Specify pre-compute experimental metadata

Each owlsim run will have a number of parameters that are necessary to record in order to repeat the analysis. These need to be specified, and added to the schema of experimental metadata to be stored in DISCO per run. Not sure if they should be static, or just key-value pairs per experiment.

Suggestions:

*date of this experiment
*date ontologies were downloaded (or version number?)
*purl for ontologies used
*date data was downloaded
*url of data source(s) (the data will be coming from DISCO itself, it out to be a REST query, or perhaps we will have a special service)
*command run
*properties (in properties file? stored in db individually, or extraneous file? this includes thresholds.)
*any precomputed lcs cache used. this will need to be stored somewhere for future access, but also possibly itself stored in a table. perhaps we should have an ftp site for this? initially closed, but then opened when we really get the wheels turning. the lcs_cache generation itself probably is an experiment!
*algorithm/software version number (can use svn revision num for now)
*cluster information (do we want to keep track of this? such as number of nodes, number of entities per batch, etc.)
*summary IC stats (mean(min), mean(max), mean(mean), min(min), max(max), etc. per disease, per namespace, and overall)
*summary annot count stats (min/max/mean n per disease, per namespace, overall)
*entity 1 type (such as disease, genotype, etc.)
*species 1
*entity 2 type
*species 2
*annotation type used for comparison (such as phenotype, expr location)
*relations used
*comments