useragent.c

   1 /*
   2  * AUTHOR: Pedro Lineu Orso                         pedro.orso@gmail.com
   3  *                                                            1998, 2010
   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    char tmp2[MAXLEN];
  41    int  agentot=0, agentot2=0, agentdif=0, cont=0, nagent;
  42    unsigned long totregsl=0;
  43    int cstatus;
  44    char ftime[128];
  45    struct getwordstruct gwarea;
  46
  47    ip[0]='\0';
  48    data[0]='\0';
  49    agent[0]='\0';
  50    user[0]='\0';
  51    user_old[0]='\0';
  52    agent_old[0]='\0';
  53    ipbefore[0]='\0';
  54    namebefore[0]='\0';
  55
  56    sprintf(tmp3,"%s/squagent.unsort",TempDir);
  57    sprintf(tmp2,"%s/squagent.log",TempDir);
  58
  59    if((fp_in=fopen(UserAgentLog,"r"))==NULL) {
  60       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],UserAgentLog);
  61       exit(1);
  62    }
  63
  64    if((fp_ou=fopen(tmp3,"w"))==NULL) {
  65       fprintf(stderr, "SARG: (email) %s: %s\n",text[45],tmp3);
  66       exit(1);
  67    }
  68
  69    if(debug) {
  70       debuga("%s: %s",text[66],UserAgentLog);
  71    }
  72
  73    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
  74       totregsl++;
  75       getword_start(&gwarea,buf);
  76       if (getword(ip,sizeof(ip),&gwarea,' ')<0 || getword_skip(MAXLEN,&gwarea,'[')<0 ||
  77           getword(data,sizeof(data),&gwarea,' ')<0) {
  78          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  79          exit(1);
  80       }
  81       if(totregsl == 1)
  82          strcpy(idate,data);
  83       if (getword_skip(MAXLEN,&gwarea,'"')<0 || getword(agent,sizeof(agent),&gwarea,'"')<0) {
  84          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  85          exit(1);
  86       }
  87       strcpy(warea,agent);
  88       strup(warea);
  89       if(strstr(warea,"SCRIPT") != 0 || strstr(warea,"ONLOAD") != 0)
  90          baddata();
  91
  92       if(gwarea.current[0]!='\0') {
  93          if (getword_skip(MAXLEN,&gwarea,' ')<0 || getword(user,sizeof(user),&gwarea,'\n')<0) {
  94             printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",UserAgentLog);
  95             exit(1);
  96          }
  97       }
  98
  99       if(user[0] == '-')
 100          strcpy(user,ip);
 101       if(user[0] == '\0')
 102          strcpy(user,ip);
 103
 104       fprintf(fp_ou,"%s\t%s\t%s\n",ip,agent,user);
 105       user[0]='\0';
 106    }
 107
 108    strcpy(fdate,data);
 109
 110    if(debug) {
 111       debuga("   %s: %ld",text[10],totregsl);
 112    }
 113
 114    fclose(fp_in);
 115    fclose(fp_ou);
 116
 117    if(debug) {
 118       debuga("%s: %s",text[54],tmp2);
 119    }
 120
 121    sprintf(csort,"sort -n -t \"\t\" -k 3,3 -k 2,2 -k 1,1 -o \"%s\" \"%s\"",tmp2,tmp3);
 122    cstatus=system(csort);
 123    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 124       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 125       fprintf(stderr, "SARG: sort command: %s\n",csort);
 126       exit(1);
 127    }
 128    if((fp_in=fopen(tmp2,"r"))==NULL) {
 129       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp2);
 130       fprintf(stderr, "SARG: sort command: %s\n",csort);
 131       exit(1);
 132    }
 133
 134    unlink(tmp3);
 135
 136    sprintf(hfile,"%s/%s/useragent.html", outdir,period);
 137    if((fp_ht=fopen(hfile,"w"))==NULL) {
 138       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],hfile);
 139       exit(1);
 140    }
 141
 142    if(debug)
 143       debuga("%s",text[72]);
 144
 145    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);
 146    css(fp_ht);
 147    fputs("</head>\n",fp_ht);
 148    //if(FontFace[0] != '\0') fprintf(fp_ht,"<font face=\"%s\">\n",FontFace);
 149    fprintf(fp_ht,"<body bgcolor=\"%s\" text=\"%s\" background=\"%s\">\n",BgColor,TxColor,BgImage);
 150    write_logo_image(fp_ht);
 151
 152    if(strcmp(IndexTree,"date") == 0)
 153       show_sarg(fp_ht, "../../..");
 154    else
 155       show_sarg(fp_ht,"..");
 156    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 157    fprintf(fp_ht,"<tr><th align=\"center\" colspan=\"2\"><b><font color=\"%s\" size=\"+1\">%s</font></b></th></tr>\n",TiColor,text[105]);
 158
 159    fputs("<tr><td></td></tr><tr><td></td></tr>",fp_ht);
 160    fputs("</table></div>\n",fp_ht);
 161
 162    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 163
 164    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);
 165
 166    fputs("</table></div>\n",fp_ht);
 167    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 168    fputs("<tr><td></td><td></td></tr>",fp_ht);
 169
 170    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]);
 171
 172    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 173       getword_start(&gwarea,buf);
