort);
 213                 exit(EXIT_FAILURE);
 214         }
 215
 216         if (unlink(general3)) {
 217                 debuga(_("Cannot delete %s - %s\n"),general3,strerror(errno));
 218                 exit(EXIT_FAILURE);
 219         }
 220
 221         if((fp_ou=fopen(report,"w"))==NULL) {
 222                 debuga(_("(topsites) Cannot open log file %s\n"),report);
 223                 exit(EXIT_FAILURE);
 224         }
 225
 226         write_html_header(fp_ou,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Top sites"),HTML_JS_SORTTABLE);
 227         fputs("<tr><td class=\"header_c\">",fp_ou);
 228         fprintf(fp_ou,_("Period: %s"),period.html);
 229         fputs("</td></tr>\n",fp_ou);
 230         fputs("<tr><th class=\"header_c\">",fp_ou);
 231         fprintf(fp_ou,_("Top %d sites"),TopSitesNum);
 232         fputs("</th></tr>\n",fp_ou);
 233         close_html_header(fp_ou);
 234
 235         fputs("<div class=\"report\"><table cellpadding=\"1\" cellspacing=\"2\"",fp_ou);
 236         if (SortTableJs[0]) fputs(" class=\"sortable\"",fp_ou);
 237         fputs(">\n",fp_ou);
 238         fprintf(fp_ou,"<thead><tr><th class=\"header_l\">%s</th><th class=\"header_l",_("NUM"));
 239         if (SortTableJs[0]) fputs(" sorttable_alpha",fp_ou);
 240         fprintf(fp_ou,"\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th></tr></thead>\n",_("ACCESSED SITE"),_("CONNECT"),_("BYTES"),_("TIME"));
 241
 242         regs=0;
 243         ntopsites = 0;
 244
 245         if ((line=longline_create())==NULL) {
 246                 debuga(_("Not enough memory to read file %s\n"),sites);
 247                 exit(EXIT_FAILURE);
 248         }
 249
 250         while(regs<TopSitesNum && (buf=longline_read(fp_in,line))!=NULL) {
 251                 getword_start(&gwarea,buf);
 252                 if (getword_atoll(&nacc,&gwarea,'\t')<0) {
 253                         debuga(_("Maybe you have a broken record or garbage in your %s file\n"),sites);
 254                         exit(EXIT_FAILURE);
 255                 }
 256                 if (nacc == 0) continue;
 257                 if (getword_atoll(&nbytes,&gwarea,'\t')<0 || getword_atoll(&ntime,&gwarea,'\t')<0) {
 258                         debuga(_("Maybe you have a broken record or garbage in your %s file\n"),sites);
 259                         exit(EXIT_FAILURE);
 260                 }
 261                 if (getword_ptr(buf,&url,&gwarea,'\t')<0) {
 262                         debuga(_("The url is invalid in file %s\n"),sites);
 263                         exit(EXIT_FAILURE);
 264                 }
 265
 266                 twork1=nacc;
 267                 twork2=nbytes;
 268                 twork3=ntime;
 269
 270                 fprintf(fp_ou,"<tr><td class=\"data\">%d</td><td class=\"data2\">",++regs);
 271
 272                 if(BlockIt[0] != '\0' && url[0]!=ALIAS_PREFIX) {
 273                         fprintf(fp_ou,"<a href=\"%s%s?url=\"",wwwDocumentRoot,BlockIt);
 274                         output_html_url(fp_ou,url);
 275                         fputs("\"><img src=\"../images/sarg-squidguard-block.png\"></a>&nbsp;",fp_ou);
 276                 }
 277
 278                 output_html_link(fp_ou,url,100);
 279                 fputs("</td><td class=\"data\"",fp_ou);
 280                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork1);
 281                 fprintf(fp_ou,">%s</td>",fixnum(twork1,1));
 282                 fputs("<td class=\"data\"",fp_ou);
 283                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork2);
 284                 fprintf(fp_ou,">%s</td>",fixnum(twork2,1));
 285                 fputs("<td class=\"data\"",fp_ou);
 286                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork3);
 287                 fprintf(fp_ou,">%s</td></tr>\n",fixtime(twork3));
 288         }
 289         fclose(fp_in);
 290         longline_destroy(&line);
 291
 292         fputs("</table></div>\n",fp_ou);
 293         if (write_html_trailer(fp_ou)<0)
 294                 debuga(_("Write error in file %s\n"),report);
 295         if (fclose(fp_ou)==EOF)
 296                 debuga(_("Failed to close file %s - %s\n"),report,strerror(errno));
 297
 298         return;
 299 }