W³_Live User Manual

$ wget -c http://www.cs.rpi.edu/~morria/w3live/code/w3live-gui-0.9.tar.gz;
$ tar xzfv w3live-gui-0.9.tar.gz;
$ cd w3live-gui-0.9;
$ ./configure;
$ make;

Installation required the GNU regex library as well as W3C's libwww. On systems such as the RPI Computer Science Departments Solaris network run configure with the following options

./configure --with-rx=/cs/puninj/WWWPAL/NEW/ALIB/RX/rx-1.5/rx/ \
	--with-www_lib=/cs/puninj/WWWPAL/NEW/ALIB/LIBWWW/cvs2/libwww/include/w3c-libwww

Instantiation

The W3Live GUI has no command line options. To run, simply execute the w3live binary. After compiling the w3live program will be in w3live-gui-0.9/src/.

Usage

Upon running w3live, you will be given a connection dialog. If a W3Live daemon is already running, you may enter its host and port and click "Connect".

Figure 1: W3Live Connection Dialog

If no W3Live daemon is currently active, you may start a new one by clicking on "Start New". You will be given a dialog asking for daemon parameters.

Figure 2: W3Live - New Daemon Dialog

Click launch to start the daemon. You will be given a window which displays log data as it is read.

Once a W3Live daemon is available to connect to, enter its host and port and click "Connect". You will be given a display listing all active users on the given web server. For each active user listed, you may click on the "Graph" tab and select the user from the drop down menu. This will display that user's path through your site. In the "Graph" tab you may also click on "Webgraph" to see a visualization of the paths of all users as they have traversed your site.

Figure 3: W3Live - Main Screen

Figure 4: W3Live - User Path Visualization

W3Live - Command Line

Download

The current version W3Live (console program) is 0.9. The package may be downloaded from

• w3live-0.9.tar.gz

Installation

Quick Installation

$ wget -c http://www.cs.rpi.edu/~morria/w3live/code/w3live-0.9.tar.gz;
$ tar xzfv w3live-0.9.tar.gz;
$ cd w3live-0.9
$ ./configure
$ make

Installation required the GNU regex library as well as W3C's libwww. On systems such as the RPI Computer Science Departments Solaris network run configure with the following options

./configure --with-rx=/cs/puninj/WWWPAL/NEW/ALIB/RX/rx-1.5/rx/ \
	--with-www_lib=/cs/puninj/WWWPAL/NEW/ALIB/LIBWWW/cvs2/libwww/include/w3c-libwww

Instantiation

W3Live comes with two programs. w3ld is the W3Live Daemon which collects and analyzes log data. After a successful compilation it will be available in w3live-0.9/w3ld/. To interact with the daemon the program w3lc is written. After compilation it will be available in w3live-0.9/w3lc/.

The W3Live daemon has the following command line options

Usage: w3ld [--version] [--help] [--debug]
        --base=<BASE> --domain=<DOMAIN> [--port=<PORT>]
        [--prefix=<PREFIX>] [--graph=<XGMML>] [--log=<FILE>]
 
  -v,--version           Output version info and quit
  -h,--help              Output this help info and quit
  -p,--port              Port to launch server on (default 1030)
  -b,--base              Base URL of web server being watched
  -f,--prefix            Prefix of site URL's (defaults to '/')
  -d,--domain            Domain of site (example: .cs.rpi.edu)
  -g,--graph             XGMML graph of site

For example, you may run w3ld as

$ w3ld --base=http://www.cs.rpi.edu --domain=.cs.rpi.edu

w3lc, The W3Live client is used for connecting to w3ld and obtaining information. It is instantiated as follows:

Usage: w3lc [--version] [--help] [--host=<HOST>] [--port=<PORT>]
                                                                                                      
  --help,-h     Print this help information
  --version,-v  Print w3lc's version
  --host,-d     Specify a w3live host to contact (default: localhost)
  --port,-p     Specify which port to contact w3live on (default: 1030)

The host defaults to localhost and the port defaults to 1030. If the daemon is running on another host, you might run w3lc as follows:

$ w3lc --host=wwwpal.cs.rpi.edu --port=1035

w3ld - The W3Live Daemon

w3ld read "extended log" data from stdin.

