Return filename of the data file with simulation results

[ikajic]

This PR is a first stab at returning the filename of the data file being stored after running a simulation. This is achieved by adding the filename in the list of default items in the dictionary result maintained in the Benchmark.run().

I personally found this useful when running tons of simulations and wanting to read only a set of the data files which by default are stored in ./data using the Data class. This seemed to me a clearer solution than always reading the whole directory and extracting the files I'm interested in.

I'd be happy to improve this further if it is something generally useful.

[tcstewar]

Nifty. I've often done a similar thing with the data_dir option -- you can tell the system to store the data in a directory other than data.

[tcstewar]

Hmm.. do you actually need the result['filename'] column for anything? I guess I'm not seeing how you use this. How do you keep track of which filenames you want to include?

[ikajic]

I was able to do what I want by creating a new dir for every set of simulations, as you suggested. Since this doesn't require any changes to the existing code, it's a way to go for now. If you don't think the feature in the current PR is useful, feel free to close it, or we can leave it and see if it shows up in the future and then decide.