useragent.c

   1 /*
   2  * AUTHOR: Pedro Lineu Orso                         pedro.orso@gmail.com
   3  *                                                            1998, 2009
   4  * SARG Squid Analysis Report Generator      http://sarg.sourceforge.net
   5  *
   6  * SARG donations:
   7  *      please look at http://sarg.sourceforge.net/donations.php
   8  * ---------------------------------------------------------------------
   9  *
  10  *  This program is free software; you can redistribute it and/or modify
  11  *  it under the terms of the GNU General Public License as published by
  12  *  the Free Software Foundation; either version 2 of the License, or
  13  *  (at your option) any later version.
  14  *
  15  *  This program is distributed in the hope that it will be useful,
  16  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  *  GNU General Public License for more details.
  19  *
  20  *  You should have received a copy of the GNU General Public License
  21  *  along with this program; if not, write to the Free Software
  22  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  23  *
  24  */
  25
  26 #include "include/conf.h"
  27 #include "include/defs.h"
  28
  29 void useragent(void)
  30 {
  31
  32    FILE *fp_in = NULL, *fp_ou = NULL, *fp_ht = NULL;
  33    char ip[MAXLEN], data[MAXLEN], agent[MAXLEN], user[MAXLEN];
  34    char ipbefore[MAXLEN], namebefore[MAXLEN];
  35    char tagent[MAXLEN];
  36    char user_old[MAXLEN]="$#%0a3bc6";
  37    char agent_old[MAXLEN]="$#%0a3bc6";
  38    char hfile[MAXLEN];
  39    char idate[MAXLEN], fdate[MAXLEN];
  40    int  agentot=0, agentot2=0, agentdif=0, cont=0, nagent;
  41    unsigned long totregsl=0;
  42    int cstatus;
  43    char ftime[128];
  44    struct getwordstruct gwarea;
  45
  46    ip[0]='\0';
  47    data[0]='\0';
  48    agent[0]='\0';
  49    user[0]='\0';
  50    user_old[0]='\0';
  51    agent_old[0]='\0';
  52    ipbefore[0]='\0';
  53    namebefore[0]='\0';
  54
  55    sprintf(tmp3,"%s/squagent.unsort",TempDir);
  56    sprintf(tmp2,"%s/squagent.log",TempDir);
  57
  58    if((fp_in=fopen(UserAgentLog,"r"))==NULL) {
  59       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],UserAgentLog);
  60       exit(1);
  61    }
  62
  63    if((fp_ou=fopen(tmp3,"w"))==NULL) {
  64       fprintf(stderr, "SARG: (email) %s: %s\n",text[45],tmp3);
  65       exit(1);
  66    }
  67
  68    if(debug) {
  69       debuga("%s: %s",text[66],UserAgentLog);
  70    }
  71
  72    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
  73       totregsl++;
  74       getword_start(&gwarea,buf);
  75       if (getword(ip,sizeof(ip),&gwarea,' ')<0 || getword_skip(MAXLEN,&gwarea,'[')<0 ||
  76           getword(data,sizeof(data),&gwarea,' ')<0) {
  77          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  78          exit(1);
  79       }
  80       if(totregsl == 1)
  81          strcpy(idate,data);
  82       if (getword_skip(MAXLEN,&gwarea,'"')<0 || getword(agent,sizeof(agent),&gwarea,'"')<0) {
  83          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  84          exit(1);
  85       }
  86       strcpy(warea,agent);
  87       strup(warea);
  88       if(strstr(warea,"SCRIPT") != 0 || strstr(warea,"ONLOAD") != 0)
  89          baddata();
  90
  91       if(strlen(gwarea.current)) {
  92          if (getword_skip(MAXLEN,&gwarea,' ')<0 || getword(user,sizeof(user),&gwarea,'\n')<0) {
  93             printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  94             exit(1);
  95          }
  96       }
  97
  98       if(user[0] == '-')
  99          strcpy(user,ip);
 100       if(strlen(user) == 0)
 101          strcpy(user,ip);
 102
 103       fprintf(fp_ou,"%s\t%s\t%s\n",ip,agent,user);
 104       user[0]='\0';
 105    }
 106
 107    strcpy(fdate,data);
 108
 109    if(debug) {
 110       debuga("   %s: %ld",text[10],totregsl);
 111    }
 112
 113    fclose(fp_in);
 114    fclose(fp_ou);