The extended log format is described by the w3c as W3C Working Draft WD-logfile-960323

In apache, the extended log format is as follows

LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" extended

You 'normal' way of feeding log data into w3ld is by reading the end of the log file as it is written.

tail -f /usr/local/apache/logs/extended_log | ./w3ld --base=http://www.example.com \
	--domain=.example.com

A slightly more elegant approach with Apache is feeding log data into w3ld bypassing the log file. In the Apache configuration file (commonly httpd.conf) you may enter the following

LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" extended
CustomLog "|/usr/local/bin/w3ld --base=http://www.example.com --domain=.example.com --port=1030" extended

Once w3ld is running, it accepts connections from w3lc (the W3Live client) on the default port of 1030, or another port specified with --port=<Port>

w3lc - The W3Live client

Once the W3Live daemon is running (described in the previous section) you may connect to the daemon using w3lc - the W3Live client.

The commands listed in the Command Reference allow you to interact with the log data as it is read. An example session is given below:

$ w3lc
(w3live) list_users()
alb-24-194-34-103.nycap.rr.com
cpe-66-190-198-193.roa.nc.charter.com
 
 (w3live) user_start_time(cpe-66-190-198-193.roa.nc.charter.com)
Thu Oct 30 18:09:06 2003
(w3live) user_entry_page(cpe-66-190-198-193.roa.nc.charter.com)
http://www.organix.tv/images/photo/Summer-2002/k2/64x48/t_k2_6.jpg
(w3live) user_path_length(cpe-66-190-198-193.roa.nc.charter.com)
8(w3live) shutdown()
$

W3Live Command Reference

• webgraph()

  Current site graph

• version()

  Current w3live version

• user_webgraph()

  Graph of current users

• user_start_time(User Name)

  Time at which user first accessed the site

• user_referer(User Name)

  Referer for a given user

• user_path_length(User Name)

  Length of the given user's path through the site

• user_path(User Name)

  A users path info with tab delimited URL, MIME, http code

• user_entry_page(User Name)

  rst document requested from a given user

• user_end_time(User Name)

  Time at which user last accessed the site

• user_current_doc(User Name)

  The last document accessed by a given user

• ser_avg_time_doc(User Name)

  Average time spent per document by a given user

• user_accesses(User Name)

  The number of requests from a given user

• total_accesses()

  The total number of accesses to the site

• shutdown()

  Shut down the w3live server

• num_active_users()

  Number of currently active users

• list_users()

  Get a list of current users (those active on the site)

• list_ua()

  List of User Agents

• list_spiders()

  Get a list of spiders who have accessed your site

• list_sites()

  Get a list of all active website users

• list_platform()

  List of Computer Platforms

• list_int_ref()

  List of internal referers

• list_graph_algos()

  List all available graph drawing algorithms

• list_ext_ref()

  List of external referers

• list_entry()

  List of site entry pages

• list_domains()

  Get a list of all active domains

• list_browser()

  List of Browsers

• links_webgraph()

  Graph of links from log

• help(Name of a function)

  Get help information on a given command

• graph_size(width, height)

  Set Graph Size

• graph_scale(integer between -7 and 20 inclusive)

  Set Graph Scale

• graph_algo(algorithm name)

  Run a graph drawing algorithm

• disconnect()

  Disconnect from the server

• base_webhost()

  Get the base webhost setting

• avg_path_length()

  Average length of all active user's paths

Authors

W3Live was written by Andrew Morrison with much help from Dr. Mukkai Krishnamoorthy and Dr. John Punin.

Graphpack was written by Dr. Mukkai Krishnamoorthy. Webloglib was written by Dr. John Punin.