]> git.ipfire.org Git - thirdparty/sarg.git/blame - useragent.c
Tag the released 2.3 version.
[thirdparty/sarg.git] / useragent.c
CommitLineData
25697a35 1/*
94ff9470 2 * SARG Squid Analysis Report Generator http://sarg.sourceforge.net
1164c474 3 * 1998, 2010
25697a35
GS
4 *
5 * SARG donations:
6 * please look at http://sarg.sourceforge.net/donations.php
1164c474
FM
7 * Support:
8 * http://sourceforge.net/projects/sarg/forums/forum/363374
25697a35
GS
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"
5f3cfd1d 28#include "include/defs.h"
25697a35 29
32e71fa4 30void useragent(void)
25697a35
GS
31{
32
33 FILE *fp_in = NULL, *fp_ou = NULL, *fp_ht = NULL;
06b39c87 34 char buf[MAXLEN];
d6e703cc
FM
35 char ip[MAXLEN], data[MAXLEN], agent[MAXLEN], user[MAXLEN];
36 char ipbefore[MAXLEN], namebefore[MAXLEN];
37 char tagent[MAXLEN];
38 char user_old[MAXLEN]="$#%0a3bc6";
39 char agent_old[MAXLEN]="$#%0a3bc6";
25697a35
GS
40 char hfile[MAXLEN];
41 char idate[MAXLEN], fdate[MAXLEN];
06ced858 42 char tmp2[MAXLEN];
2240dcea 43 char tmp3[MAXLEN];
012ba254 44 char day[4],month[5],year[5], wdate[20];
2240dcea 45 char csort[MAXLEN];
dcb54d06 46 int agentot=0, agentot2=0, agentdif=0, cont=0, nagent;
25697a35 47 unsigned long totregsl=0;
456d78a5 48 int cstatus;
012ba254
FM
49 int ndate;
50 double perc;
51 struct getwordstruct gwarea, gwarea1;
25697a35
GS
52
53 ip[0]='\0';
54 data[0]='\0';
55 agent[0]='\0';
56 user[0]='\0';
57 user_old[0]='\0';
58 agent_old[0]='\0';
d6e703cc
FM
59 ipbefore[0]='\0';
60 namebefore[0]='\0';
25697a35 61
d6e703cc 62 sprintf(tmp3,"%s/squagent.unsort",TempDir);
25697a35
GS
63 sprintf(tmp2,"%s/squagent.log",TempDir);
64
65 if((fp_in=fopen(UserAgentLog,"r"))==NULL) {
9f70c14e 66 debuga(_("(useragent) Cannot open file %s\n"),UserAgentLog);
06b39c87 67 exit(EXIT_FAILURE);
25697a35
GS
68 }
69
d6e703cc 70 if((fp_ou=fopen(tmp3,"w"))==NULL) {
9f70c14e 71 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
06b39c87 72 exit(EXIT_FAILURE);
25697a35
GS
73 }
74
75 if(debug) {
10210234 76 debuga(_("Reading useragent log: %s\n"),UserAgentLog);
25697a35
GS
77 }
78
79 while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
80 totregsl++;
9c7c6346
FM
81 getword_start(&gwarea,buf);
82 if (getword(ip,sizeof(ip),&gwarea,' ')<0 || getword_skip(MAXLEN,&gwarea,'[')<0 ||
83 getword(data,sizeof(data),&gwarea,' ')<0) {
9f70c14e 84 debuga(_("Maybe you have a broken record or garbage in your %s file\n"),UserAgentLog);
06b39c87 85 exit(EXIT_FAILURE);
4bcb77cf 86 }
012ba254
FM
87 getword_start(&gwarea1,data);
88 if (getword(day,sizeof(day),&gwarea1,'/')<0 || getword(month,sizeof(month),&gwarea1,'/')<0 ||
89 getword(year,sizeof(year),&gwarea1,':')<0) {
9f70c14e 90 debuga(_("Maybe you have a broken date in your %s file\n"),UserAgentLog);