 115
 116    if(debug) {
 117       debuga("%s: %s",text[54],tmp2);
 118    }
 119
 120    sprintf(csort,"sort -n -t \"\t\" -k 3,3 -k 2,2 -k 1,1 -o \"%s\" \"%s\"",tmp2,tmp3);
 121    cstatus=system(csort);
 122    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 123       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 124       fprintf(stderr, "SARG: sort command: %s\n",csort);
 125       exit(1);
 126    }
 127    if((fp_in=fopen(tmp2,"r"))==NULL) {
 128       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp2);
 129       fprintf(stderr, "SARG: sort command: %s\n",csort);
 130       exit(1);
 131    }
 132
 133    unlink(tmp3);
 134
 135    sprintf(hfile,"%s/%s/useragent.html", outdir,period);
 136    if((fp_ht=fopen(hfile,"w"))==NULL) {
 137       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],hfile);
 138       exit(1);
 139    }
 140
 141    if(debug)
 142       debuga("%s",text[72]);
 143
 144    fprintf(fp_ht, "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01 Transitional//EN\" \"http://www.w3.org/TR/html4/loose.dtd\">\n<html>\n<head>\n  <meta http-equiv=\"Content-Type\" content=\"text/html; charset=%s\">\n",CharSet);
 145    css(fp_ht);
 146    fputs("</head>\n",fp_ht);
 147    //if(FontFace[0] != '\0') fprintf(fp_ht,"<font face=\"%s\">\n",FontFace);
 148    fprintf(fp_ht,"<body bgcolor=\"%s\" text=\"%s\" background=\"%s\">\n",BgColor,TxColor,BgImage);
 149    write_logo_image(fp_ht);
 150
 151    if(strcmp(IndexTree,"date") == 0)
 152       show_sarg(fp_ht, "../../..");
 153    else
 154       show_sarg(fp_ht,"..");
 155    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 156    fprintf(fp_ht,"<tr><th align=\"center\" colspan=\"2\"><b><font color=\"%s\" size=\"+1\">%s</font></b></th></tr>\n",TiColor,text[105]);
 157
 158    fputs("<tr><td></td></tr><tr><td></td></tr>",fp_ht);
 159    fputs("</table></div>\n",fp_ht);
 160
 161    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 162
 163    fprintf(fp_ht,"<tr><td align=\"right\" bgcolor=\"%s\">%s:<td align=\"left\" bgcolor=\"%s\">%s - %s</td></td></tr>\n",HeaderBgColor,text[89],TxBgColor,idate,fdate);
 164
 165    fputs("</table></div>\n",fp_ht);
 166    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 167    fputs("<tr><td></td><td></td></tr>",fp_ht);
 168
 169    fprintf(fp_ou,"<tr><th align=\"left\" bgcolor=\"%s\"><font color=\"%s\">%s</font></th><th bgcolor=\"%s\" align=\"left\"><font color=\"%s\">%s</font></th></tr>\n",HeaderBgColor,HeaderColor,text[98],HeaderBgColor,HeaderColor,text[106]);
 170
 171    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 172       getword_start(&gwarea,buf);
 173       if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
 174          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp2);
 175          exit(1);
 176       }
 177
 178       if(strcmp(Ip2Name,"yes") == 0) {
 179          if(strcmp(ip,ipbefore) != 0) {
 180             strcpy(ipbefore,ip);
 181             ip2name(ip,sizeof(ip));
 182             strcpy(namebefore,ip);
 183          } else strcpy(ip,namebefore);
 184       }
 185
 186       if (getword(agent,sizeof(agent),&gwarea,'\t')<0 ||
 187           getword(user,sizeof(user),&gwarea,'\t')<0) {
 188          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp2);
 189          exit(1);
 190       }
 191
 192       if(strcmp(user,user_old) != 0) {
 193          fprintf(fp_ht,"<tr><td align=\"left\" bgcolor=\"%s\">%s</td><td align=\"left\" bgcolor=\"%s\">%s</td></tr>\n",TxBgColor,user,TxBgColor,agent);
 194          strcpy(user_old,user);
 195          strcpy(agent_old,agent);
 196       } else {
 197          if(strcmp(agent,agent_old) != 0) {
 198             fprintf(fp_ht,"<tr><td></td><td align=\"left\" bgcolor=\"%s\">%s</td></tr>\n",TxBgColor,agent);
 199             strcpy(agent_old,agent);
 200          }
 201       }
 202    }
 203
 204    fputs("</table>\n",fp_ht);
 205    fclose(fp_in);
 206    fclose(fp_ou);
 207
