Report Preparation - Who designs the reports?


#1

I’m still making my way through the documentation and code so please excuse me if I’ve missed something obvious.

My background is running a Risk system which does batch EOD reports. These are usually run once a day and the definition doesn’t change much.

So the users can see these reports the IT support staff need to do various things. We need to make sure that the securities/portfolios/historical time series/etc are fed into the system. This is reasonably straight forward from the existing documentation. I can see from the Java how I might create loaders and integrate with other data masters.

Now “reports” are “Views” in OG terminology. The whole process of running a view involves a “View Processor” (http://docs.opengamma.com/display/DOC/Views+and+Results ) which takes a “View Definition” and prepares the data specified.

I’m used to a system where the “View Definition” is created by users - the Risk Managers - through a web based UI. However I can’t see anything similar in OG.

In OG “They [ViewDefinitions] also can be created on-the-fly, programmatically, through the OpenGamma Tools for Microsoft Excel, or through OpenGamma’s R integration.”

I’m not keen on using Excel for defining reports - though I see how you might want to use it for one off reports. I don’t know anything about “R”.

So what options are left for users to create report/view definitions?

If users can’t, how do developers or other IT persons?

Cheers

Alex McLintock


#2

Hi, sorry we missed replying to this one.

The short answer is that we are working on a better web based UI. This includes view definition editing. You may be able to see it from a recent git checkout at url /jax/bundles/fm/prototype/index.ftl#/configs/ (liable to change)

Obviously, view definitions can also be created using Java code, simply by creating and populating the ViewDefinition class.


#3

Hi, Alex,

Another thing to bear in mind is that many of the reports that would commonly be part of an EOD run would logically come out of a batch process. So you’d set up a View Definition which contains all the metrics you’d need for the report, and then run that as part of an overnight process. Then you’d use your normal reporting stack (something like Mondrian or Microsoft Olap Services for cubing, and something like BIRT or Jasper on top) to generate the actual PDF reports.

At the moment we’ve focused far more on the data management and calculation support, but are looking to add more structured reporting integration and samples in the future.


#4

Just to add that the 0.9.0 release now contains much better view definition editing