Java-Download - Http Log Parser

bullet1 

Compiled Classes
(48 KB)
Updated on 3/14/03

bullet1 

Sourcecode and javadoc
(133 KB)

Purpose:

This log analyzer reads a given http log and counts how many requests come from a browser or search robot. It also allows to set filter (e.g. for counting only html page requests or images) or to count search requests from search engines.

Highlights compared to other log analyzers are:

  • summarizes browsers by generation, e.g. MSIE 5.5 or 6.x
  • analysis by which search engine and what search keywords your page was accessed (improved in version 1.1.0). You can also set filters.

 

bullet1 

Installation

bullet1 

Usage

bullet1 

Planned Extensions

bullet1 

For Developers

1. Installation

Include the file logparser.jar in your classpath

2. Usage

If you want to use the graphical interface:
   java -jar logparser.jar

If you prefer the text mode
  java de.hawlitzek.logparser.FHDetailHttpLogParser <file> [<filter>]
   <filter>: [ALL | MAIN | HTML | IMG | REF] (filterString)*
 

3. Planned Extensions

We would like to support new log file formats (feel free to support your own). The program should parse more than one file in a single run.
I'm using Micrsosoft Excel for the visualization of the results, so I do not plan to provide any charts. But it would be nice to have a copy/paste or file export support for the data transfer to Excel.

4. For Developers

This tool is freeware and Open Source. You can use or modify it as you want. We would be pleased if you send your improvements back to java@hawlitzek-consulting.de

The package de.hawlitzek.logparser contains the parsing classes. If you want to add other log formats (I'm just using the one from my provider) or additional statistics, you can extend the given classes.
Most important is the class LogEntry which represents one line in the http log. It parses the line into the fields for date, ip address, referrer etc. It also provides functionality to filter log entries by type, occurring names and search engines.
We provided two parser classes: The first one, FHBrowserAnalyzingHttpLogParser, is a very simple log analyzer which only counts how many request come from a browser or search robot. The second one, FHDetailHttpLogParser, provides more statistics and filters.
There is another package named de.hawlitzek.logparser.gui which contains a graphical user interface for FHDetailHttpLogParser. It is generated with IBM VisualAge for Java. Please do not manually change this, but write me an email (java@hawlitzek-consulting.de), because manual changes could be overwritten by the generator.
We provided the gui in two languages: English and German. If want to add additional languages, just translate the file logparsergui.properties to your language, rename it according to the ressource naming standard, e.g. logparsergui_it_IT.properties and add it to the jar file.
 

© 2003 Hawlitzek IT-Consulting, Marketing&Design Kirsten Literski-Hawlitzek, Mail an Webmaster

Seitenanfang