06b39c87 91 exit(EXIT_FAILURE);
012ba254
FM
92 }
93 buildymd(day,month,year,wdate);
94 ndate=atoi(wdate);
c0ec9cc7
FM
95 if (ndate<dfrom) continue;
96 if (ndate>duntil) break;
25697a35
GS
97 if(totregsl == 1)
98 strcpy(idate,data);
012ba254 99 strcpy(fdate,data);
9c7c6346 100 if (getword_skip(MAXLEN,&gwarea,'"')<0 || getword(agent,sizeof(agent),&gwarea,'"')<0) {
9f70c14e 101 debuga(_("Maybe you have a broken useragent entry in your %s file\n"),UserAgentLog);
06b39c87 102 exit(EXIT_FAILURE);
4bcb77cf 103 }
d6e703cc
FM
104 strcpy(warea,agent);
105 strup(warea);
106 if(strstr(warea,"SCRIPT") != 0 || strstr(warea,"ONLOAD") != 0)
107 baddata();
108
0a4d2bda 109 if(gwarea.current[0]!='\0') {
9c7c6346 110 if (getword_skip(MAXLEN,&gwarea,' ')<0 || getword(user,sizeof(user),&gwarea,'\n')<0) {
9f70c14e 111 debuga(_("Maybe you have a broken record or garbage in your %s file\n"),UserAgentLog);
06b39c87 112 exit(EXIT_FAILURE);
4bcb77cf 113 }
25697a35
GS
114 }
115
116 if(user[0] == '-')
117 strcpy(user,ip);
0a4d2bda 118 if(user[0] == '\0')
25697a35
GS
119 strcpy(user,ip);
120
dcb54d06 121 fprintf(fp_ou,"%s\t%s\t%s\n",ip,agent,user);
25697a35 122 user[0]='\0';
c0ec9cc7 123 useragent_count++;
25697a35
GS
124 }
125
25697a35 126 if(debug) {
10210234 127 debuga(_(" Records read: %ld\n"),totregsl);
25697a35
GS
128 }
129
130 fclose(fp_in);
131 fclose(fp_ou);
132
25697a35 133 if(debug) {
10210234 134 debuga(_("Sorting file: %s\n"),tmp2);
25697a35
GS
135 }
136
dcb54d06 137 sprintf(csort,"sort -n -t \"\t\" -k 3,3 -k 2,2 -k 1,1 -o \"%s\" \"%s\"",tmp2,tmp3);
456d78a5
FM
138 cstatus=system(csort);
139 if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
10210234
FM
140 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
141 debuga(_("sort command: %s\n"),csort);
06b39c87 142 exit(EXIT_FAILURE);
456d78a5 143 }
25697a35 144 if((fp_in=fopen(tmp2,"r"))==NULL) {
fcdc0918 145 debuga(_("(useragent) Cannot open file %s\n"),tmp2);
10210234 146 debuga(_("sort command: %s\n"),csort);
06b39c87 147 exit(EXIT_FAILURE);
25697a35
GS
148 }
149
456d78a5
FM
150 unlink(tmp3);
151
d5d021c5 152 snprintf(hfile,sizeof(hfile),"%s/useragent.html", outdirname);
25697a35 153 if((fp_ht=fopen(hfile,"w"))==NULL) {
9f70c14e 154 debuga(_("(useragent) Cannot open file %s\n"),hfile);
06b39c87 155 exit(EXIT_FAILURE);
25697a35 156 }
120d768c 157
25697a35 158 if(debug)
10210234 159 debuga(_("Making Useragent report\n"));
25697a35 160
c0ec9cc7 161 write_html_header(fp_ht,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Squid Useragent's Report"));
c36c7384
FM
162 fprintf(fp_ht,"<tr><th class=\"header_c\">%s</th></tr>\n",_("Squid Useragent's Report"));
163 fprintf(fp_ht,"<tr><td class=\"header_c\">%s: %s - %s</td></tr>\n",_("Period"),idate,fdate);
