topsites.c

   1 /*
   2  * SARG Squid Analysis Report Generator      http://sarg.sourceforge.net
   3  *                                                            1998, 2012
   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         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(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general,general2);
  85                 exit(EXIT_FAILURE);
  86         }
  87         cstatus=system(csort);
  88         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
  89                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
  90                 debuga(_("sort command: %s\n"),csort);
  91                 exit(EXIT_FAILURE);
  92         }
  93
  94         if((fp_in=fopen(general2,"r"))==NULL) {
  95                 debuga(_("(topsites) Cannot open log file %s\n"),general2);
  96                 debuga(_("sort command: %s\n"),csort);
  97                 exit(EXIT_FAILURE);
  98         }
  99
 100         if((fp_ou=fopen(general3,"w"))==NULL) {
 101                 debuga(_("(topsites) Cannot open log file %s\n"),general3);
 102                 exit(EXIT_FAILURE);
 103         }
 104
 105         if ((line=longline_create())==NULL) {
 106                 debuga(_("Not enough memory to read file %s\n"),general2);
 107                 exit(EXIT_FAILURE);
 108         }
 109
 110         while((buf=longline_read(fp_in,line))!=NULL) {
 111                 ger_read(buf,&item,general2);
 112                 if(item.total) continue;
 113
 114                 if(!regs) {
 115                         url_len=strlen(item.url);
 116                         if (!ourl || url_len>=ourl_size) {
 117                                 ourl_size=url_len+1;
 118                                 ourl=realloc(ourl,ourl_size);
 119                                 if (!ourl) {
 120                                         debuga(_("Not enough memory to store the url\n"));
 121                                         exit(EXIT_FAILURE);
 122                                 }
 123                         }
 124                         strcpy(ourl,item.url);
 125                         regs++;
 126                 }
 127
 128                 if(strcmp(item.url,ourl) != 0) {
 129                         /*
 130                         This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 131                         to print a long long int unless it is exactly 64-bits long.
 132                         */
 133                         fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 134                         url_len=strlen(item.url);
 135                         if (url_len>=ourl_size) {
 136                                 ourl_size=url_len+1;
 137                                 ourl=realloc(ourl,ourl_size);
 138                                 if (!ourl) {
 139                                         debuga(_("Not enough memory to store the url\n"));
 140                                         exit(EXIT_FAILURE);
 141                                 }
 142                         }
 143                         strcpy(ourl,item.url);
 144                         strcpy(ouser,item.user);
 145                         tnacc=0;
 146                         tnbytes=0;
 147                         tntime=0;
 148                         nusers=1;
 149                 } else if (strcmp(item.user,ouser)!=0) {
 150                         strcpy(ouser,item.user);
 151                         nusers++;
 152                 }
 153
 154                 tnacc+=item.nacc;
 155                 tnbytes+=item.nbytes;
 156                 tntime+=item.nelap;
 157 #ifdef ENABLE_DOUBLE_CHECK_DATA
 158                 ttnacc+=item.nacc;
 159                 ttnbytes+=item.nbytes;
 160                 ttntime+=item.nelap;
 161 #endif
 162         }
 163         fclose(fp_in);
 164         longline_destroy(&line);
 165
 166         if (ourl) {
 167                 /*
 168                 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
 169                 to print a long long int unless it is exactly 64-bits long.
 170                 */
 171                 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
 172                 free(ourl);
 173         }
 174
 175         fclose(fp_ou);
 176
 177 #ifdef ENABLE_DOUBLE_CHECK_DATA
 178         if (ttnacc!=globstat.nacc || ttnbytes!=globstat.nbytes || ttntime!=globstat.elap) {
 179                 debuga(_("Total statistics mismatch when reading %s to produce the top sites\n"),general2);
 180                 exit(EXIT_FAILURE);
 181         }
 182 #endif
 183
 184         if (!KeepTempLog && unlink(general2)) {
 185                 debuga(_("Cannot delete \"%s\": %s\n"),general2,strerror(errno));
 186                 exit(EXIT_FAILURE);
 187         }
 188
 189         if((TopsitesSort & TOPSITE_SORT_CONNECT) != 0) {
 190                 sortf="-k 1,1 -k 2,2";
 191         } else if((TopsitesSort & TOPSITE_SORT_BYTES) != 0) {
 192                 sortf="-k 2,2 -k 1,1";
 193         } else if((TopsitesSort & TOPSITE_SORT_TIME) != 0) {
 194                 sortf="-k 3,3";
 195         } else if((TopsitesSort & TOPSITE_SORT_USER) != 0) {
 196                 sortf="-k 4,4 -k 1,1 -k 2,2";
 197         } else {
 198                 sortf="-k 2,2 -k 1,1"; //default is BYTES
 199         }
 200         if((TopsitesSort & TOPSITE_SORT_REVERSE) != 0) {
 201                 sortt="-r";
 202         } else {
 203                 sortt="";
 204         }
 205
 206         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3)>=sizeof(csort)) {
 207                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general3,sites);
 208                 exit(EXIT_FAILURE);
 209         }
 210         cstatus=system(csort);
