useragent.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 void useragent(void)
  31 {
  32         FILE *fp_in = NULL, *fp_ou = NULL, *fp_ht = NULL;
  33         char buf[MAXLEN];
  34         char ip[MAXLEN], data[MAXLEN], agent[MAXLEN], user[MAXLEN];
  35         char ipbefore[MAXLEN], namebefore[MAXLEN];
  36         char tagent[MAXLEN];
  37         char user_old[MAXLEN]="$#%0a3bc6";
  38         char agent_old[MAXLEN]="$#%0a3bc6";
  39         char hfile[MAXLEN];
  40         char idate[MAXLEN], fdate[MAXLEN];
  41         char tmp2[MAXLEN];
  42         char tmp3[MAXLEN];
  43         char day[4],month[5],year[5], wdate[20];
  44         char csort[MAXLEN];
  45         int  agentot=0, agentot2=0, agentdif=0, cont=0, nagent;
  46         unsigned long totregsl=0;
  47         int cstatus;
  48         int ndate;
  49         double perc;
  50         struct getwordstruct gwarea, gwarea1;
  51
  52         ip[0]='\0';
  53         data[0]='\0';
  54         agent[0]='\0';
  55         user[0]='\0';
  56         user_old[0]='\0';
  57         agent_old[0]='\0';
  58         ipbefore[0]='\0';
  59         namebefore[0]='\0';
  60
  61         sprintf(tmp3,"%s/squagent.int_unsort",tmp);
  62         sprintf(tmp2,"%s/squagent.int_log",tmp);
  63
  64         if((fp_in=fopen(UserAgentLog,"r"))==NULL) {
  65                 debuga(_("(useragent) Cannot open file %s\n"),UserAgentLog);
  66                 exit(EXIT_FAILURE);
  67         }
  68
  69         if((fp_ou=fopen(tmp3,"w"))==NULL) {
  70                 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
  71                 exit(EXIT_FAILURE);
  72         }
  73
  74         if(debug) {
  75                 debuga(_("Reading useragent log: %s\n"),UserAgentLog);
  76         }
  77
  78         while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
  79                 totregsl++;
  80                 getword_start(&gwarea,buf);
  81                 if (getword(ip,sizeof(ip),&gwarea,' ')<0 || getword_skip(MAXLEN,&gwarea,'[')<0 ||
  82                     getword(data,sizeof(data),&gwarea,' ')<0) {
  83                         debuga(_("Maybe you have a broken record or garbage in your %s file\n"),UserAgentLog);
  84                         exit(EXIT_FAILURE);
  85                 }
  86                 getword_start(&gwarea1,data);
  87                 if (getword(day,sizeof(day),&gwarea1,'/')<0 || getword(month,sizeof(month),&gwarea1,'/')<0 ||
  88                     getword(year,sizeof(year),&gwarea1,':')<0) {
  89                         debuga(_("Maybe you have a broken date in your %s file\n"),UserAgentLog);
  90                         exit(EXIT_FAILURE);
  91                 }
  92                 buildymd(day,month,year,wdate);
  93                 ndate=atoi(wdate);
  94                 if (ndate<dfrom) continue;
  95                 if (ndate>duntil) break;
  96                 if(totregsl == 1)
  97                         strcpy(idate,data);
  98                 strcpy(fdate,data);
  99                 if (getword_skip(MAXLEN,&gwarea,'"')<0 || getword(agent,sizeof(agent),&gwarea,'"')<0) {
 100                         debuga(_("Maybe you have a broken useragent entry in your %s file\n"),UserAgentLog);
 101                         exit(EXIT_FAILURE);
 102                 }
 103
 104                 if(gwarea.current[0]!='\0') {
 105                         if (getword_skip(MAXLEN,&gwarea,' ')<0 || getword(user,sizeof(user),&gwarea,'\n')<0) {
 106                                 debuga(_("Maybe you have a broken record or garbage in your %s file\n"),UserAgentLog);
 107                                 exit(EXIT_FAILURE);
 108                         }
 109                         if(user[0] == '-')
 110                                 strcpy(user,ip);
 111                         if(user[0] == '\0')
 112                                 strcpy(user,ip);
 113                 } else {
 114                         strcpy(user,ip);
 115                 }
 116
 117                 fprintf(fp_ou,"%s\t%s\t%s\n",ip,agent,user);
 118                 useragent_count++;
 119         }
 120
 121         if(debug) {
 122                 debuga(_("   Records read: %ld\n"),totregsl);
 123         }
 124
 125         fclose(fp_in);
 126         if (fclose(fp_ou)==EOF) {
 127                 debuga(_("Failed to close file %s - %s\n"),tmp3,strerror(errno));
 128                 exit(EXIT_FAILURE);
 129         }
 130
 131         if(debug) {
 132                 debuga(_("Sorting file: %s\n"),tmp2);
 133         }
 134
 135         if (snprintf(csort,sizeof(csort),"sort -n -t \"\t\" -k 3,3 -k 2,2 -k 1,1 -o \"%s\" \"%s\"",tmp2,tmp3)>=sizeof(csort)) {