c0ec9cc7 164 close_html_header(fp_ht);
25697a35 165
c0ec9cc7 166 fputs("<br><br>\n",fp_ht);
25697a35 167
c0ec9cc7 168 fputs("<div class=\"report\"><table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
012ba254 169 fputs("<tr><td>&nbsp;</td><td>&nbsp;</td></tr>",fp_ht);
25697a35 170
c36c7384 171 fprintf(fp_ht,"<tr><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th></tr>\n",_("USERID"),_("AGENT"));
25697a35
GS
172
173 while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
9c7c6346
FM
174 getword_start(&gwarea,buf);
175 if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
10210234 176 debuga(_("There is an invalid IP address in file %s\n"),tmp2);
06b39c87 177 exit(EXIT_FAILURE);
4bcb77cf 178 }
25697a35 179
246c8489 180 if(Ip2Name) {
d6e703cc
FM
181 if(strcmp(ip,ipbefore) != 0) {
182 strcpy(ipbefore,ip);
a1c55d8c 183 ip2name(ip,sizeof(ip));
d6e703cc
FM
184 strcpy(namebefore,ip);
185 } else strcpy(ip,namebefore);
25697a35
GS
186 }
187
10210234
FM
188 if (getword(agent,sizeof(agent),&gwarea,'\t')<0) {
189 debuga(_("There is an invalid useragent in file %s\n"),tmp2);
06b39c87 190 exit(EXIT_FAILURE);
10210234
FM
191 }
192 if (getword(user,sizeof(user),&gwarea,'\t')<0) {
193 debuga(_("There is an invalid user ID in file %s\n"),tmp2);
06b39c87 194 exit(EXIT_FAILURE);
4bcb77cf 195 }
25697a35
GS
196
197 if(strcmp(user,user_old) != 0) {
012ba254 198 fprintf(fp_ht,"<tr><td class=\"data2\">%s</td><td class=\"data2\">%s</td></tr>\n",user,agent);
25697a35
GS
199 strcpy(user_old,user);
200 strcpy(agent_old,agent);
201 } else {
202 if(strcmp(agent,agent_old) != 0) {
012ba254 203 fprintf(fp_ht,"<tr><td></td><td class=\"data2\">%s</td></tr>\n",agent);
25697a35
GS
204 strcpy(agent_old,agent);
205 }
206 }
207 }
120d768c 208
25697a35
GS
209 fputs("</table>\n",fp_ht);
210 fclose(fp_in);
25697a35 211
dcb54d06 212 sprintf(csort,"sort -t \"\t\" -k 2,2 -o \"%s\" \"%s\"",tmp3,tmp2);
456d78a5
FM
213 cstatus=system(csort);
214 if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
10210234
FM
215 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
216 debuga(_("sort command: %s\n"),csort);
06b39c87 217 exit(EXIT_FAILURE);
456d78a5 218 }
d6e703cc 219 if((fp_in=fopen(tmp3,"r"))==NULL) {
9f70c14e 220 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
10210234 221 debuga(_("sort command: %s\n"),csort);
06b39c87 222 exit(EXIT_FAILURE);
25697a35
GS
223 }
224
456d78a5
FM
225 unlink(tmp2);
226
25697a35 227 if((fp_ou=fopen(tmp2,"w"))==NULL) {
fcdc0918 228 debuga(_("(useragent) Cannot open file %s\n"),tmp2);
06b39c87 229 exit(EXIT_FAILURE);
25697a35
GS
230 }
231
232 agent_old[0]='\0';
dcb54d06 233 cont=0;
25697a35
GS
234
235 while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