 211         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 212                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 213                 debuga(_("sort command: %s\n"),csort);
 214                 exit(EXIT_FAILURE);
 215         }
 216         if((fp_in=fopen(sites,"r"))==NULL) {
 217                 debuga(_("(topsites) Cannot open log file %s\n"),sites);
 218                 debuga(_("sort command: %s\n"),csort);
 219                 exit(EXIT_FAILURE);
 220         }
 221
 222         if (!KeepTempLog && unlink(general3)) {
 223                 debuga(_("Cannot delete \"%s\": %s\n"),general3,strerror(errno));
 224                 exit(EXIT_FAILURE);
 225         }
 226
 227         if((fp_ou=fopen(report,"w"))==NULL) {
 228                 debuga(_("(topsites) Cannot open log file %s\n"),report);
 229                 exit(EXIT_FAILURE);
 230         }
 231
 232         write_html_header(fp_ou,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Top sites"),HTML_JS_SORTTABLE);
 233         fputs("<tr><td class=\"header_c\">",fp_ou);
 234         fprintf(fp_ou,_("Period: %s"),period.html);
 235         fputs("</td></tr>\n",fp_ou);
 236         fputs("<tr><th class=\"header_c\">",fp_ou);
 237         fprintf(fp_ou,_("Top %d sites"),TopSitesNum);
 238         fputs("</th></tr>\n",fp_ou);
 239         close_html_header(fp_ou);
 240
 241         fputs("<div class=\"report\"><table cellpadding=\"1\" cellspacing=\"2\"",fp_ou);
 242         if (SortTableJs[0]) fputs(" class=\"sortable\"",fp_ou);
 243         fputs(">\n",fp_ou);
 244         fprintf(fp_ou,"<thead><tr><th class=\"header_l\">%s</th><th class=\"header_l",
 245         /* TRANSLATORS: This is a column header showing the position of the entry in the sorted list. */
 246         _("NUM"));
 247         if (SortTableJs[0]) fputs(" sorttable_alpha",fp_ou);
 248         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",
 249         /* TRANSLATORS: This is a column header showing the URL of the visited sites. */
 250         _("ACCESSED SITE"),
 251         /* TRANSLATORS: This is a column header showing the number of connections to a visited site. */
 252         _("CONNECT"),
 253         /* TRANSLATORS: This is a column header showing the number of transfered bytes. */
 254         _("BYTES"),
 255         /* TRANSLATORS: This is a column header showing the time spent by the proxy processing the requests. */
 256         _("TIME"),
 257         /* TRANSLATORS: This is a column header showing the number of users who visited a sites. */
 258         _("USERS"));
 259
 260         regs=0;
 261         ntopsites = 0;
 262
 263         if ((line=longline_create())==NULL) {
 264                 debuga(_("Not enough memory to read file %s\n"),sites);
 265                 exit(EXIT_FAILURE);
 266         }
 267
 268         while(regs<TopSitesNum && (buf=longline_read(fp_in,line))!=NULL) {
 269                 getword_start(&gwarea,buf);
 270                 if (getword_atoll(&nacc,&gwarea,'\t')<0) {
 271                         debuga(_("Maybe you have a broken record or garbage in your %s file\n"),sites);
 272                         exit(EXIT_FAILURE);
 273                 }
 274                 if (nacc == 0) continue;
 275                 if (getword_atoll(&nbytes,&gwarea,'\t')<0 || getword_atoll(&ntime,&gwarea,'\t')<0) {
 276                         debuga(_("Maybe you have a broken record or garbage in your %s file\n"),sites);
 277                         exit(EXIT_FAILURE);
 278                 }
 279                 if (getword_atoi(&nusers,&gwarea,'\t')<0) {
 280                         debuga(_("The number of users is invalid in file %s\n"),sites);
 281                         exit(EXIT_FAILURE);
 282                 }
 283                 if (getword_ptr(buf,&url,&gwarea,'\t')<0) {
 284                         debuga(_("The url is invalid in file %s\n"),sites);
 285                         exit(EXIT_FAILURE);
 286                 }
 287
 288                 twork1=nacc;
 289                 twork2=nbytes;
 290                 twork3=ntime;
 291
 292                 fprintf(fp_ou,"<tr><td class=\"data\">%d</td><td class=\"data2\">",++regs);
 293
 294                 if(BlockIt[0] != '\0' && url[0]!=ALIAS_PREFIX) {
 295                         fprintf(fp_ou,"<a href=\"%s%s?url=\"",wwwDocumentRoot,BlockIt);
 296                         output_html_url(fp_ou,url);
 297                         fputs("\"><img src=\"../images/sarg-squidguard-block.png\"></a>&nbsp;",fp_ou);
 298                 }
 299
 300                 output_html_link(fp_ou,url,100);
 301                 fputs("</td><td class=\"data\"",fp_ou);
 302                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork1);
 303                 fprintf(fp_ou,">%s</td>",fixnum(twork1,1));
 304                 fputs("<td class=\"data\"",fp_ou);
 305                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork2);
 306                 fprintf(fp_ou,">%s</td>",fixnum(twork2,1));
 307                 fputs("<td class=\"data\"",fp_ou);
 308                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork3);
 309                 fprintf(fp_ou,">%s</td>",fixtime(twork3));
 310                 fputs("<td class=\"data\"",fp_ou);
 311                 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%d\"",nusers);
 312                 fprintf(fp_ou,">%s</td></tr>\n",fixnum(nusers,1));
 313         }
 314         fclose(fp_in);
 315         longline_destroy(&line);
 316
 317         fputs("</table></div>\n",fp_ou);
 318         if (write_html_trailer(fp_ou)<0)
 319                 debuga(_("Write error in file %s\n"),report);
 320         if (fclose(fp_ou)==EOF)
 321                 debuga(_("Failed to close file %s - %s\n"),report,strerror(errno));
 322
 323         return;
 324 }