 136                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),tmp2,tmp3);
 137                 exit(EXIT_FAILURE);
 138         }
 139         cstatus=system(csort);
 140         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 141                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 142                 debuga(_("sort command: %s\n"),csort);
 143                 exit(EXIT_FAILURE);
 144         }
 145         if((fp_in=fopen(tmp2,"r"))==NULL) {
 146                 debuga(_("(useragent) Cannot open file %s\n"),tmp2);
 147                 debuga(_("sort command: %s\n"),csort);
 148                 exit(EXIT_FAILURE);
 149         }
 150
 151         if (!KeepTempLog && unlink(tmp3)) {
 152                 debuga(_("Cannot delete \"%s\": %s\n"),tmp3,strerror(errno));
 153                 exit(EXIT_FAILURE);
 154         }
 155
 156         snprintf(hfile,sizeof(hfile),"%s/useragent.html", outdirname);
 157         if((fp_ht=fopen(hfile,"w"))==NULL) {
 158                 debuga(_("(useragent) Cannot open file %s\n"),hfile);
 159                 exit(EXIT_FAILURE);
 160         }
 161
 162         if(debug)
 163                 debuga(_("Making Useragent report\n"));
 164
 165         write_html_header(fp_ht,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Squid Useragent's Report"),HTML_JS_NONE);
 166         fprintf(fp_ht,"<tr><th class=\"header_c\">%s</th></tr>\n",_("Squid Useragent's Report"));
 167         fprintf(fp_ht,"<tr><td class=\"header_c\">%s: %s - %s</td></tr>\n",_("Period"),idate,fdate);
 168         close_html_header(fp_ht);
 169
 170         fputs("<br><br>\n",fp_ht);
 171
 172         fputs("<div class=\"report\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 173         fputs("<tr><td>&nbsp;</td><td>&nbsp;</td></tr>",fp_ht);
 174
 175         fprintf(fp_ht,"<tr><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th></tr>\n",_("USERID"),_("AGENT"));
 176
 177         while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 178                 getword_start(&gwarea,buf);
 179                 if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
 180                         debuga(_("There is an invalid IP address in file %s\n"),tmp2);
 181                         exit(EXIT_FAILURE);
 182                 }
 183
 184                 if(Ip2Name) {
 185                         if(strcmp(ip,ipbefore) != 0) {
 186                                 strcpy(ipbefore,ip);
 187                                 ip2name(ip,sizeof(ip));
 188                                 strcpy(namebefore,ip);
 189                         } else strcpy(ip,namebefore);
 190                 }
 191
 192                 if (getword(agent,sizeof(agent),&gwarea,'\t')<0) {
 193                         debuga(_("There is an invalid useragent in file %s\n"),tmp2);
 194                         exit(EXIT_FAILURE);
 195                 }
 196                 if (getword(user,sizeof(user),&gwarea,'\t')<0) {
 197                         debuga(_("There is an invalid user ID in file %s\n"),tmp2);
 198                         exit(EXIT_FAILURE);
 199                 }
 200
 201                 if(strcmp(user,user_old) != 0) {
 202                         fprintf(fp_ht,"<tr><td class=\"data2\">%s</td><td class=\"data2\">",user);
 203                         output_html_string(fp_ht,agent,250);
 204                         fputs("</td></tr>\n",fp_ht);
 205                         strcpy(user_old,user);
 206                         strcpy(agent_old,agent);
 207                 } else if(strcmp(agent,agent_old) != 0) {
 208                         fputs("<tr><td></td><td class=\"data2\">",fp_ht);
 209                         output_html_string(fp_ht,agent,250);
 210                         fputs("</td></tr>\n",fp_ht);
 211                         strcpy(agent_old,agent);
 212                 }
 213         }
 214
 215         fputs("</table>\n",fp_ht);
 216         fclose(fp_in);
 217
 218         if (snprintf(csort,sizeof(csort),"sort -t \"\t\" -k 2,2 -o \"%s\" \"%s\"",tmp3,tmp2)>=sizeof(csort)) {
 219                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),tmp2,tmp3);
 220                 exit(EXIT_FAILURE);
 221         }
 222         cstatus=system(csort);
 223         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 224                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 225                 debuga(_("sort command: %s\n"),csort);
 226                 exit(EXIT_FAILURE);
 227         }
 228         if((fp_in=fopen(tmp3,"r"))==NULL) {
 229                 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
 230                 debuga(_("sort command: %s\n"),csort);
 231                 exit(EXIT_FAILURE);