 174       if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
 175          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp2);
 176          exit(1);
 177       }
 178
 179       if(Ip2Name) {
 180          if(strcmp(ip,ipbefore) != 0) {
 181             strcpy(ipbefore,ip);
 182             ip2name(ip,sizeof(ip));
 183             strcpy(namebefore,ip);
 184          } else strcpy(ip,namebefore);
 185       }
 186
 187       if (getword(agent,sizeof(agent),&gwarea,'\t')<0 ||
 188           getword(user,sizeof(user),&gwarea,'\t')<0) {
 189          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp2);
 190          exit(1);
 191       }
 192
 193       if(strcmp(user,user_old) != 0) {
 194          fprintf(fp_ht,"<tr><td align=\"left\" bgcolor=\"%s\">%s</td><td align=\"left\" bgcolor=\"%s\">%s</td></tr>\n",TxBgColor,user,TxBgColor,agent);
 195          strcpy(user_old,user);
 196          strcpy(agent_old,agent);
 197       } else {
 198          if(strcmp(agent,agent_old) != 0) {
 199             fprintf(fp_ht,"<tr><td></td><td align=\"left\" bgcolor=\"%s\">%s</td></tr>\n",TxBgColor,agent);
 200             strcpy(agent_old,agent);
 201          }
 202       }
 203    }
 204
 205    fputs("</table>\n",fp_ht);
 206    fclose(fp_in);
 207    fclose(fp_ou);
 208
 209    sprintf(csort,"sort -t \"\t\" -k 2,2 -o \"%s\" \"%s\"",tmp3,tmp2);
 210    cstatus=system(csort);
 211    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 212       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 213       fprintf(stderr, "SARG: sort command: %s\n",csort);
 214       exit(1);
 215    }
 216    if((fp_in=fopen(tmp3,"r"))==NULL) {
 217       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp3);
 218       fprintf(stderr, "SARG: sort command: %s\n",csort);
 219       exit(1);
 220    }
 221
 222    unlink(tmp2);
 223
 224    if((fp_ou=fopen(tmp2,"w"))==NULL) {
 225       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp2);
 226       exit(1);
 227    }
 228
 229    agent_old[0]='\0';
 230    cont=0;
 231
 232    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 233       getword_start(&gwarea,buf);
 234       if (getword(ip,sizeof(ip),&gwarea,'\t')<0 ||
 235           getword(agent,sizeof(agent),&gwarea,'\t')<0) {
 236          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp);
 237          exit(1);
 238       }
 239
 240       if(!cont) {
 241          cont++;
 242          strcpy(agent_old,agent);
 243       }
 244
 245       if(strcmp(agent,agent_old) != 0) {
 246          agentdif++;
 247          fprintf(fp_ou,"%06d %s\n",agentot,agent_old);
 248          strcpy(agent_old,agent);
 249          agentot2+=agentot;
 250          agentot=0;
 251       }
 252       agentot++;
 253    }
 254    agentdif++;
 255    fprintf(fp_ou,"%06d %s\n",agentot,agent);
 256    agentot2+=agentot;
 257
 258    fclose(fp_in);
 259    fclose(fp_ou);
 260
 261    unlink(tmp3);
 262
 263    sprintf(csort,"sort -n -r -k 1,1 -o \"%s\" \"%s\"",tmp3,tmp2);
 264    cstatus=system(csort);
 265    if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
 266       fprintf(stderr, "SARG: sort command return status %d\n",WEXITSTATUS(cstatus));
 267       fprintf(stderr, "SARG: sort command: %s\n",csort);
 268       exit(1);
 269    }
 270    if((fp_in=fopen(tmp3,"r"))==NULL) {
 271       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],tmp3);
 272       fprintf(stderr, "SARG: sort command: %s\n",csort);
 273       exit(1);
 274    }
 275
 276    unlink(tmp2);
 277
 278    if((fp_ht=fopen(hfile,"a"))==NULL) {
 279       fprintf(stderr, "SARG: (useragent) %s: %s\n",text[45],hfile);
 280       exit(1);
 281    }
 282
 283    fputs("<br><br>\n",fp_ht);
 284
 285    fputs("<div align=\"center\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
 286    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);
 287
 288    perc=0.;
 289    while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
 290       fixendofline(buf);
 291       getword_start(&gwarea,buf);
 292       if (getword(tagent,sizeof(tagent),&gwarea,' ')<0) {
 293          printf("SARG: Maybe you have a broken record or garbage in your %s file.\n",tmp);
 294          exit(1);
 295       }
 296       nagent=atoi(tagent);
 297       if(agentot2>0) {
 298          perc=nagent * 100;
 299          perc=perc / agentot2;
 300       }
 301
 302       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);
 303    }
 304
 305    fputs("</table></html>\n",fp_ht);
 306    if(ShowSargInfo) {
 307       zdate(ftime, sizeof(ftime), DateFormat);
 308       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);
 309    }
 310
 311    fputs("</table>\n</body>\n</html>\n",fp_ht);
 312    fclose(fp_in);
 313    fclose(fp_ht);
 314
 315    unlink(tmp3);
 316
 317    return;
 318
 319 }