readme

harvest是一个下载html网页得机器人

Since Eric directed my attention to the access statistics, I dug up the
seemingly orphaned project BrokerStats and prepared this release.

What I have done is to make this archive is complete because the original
archive, distributed as BrokerStats.tar.gz was not packaged correctly.

Besides this, no changes were made.

BrokerStats needs work to become useful and even more work to reach
release quality.

Most notable problems are:
- year 2000 problem
- it would be nice to get rid of cgi_lib.pl
- code needs cleanup

BrokerStats.cgi
v0.6
---------------

This is the first release of these file... so i don't know very well what should i write.
If you find something  confuse, please mail me ([e-mail]), and may be better explained in the next version.

This script is useful to get statiscs about brokers in Harvest. It works as cgi-bin, because i think that WWW is a easier interface than the command line for most users... but with little changes can be run from the command line.

I used part of a script from Eric Lease Morgan, as he says:

"I have no problem with you incorporating my code into your. Enjoy. Use it as
you will. My philosophy is, "If you don't want it copied, then don't put it on
the 'Net." Enjoy.

--
Eric Lease Morgan"

I also took some modifications from Javi Massa to use Eric Lease script in the group IRIS-INDEX.

Feel free to use this script and distribute it if you think that it's useful. As I took part of the script from other people, I can't tell anybody not to copy it, furthermore... I will be pleased if somebody use it!.

Anyway, I would like to hear from any modifications made, in order to add them to the script.



----------------------------------------------------

And now, some tips about the program:

- Installation

There is no much to say about installation... if you know what a cgi-bin is, and you know something about perl, you will find it very easy.
You just need to copy the program in the  cgi-bin directory of your HTTP server.. The HTML page can be anywhere in your server. You can even copy that page in your broker admin directory.


Like in other perl scripts, be sure that the first line of the script is your correct perl location.
It's also important to copy the cgi_lib.pl to your perl lib directory (should also work fine if you leave it in the same directory as the script)

The last important thing is that your broker.out file should be readable for the HTTP server.


- Support for various brokers:

The program can manage various brokers, that is what i needed at CIPRES, but if you plan to use it for just one broker, you can fix the broker name at the $broker variable at the beginning of the program. Another possibility is to fix the name as a hidden variable in the HTML page.

It's in the HTML page where you should give the brokers names, in order to get combined with variables $logprefix and $brokerprefix, obtaining the right values to broker.out and broker url.

- cgi_lib.pl

It's a perl library that the program uses. It should be with the rest of perl librays (in /usr/lib/perl5, in Linux RedHat) or in the same directory as the script file.
You can also copy the function ReadParse (and other functions used by ReadParse) into the source of BrokerStats.cgi.

- maxdetail

This variable is used to limit the number of detailed results in the list of querys sent to the broker. I think that querys made only one time are not relevant, but you can show them if you want.








Madrid, January de 1998

------------------------------------------------------------------
Hermann Straus Carrasco
[e-mail]  -  http://www.cipres.upm.es/~hstraus
[e-mail]