 232         }
 233
 234         if (!KeepTempLog && unlink(tmp2)) {
 235                 debuga(_("Cannot delete \"%s\": %s\n"),tmp2,strerror(errno));
 236                 exit(EXIT_FAILURE);
 237         }
 238
 239         if((fp_ou=fopen(tmp2,"w"))==NULL) {
 240                 debuga(_("(useragent) Cannot open file %s\n"),tmp2);
 241                 exit(EXIT_FAILURE);
 242         }
 243
 244         agent_old[0]='\0';
 245         cont=0;
 246
 247         while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 248                 getword_start(&gwarea,buf);
 249                 if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
 250                         debuga(_("There is an invalid IP address in file %s\n"),tmp3);
 251                         exit(EXIT_FAILURE);
 252                 }
 253                 if (getword(agent,sizeof(agent),&gwarea,'\t')<0) {
 254                         debuga(_("There is an invalid useragent in file %s\n"),tmp3);
 255                         exit(EXIT_FAILURE);
 256                 }
 257
 258                 if(!cont) {
 259                         cont++;
 260                         strcpy(agent_old,agent);
 261                 }
 262
 263                 if(strcmp(agent,agent_old) != 0) {
 264                         agentdif++;
 265                         fprintf(fp_ou,"%06d %s\n",agentot,agent_old);
 266                         strcpy(agent_old,agent);
 267                         agentot2+=agentot;
 268                         agentot=0;
 269                 }
 270                 agentot++;
 271         }
 272         agentdif++;
 273         fprintf(fp_ou,"%06d %s\n",agentot,agent);
 274         agentot2+=agentot;
 275
 276         fclose(fp_in);
 277         if (fclose(fp_ou)==EOF) {
 278                 debuga(_("Failed to close file %s - %s\n"),tmp3,strerror(errno));
 279                 exit(EXIT_FAILURE);
 280         }
 281
 282         if (!KeepTempLog && unlink(tmp3)) {
 283                 debuga(_("Cannot delete \"%s\": %s\n"),tmp3,strerror(errno));
 284                 exit(EXIT_FAILURE);
 285         }
 286
 287         if (snprintf(csort,sizeof(csort),"sort -n -r -k 1,1 -o \"%s\" \"%s\"",tmp3,tmp2)>=sizeof(csort)) {
 288                 debuga(_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),tmp2,tmp3);
 289                 exit(EXIT_FAILURE);
 290         }
 291         cstatus=system(csort);
 292         if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 293                 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
 294                 debuga(_("sort command: %s\n"),csort);
 295                 exit(EXIT_FAILURE);
 296         }
 297         if((fp_in=fopen(tmp3,"r"))==NULL) {
 298                 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
 299                 debuga(_("sort command: %s\n"),csort);
 300                 exit(EXIT_FAILURE);
 301         }
 302
 303         if (!KeepTempLog && unlink(tmp2)) {
 304                 debuga(_("Cannot delete \"%s\": %s\n"),tmp2,strerror(errno));
 305                 exit(EXIT_FAILURE);
 306         }
 307
 308         fputs("<br><br>\n",fp_ht);
 309
 310         fputs("<table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 311         fprintf(fp_ht,"<tr><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_c\">%%</th></tr>\n",_("AGENT"),_("TOTAL"));
 312
 313         perc=0.;
 314         while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 315                 fixendofline(buf);
 316                 getword_start(&gwarea,buf);
 317                 if (getword(tagent,sizeof(tagent),&gwarea,' ')<0) {
 318                         debuga(_("There is an invalid useragent in file %s\n"),tmp3);
 319                         exit(EXIT_FAILURE);
 320                 }
 321                 nagent=atoi(tagent);
 322                 perc=(agentot2>0) ? nagent * 100. / agentot2 : 0.;
 323
 324                 fputs("<tr><td class=\"data2\">",fp_ht);
 325                 output_html_string(fp_ht,gwarea.current,250);
 326                 fprintf(fp_ht,"</td><td class=\"data\">%d</td><td class=\"data\">%3.2lf</td></tr>\n",nagent,perc);
 327         }
 328         fclose(fp_in);
 329
 330         fputs("</table></div>\n",fp_ht);
 331         if (write_html_trailer(fp_ht)<0)
 332                 debuga(_("Write error in file %s\n"),hfile);
 333         if (fclose(fp_ht)==EOF)
 334                 debuga(_("Failed to close file %s - %s\n"),hfile,strerror(errno));
 335
 336         if (!KeepTempLog && unlink(tmp3)) {
 337                 debuga(_("Cannot delete \"%s\": %s\n"),tmp3,strerror(errno));
 338                 exit(EXIT_FAILURE);
 339         }
 340
 341         return;
 342 }