Hi Marshall, I have successfully created an annotator and used it to get some analytics from the web-server logs. Then again, can I request you to have a quick look at the problem-description and the solution-used (as given below) ?
I would like to know if I am making use of the Annotator properly. [The goal is to generate user-behaviour patterns using UIMA] ------------------------------------------------------------------------------ Problem Statement: To generate the User-Behaviour patterns [like, the paths used by an user] from the web-server logs. (say, IIS-W3C-log-format) [Here the IPAddress is assumed as the user] ------------------------------------------------------------------------------ My solution to the problem description mentioned above involves the following steps[1 - 7]: Step 1 The latest stable version of the UIMA Java framework is down-loaded from the Apache web-site using the Subclipse. The UIMA is set-up on the Eclipse IDE for development and it is built using the Maven. Step 2 An Annotator is developed that consumes the web-server's log-file. Currently our annotator supports the IIS W3C-format for the web-server logs. ===> [Here I have used an Annotator very similar to the Room-Number Annotator] Step 3 A PEAR package is created for the annotator. Step 4 This PEAR package is fed as an input to the Simple-Server(a UIMA REST Service). The output of this service is an XML file that contains the data needed for our analysis. Step 5 An XML-DOM Parser parses the above XML output and store the output in a MySQL database. Step 6 The annotator and the parser will be provided as a JAR file for the end-user to be run using a scheduled-job, so that the processing of the web-server logs is automated. Step 7 The Resin web-server is used to run the JSP-based application to display the reports based on the analytics. ------------------------------------------------------------------------------- Doubts: If this is an acceptable way of using the Annotator, does it mean that in this specific scenario even a Shell-Script would do an equally good job ? ------------------------------------------------------------------------------- Can you please suggest any modifications that I can do to make my solution unique to UIMA ? Thankyou so much for your time. Kind Regards, Balkrishnan.V
