|
|
Using Scratchy
Scratchy currently consists of 2 executable python scripts, namely parse and report. Under
normal usage, there is no need to execute report because the
parse program automatically generates reports as necessary.
Typically, the report script can be executed to rebuild an
existing report, perhaps using a different config
file.
Prerequisites
In order to run the excutables within Scratchy you will need to obtain
version 2.1 or greater of Python,
Most Linux/Unix distributions include Python, however, depending on your
distro the version may not be as recent as 2.1. You can determine the
version of Python installed on your system by executing the following
command:
$ python -V
Python 2.1
$
Scratchy also requires a SQL database. Currently, MySQL and SQLite are
supported. Please see choosing the SQL database
for more information.
NOTE: SQLite is not currently supported although the latest
Scratchy release contains preliminary support it is considerer experiental
and not fully documented (and hasn't been tested, in other words, don't use it!).
If you wish for Scratchy to produce bar charts of the parsed data then you will also need the ChartDirector Python API
(and optionally, the ChartDirectorScratchy library). The fully-functional ChartDirector trial library adds a yellow banner
to each chart it creates. The fine folks at Advanced Software Engineering
have created a custom replacement library for the Scratchy project which replaces the yellow banner with a more pleasant
transparent image. This replacement library should be installed after ChartDirector is installed and is available for download.
Setup
First off, you will need an Apache server log file (typically, located in
/var/log/apache or /usr/local/apache/logs). Currently, only the combined
log format is supported (others may work, but I haven't tried them yet).
Scratchy ships with a file named config.dist you should copy this
file to config if you have only one web server that you are
monitoring. If you have multiple distinct servers (such as your default
site and then a dedicated search server or user server) you may wish to
parse these into separate reports. It's up to you. If you wish to parse
them all into one report then you will only need one config file. If
however you wish to have distinct reports then you will want to create a
config for each report. Under this scenario you may want a config_www,
config_search and a config_user.
The config file, at the very least contains central information for the
parser and reporter scripts. The config file may also contain other
preferences such as color schemes and data limits for the reporter. You
may want to explore the settings of the Config
file.
Create the database
Assuming you have created one or more config files you must now create the
SQL database. To do this, go to the Scratchy/scripts directory and
execute the following:
$ cd scripts
$ python create_database.py -c ../your-config-file
You are now ready to run the parser.
|