 208    sprintf(csort,"sort -t \"\t\" -k 2,2 -o \"%s\" \"%s\"",tmp3,tmp2);
 209    cstatus=system(csort);
 210    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 211       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 212       fprintf(stderr, "SARG: sort command: %s\n",csort);
 213       exit(1);
 214    }
 215    if((fp_in=fopen(tmp3,"r"))==NULL) {
 216       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp3);
 217       fprintf(stderr, "SARG: sort command: %s\n",csort);
 218       exit(1);
 219    }
 220
 221    unlink(tmp2);
 222
 223    if((fp_ou=fopen(tmp2,"w"))==NULL) {
 224       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp2);
 225       exit(1);
 226    }
 227
 228    agent_old[0]='\0';
 229    cont=0;
 230
 231    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 232       getword_start(&gwarea,buf);
 233       if (getword(ip,sizeof(ip),&gwarea,'\t')<0 ||
 234           getword(agent,sizeof(agent),&gwarea,'\t')<0) {
 235          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp);
 236          exit(1);
 237       }
 238
 239       if(!cont) {
 240          cont++;
 241          strcpy(agent_old,agent);
 242       }
 243
 244       if(strcmp(agent,agent_old) != 0) {
 245          agentdif++;
 246          fprintf(fp_ou,"%06d %s\n",agentot,agent_old);
 247          strcpy(agent_old,agent);
 248          agentot2+=agentot;
 249          agentot=0;
 250       }
 251       agentot++;
 252    }
 253    agentdif++;
 254    fprintf(fp_ou,"%06d %s\n",agentot,agent);
 255    agentot2+=agentot;
 256
 257    fclose(fp_in);
 258    fclose(fp_ou);
 259
 260    unlink(tmp3);
 261
 262    sprintf(csort,"sort -n -r -k 1,1 -o \"%s\" \"%s\"",tmp3,tmp2);
 263    cstatus=system(csort);
 264    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 265       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 266       fprintf(stderr, "SARG: sort command: %s\n",csort);
 267       exit(1);
 268    }
 269    if((fp_in=fopen(tmp3,"r"))==NULL) {
 270       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp3);
 271       fprintf(stderr, "SARG: sort command: %s\n",csort);
 272       exit(1);
 273    }
 274
 275    unlink(tmp2);
 276
 277    if((fp_ht=fopen(hfile,"a"))==NULL) {
 278       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],hfile);
 279       exit(1);
 280    }
 281
 282    fputs("<br><br>\n",fp_ht);
 283
 284    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 285    fprintf(fp_ht,"<tr><th align=\"left\" bgcolor=\"%s\"><font color=\"%s\">%s</font></th><th bgcolor=\"%s\" align=\"left\"><font color=\"%s\">%s</font></th><th align=\"center\" bgcolor=\"%s\"><font color=\"%s\">%%</font></th></tr>\n",HeaderBgColor,HeaderColor,text[106],HeaderBgColor,HeaderColor,text[107],HeaderBgColor,HeaderColor);
 286
 287    perc=0.;
 288    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 289       fixendofline(buf);
 290       getword_start(&gwarea,buf);
 291       if (getword(tagent,sizeof(tagent),&gwarea,' ')<0) {
 292          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp);
 293          exit(1);
 294       }
 295       nagent=atoi(tagent);
 296       if(agentot2>0) {
 297          perc=nagent * 100;
 298          perc=perc / agentot2;
 299       }
 300
 301       fprintf(fp_ht,"<tr><td align=\"left\" bgcolor=\"%s\">%s</td><td align=\"right\" bgcolor=\"%s\">%d</td><td align=\"right\" bgcolor=\"%s\">%3.2f</td></tr>\n",TxBgColor,gwarea.current,TxBgColor,nagent,TxBgColor,perc);
 302    }
 303
 304    fputs("</table></html>\n",fp_ht);
 305    if(strcmp(ShowSargInfo,"yes") == 0) {
 306       zdate(ftime, sizeof(ftime), DateFormat);
 307       fprintf(fp_ht,"<br><br><div align=\"center\"><font size=\"-2\">%s <a href=\"%s\">%s-%s</a> %s %s</font></div>\n",text[108],URL,PGM,VERSION,text[109],ftime);
 308    }
 309
 310    fputs("</table>\n</body>\n</html>\n",fp_ht);
 311    fclose(fp_in);
 312    fclose(fp_ht);
 313
 314    unlink(tmp3);
 315
 316    return;
 317
 318 }