9c7c6346 236 getword_start(&gwarea,buf);
10210234
FM
237 if (getword(ip,sizeof(ip),&gwarea,'\t')<0) {
238 debuga(_("There is an invalid IP address in file %s\n"),tmp);
06b39c87 239 exit(EXIT_FAILURE);
10210234
FM
240 }
241 if (getword(agent,sizeof(agent),&gwarea,'\t')<0) {
242 debuga(_("There is an invalid useragent in file %s\n"),tmp);
06b39c87 243 exit(EXIT_FAILURE);
4bcb77cf 244 }
25697a35
GS
245
246 if(!cont) {
247 cont++;
248 strcpy(agent_old,agent);
249 }
120d768c 250
25697a35
GS
251 if(strcmp(agent,agent_old) != 0) {
252 agentdif++;
354c1a68 253 fprintf(fp_ou,"%06d %s\n",agentot,agent_old);
25697a35
GS
254 strcpy(agent_old,agent);
255 agentot2+=agentot;
dcb54d06 256 agentot=0;
25697a35 257 }
dcb54d06 258 agentot++;
25697a35
GS
259 }
260 agentdif++;
354c1a68 261 fprintf(fp_ou,"%06d %s\n",agentot,agent);
25697a35
GS
262 agentot2+=agentot;
263
264 fclose(fp_in);
265 fclose(fp_ou);
266
d6e703cc 267 unlink(tmp3);
25697a35 268
9a2efbd0 269 sprintf(csort,"sort -n -r -k 1,1 -o \"%s\" \"%s\"",tmp3,tmp2);
456d78a5
FM
270 cstatus=system(csort);
271 if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
10210234
FM
272 debuga(_("sort command return status %d\n"),WEXITSTATUS(cstatus));
273 debuga(_("sort command: %s\n"),csort);
06b39c87 274 exit(EXIT_FAILURE);
456d78a5 275 }
d6e703cc 276 if((fp_in=fopen(tmp3,"r"))==NULL) {
9f70c14e 277 debuga(_("(useragent) Cannot open file %s\n"),tmp3);
10210234 278 debuga(_("sort command: %s\n"),csort);
06b39c87 279 exit(EXIT_FAILURE);
25697a35
GS
280 }
281
456d78a5
FM
282 unlink(tmp2);
283
c0ec9cc7 284 fputs("<br><br>\n",fp_ht);
25697a35 285
c0ec9cc7 286 fputs("<table cellpadding=\"0\" cellspacing=\"0\">\n",fp_ht);
c36c7384 287 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"));
25697a35 288
dcb54d06 289 perc=0.;
25697a35 290 while(fgets(buf,sizeof(buf),fp_in)!=NULL) {
120d768c 291 fixendofline(buf);
9c7c6346
FM
292 getword_start(&gwarea,buf);
293 if (getword(tagent,sizeof(tagent),&gwarea,' ')<0) {
10210234 294 debuga(_("There is an invalid useragent in file %s\n"),tmp);
06b39c87 295 exit(EXIT_FAILURE);
4bcb77cf 296 }
dcb54d06 297 nagent=atoi(tagent);
012ba254 298 perc=(agentot2>0) ? nagent * 100. / agentot2 : 0.;
25697a35 299
012ba254 300 fprintf(fp_ht,"<tr><td class=\"data2\">%s</td><td class=\"data\">%d</td><td class=\"data\">%3.2lf</td></tr>\n",gwarea.current,nagent,perc);
25697a35 301 }
25697a35 302 fclose(fp_in);
c0ec9cc7
FM
303
304 fputs("</table></div>\n",fp_ht);
fa6552b0
FM
305 if (write_html_trailer(fp_ht)<0)
306 debuga(_("Write error in file %s\n"),hfile);
307 if (fclose(fp_ht)==EOF)
308 debuga(_("Failed to close file %s - %s\n"),hfile,strerror(errno));
25697a35 309
120d768c 310 unlink(tmp3);
25697a35
GS
311
312 return;
120d768c 313
25697a35 314}